From subscript at free.fr Sun Aug 1 14:15:19 2004 From: subscript at free.fr (wwp) Date: Sun, 1 Aug 2004 16:15:19 +0200 Subject: extfs: .cab and .ace Message-ID: <20040801161519.55d64dad@tethys.solarsys.org> Hi folks, does anyone know about patches for .cab and .ace support to MC's extfs (maybe thru' cabextract and unace)? Regards, -- wwp From oskar at osk.mine.nu Mon Aug 2 10:08:44 2004 From: oskar at osk.mine.nu (Oskar Liljeblad) Date: Mon, 2 Aug 2004 12:08:44 +0200 Subject: hidden files feature request Message-ID: <20040802100844.GA4982@oskar> Hi! I have dotfiles visible ("show Hidden files") because I don't like it when a directory seems empty and is not. I think most experienced MC users use this option (or am I wrong?). But in directories such as $HOME there are usually quite a few dotfiles. This means that it's somewhat messy to navigate and modify the home directory. So I thought about this a little. What about adding some kind of indicator if the directory has hidden files (dot- or backup files)? And on top of this, add a ctrl keyboard bind to allow these files to be toggled visible or hidden. Or maybe add a directory entry saying "(hidden files)". Pressing enter over this directory entry would reveal the hidden files. Any thoughts? I was thinking about implementing this myself, but I probably won't find the time so I decided to write to the list first. Maybe someone else with time and inspiration might want to code this. Regards, Oskar Liljeblad (oskar at osk.mine.nu) From szaszg at hu.inter.net Wed Aug 4 05:58:45 2004 From: szaszg at hu.inter.net (Gergely Sza'sz,,,) Date: Wed, 4 Aug 2004 06:58:45 +0100 Subject: External scripting In-Reply-To: <20040802160117.98C5D3B08DE@menubar.gnome.org> References: <20040802160117.98C5D3B08DE@menubar.gnome.org> Message-ID: <20040804055845.GA15160@hu.inter.net> Hello! Here is a new patch. I make a minor change in code: when close_error_pipe is return 1 (some message on stderror from external script/command), and no original highliting and the 'Whole file if no highlite' option is checked then first unhighlite the temp. highlited file and next return. Gergely Szasz Here is the new diff file: diff -urx '*.gz' mc-4.6.0-pre2/edit/edit.orig/edit.c mc-4.6.0-pre2/edit/edit.c --- mc-4.6.0-pre2/edit/edit.orig/edit.c 2004-06-07 21:36:23.000000000 +0100 +++ mc-4.6.0-pre2/edit/edit.c 2004-08-03 22:02:41.000000000 +0100 @@ -2555,6 +2555,9 @@ case CK_Mail: edit_mail_dialog (edit); break; + case CK_Escript: + edit_escript_dialog (edit); + break; case CK_Shell: view_other_cmd (); break; diff -urx '*.gz' mc-4.6.0-pre2/edit/edit.orig/edit.h mc-4.6.0-pre2/edit/edit.h --- mc-4.6.0-pre2/edit/edit.orig/edit.h 2004-06-07 21:36:24.000000000 +0100 +++ mc-4.6.0-pre2/edit/edit.h 2004-08-03 22:02:41.000000000 +0100 @@ -253,6 +253,7 @@ void edit_insert_indent (WEdit *edit, int indent); void edit_options_dialog (void); void edit_mail_dialog (WEdit *edit); +void edit_escript_dialog (WEdit *edit); void format_paragraph (WEdit *edit, int force); /* either command or char_for_insertion must be passed as -1 */ diff -urx '*.gz' mc-4.6.0-pre2/edit/edit.orig/editcmd.c mc-4.6.0-pre2/edit/editcmd.c --- mc-4.6.0-pre2/edit/edit.orig/editcmd.c 2004-06-07 21:36:33.000000000 +0100 +++ mc-4.6.0-pre2/edit/editcmd.c 2004-08-03 22:02:41.000000000 +0100 @@ -885,6 +885,43 @@ #define space_width 1 static void +edit_insert_column (WEdit * edit, unsigned char data, int col, int width) +{ + if (data == '\n') { /* fill in and move to next line */ + int l; + long p; + + if (edit_get_byte (edit, edit->curs1) != '\n') { + l = width - (edit_get_col (edit) - col); + while (l > 0) { + edit_insert (edit, ' '); + l -= space_width; + } + } + for (p = edit->curs1;; p++) { + if (p == edit->last_byte) { + edit_cursor_move (edit, edit->last_byte - edit->curs1); + edit_insert_ahead (edit, '\n'); + p++; + break; + } + if (edit_get_byte (edit, p) == '\n') { + p++; + break; + } + } + edit_cursor_move (edit, edit_move_forward3 (edit, p, col, 0) - edit->curs1); + l = col - edit_get_col (edit); + while (l >= space_width) { + edit_insert (edit, ' '); + l -= space_width; + } + return; + } + edit_insert (edit, data); +} + +static void edit_insert_column_of_text (WEdit * edit, unsigned char *data, int size, int width) { long cursor; @@ -892,41 +929,31 @@ cursor = edit->curs1; col = edit_get_col (edit); for (i = 0; i < size; i++) { - if (data[i] == '\n') { /* fill in and move to next line */ - int l; - long p; - if (edit_get_byte (edit, edit->curs1) != '\n') { - l = width - (edit_get_col (edit) - col); - while (l > 0) { - edit_insert (edit, ' '); - l -= space_width; - } - } - for (p = edit->curs1;; p++) { - if (p == edit->last_byte) { - edit_cursor_move (edit, edit->last_byte - edit->curs1); - edit_insert_ahead (edit, '\n'); - p++; - break; - } - if (edit_get_byte (edit, p) == '\n') { - p++; - break; - } - } - edit_cursor_move (edit, edit_move_forward3 (edit, p, col, 0) - edit->curs1); - l = col - edit_get_col (edit); - while (l >= space_width) { - edit_insert (edit, ' '); - l -= space_width; - } - continue; - } - edit_insert (edit, data[i]); + edit_insert_column (edit, data[i], col, width); } edit_cursor_move (edit, cursor - edit->curs1); } +static void +edit_insert_column_from_file (WEdit * edit, char *filename, int width) +{ + int file; + + long cursor; + int i, col; + unsigned char data; + cursor = edit->curs1; + col = edit_get_col (edit); + + if ((file = open (filename, O_RDONLY | O_BINARY)) == -1) + return; + while ((i = read (file, &data, 1)) > 0) { + edit_insert_column (edit, data, col, width); + } + + edit_cursor_move (edit, cursor - edit->curs1); +} + void edit_block_copy_cmd (WEdit *edit) @@ -939,11 +966,11 @@ x = edit->curs_col; if (eval_marks (edit, &start_mark, &end_mark)) return; - if (column_highlighting) +/* if (column_highlighting) if ((x >= edit->column1 && x < edit->column2) || (x > edit->column2 && x <= edit->column1)) return; - +*/ copy_buf = edit_get_block (edit, start_mark, end_mark, &size); /* all that gets pushed are deletes hence little space is used on the stack */ @@ -1484,7 +1511,7 @@ } } else { /* regexp matching */ long offset = 0; - int found_start, match_bol, move_win = 0; + int found_start, match_bol, move_win = 0; while (start + offset < last_byte) { match_bol = (offset == 0 || (*get_byte) (data, start + offset - 1) == '\n'); @@ -2138,6 +2165,49 @@ return r; } +typedef void (*edit_iter_block_cb_t) (WEdit *edit, long start, long idx, + long finish, int pad, int ch, void *data); + +/* return next character of highlited block... */ +static void +edit_iter_block (WEdit *edit, long start, long finish, int pad, + edit_iter_block_cb_t cb, void *data) +{ + long idx = start; + + if (column_highlighting) { + int col1, col2; + + col1 = min (edit->column1, edit->column2); + col2 = max (edit->column1, edit->column2); + + if (edit_move_forward3 (edit, edit_bol (edit, idx), 0, idx) == col2) + idx++; + /* copy from buffer, excluding chars that are out of the column 'margins' */ + while (idx < finish) { + long i; + int c, x; + c = edit_get_byte (edit, idx); + x = edit_move_forward3 (edit, edit_bol (edit, idx), 0, idx); + if (x >= col1 && x < col2) { + if (c == '\n' && pad) + for (i = edit_move_forward3 (edit, edit_bol (edit, idx), + 0, idx); i < col2; i++) + cb (edit, start, idx, finish, pad, ' ', data); + cb (edit, start, idx, finish, pad, c, data); + } + else if (x == col2) + cb (edit, start, idx, finish, pad, '\n', data); + idx++; + } + } else { + while (idx < finish) { + cb (edit, start, idx, finish, pad, edit_get_byte (edit, idx), data); + idx++; + } + } +} + /* save block, returns 1 on success */ int edit_save_block (WEdit * edit, const char *filename, long start, @@ -2343,7 +2413,7 @@ exp = old ? old : ""; - exp = input_dialog (_(" Run Sort "), + exp = input_dialog (_(" Run Sort "), _(" Enter sort options (see manpage) separated by whitespace: "), exp); if (!exp) @@ -2355,12 +2425,12 @@ e = system (catstrs (" sort ", exp, " ", home_dir, BLOCK_FILE, " > ", home_dir, TEMP_FILE, 0)); if (e) { if (e == -1 || e == 127) { - edit_error_dialog (_(" Sort "), + edit_error_dialog (_(" Sort "), get_sys_error (_(" Cannot execute sort command "))); } else { char q[8]; sprintf (q, "%d ", e); - edit_error_dialog (_(" Sort "), + edit_error_dialog (_(" Sort "), catstrs (_(" Sort returned non-zero: "), q, 0)); } return -1; @@ -2608,6 +2678,196 @@ } +#define EXT_SCRIPT_BASIC 0 +#define EXT_SCRIPT_FILTER 1 + +#define EXT_SCRIPT_REPLACE 0 +#define EXT_SCRIPT_INSERT 1 +#define EXT_SCRIPT_DISCARD 2 + +void edit_ext_script_cb (WEdit *edit, long start, long idx, + long finish, int pad, int ch, void *data) +{ + FILE *p = data; + putc (ch, p); +} + +void +edit_ext_script_cmd (WEdit *edit, const char *scr, int type, + int sout, int ifnhi, int pad) +{ + long start_mark, end_mark; + FILE *script_home = NULL; + FILE *p = NULL; /* pipe */ + char *s = NULL; /* script */ + char *b = NULL; /* block file */ + char *q = NULL; /* quoted name */ + + if (eval_marks (edit, &start_mark, &end_mark)) { /* no highlite */ + if (ifnhi) { + edit_push_markers (edit); + edit_set_markers (edit, 0, edit->last_byte, 0, 0); + eval_marks (edit, &start_mark, &end_mark); + ifnhi = -1; + } else + sout = EXT_SCRIPT_DISCARD; + } + + b = catstrs (home_dir, BLOCK_FILE, 0); /* block file */ + q = name_quote (edit->filename, 0); + open_error_pipe (); + if (type == EXT_SCRIPT_BASIC) { + + s = catstrs (home_dir, EDIT_DIR "/", scr, 0); /* user script */ + if (!(script_home = fopen (s, "r"))) { + s = catstrs (home_dir, EDIT_DIR "/edit.", scr, ".rc", 0); /* user script short name */ + if (!(script_home = fopen (s, "r"))) { + s = catstrs (mc_home, scr, 0); /* system wide script */ + if (!(script_home = fopen (s, "r"))) { + s = catstrs (mc_home, "edit.", scr, ".rc", 0); /* system wide script */ + } + } + } + if (script_home) + fclose (script_home); + + edit_save_block (edit, b, start_mark, end_mark); + /* + * Run script. + * Initial space is to avoid polluting bash history. + * Arguments: + * $1 - name of the edited file (to check its extension etc). + * $2 - file containing the current block. + * $3 - file where error messages should be put + * (for compatibility with old scripts). + * $0 $1 $2 $3 */ + system (catstrs (" ", s, " ", q, " ", b, " /dev/null", NULL)); + + } else if (type == EXT_SCRIPT_FILTER) { + s = catstrs (" ", scr, "| cat >", b, 0); + if (*(s+1) == '|') /* oops, empty scr !!!*/ + *(s+1) = ' '; /* :-) */ + /* ok save stdout to BLOCK_FILE :-) */ + p = popen (s, "w"); + if (p) { + edit_iter_block (edit, start_mark, end_mark, pad, + edit_ext_script_cb, p); + pclose (p); + } + } + + g_free (q); + if (close_error_pipe (0, 0)) { + if (ifnhi == -1) + edit_set_markers (edit, 0, 0, 0, 0); + return; + } + + edit_refresh_cmd (edit); + edit->force |= REDRAW_COMPLETELY; + + /* insert result block */ + + switch (sout) { + case EXT_SCRIPT_REPLACE: + if (edit_block_delete_cmd (edit)) + return; + if (column_highlighting) /* Hmm.. Column selected */ + edit_cursor_move (edit, start_mark - edit->curs1); + case EXT_SCRIPT_INSERT: + if (column_highlighting) /* Hmm.. Column selected */ + edit_insert_column_from_file (edit, b, + max (edit->column2 - edit->column1, + edit->column1 - edit->column2)); + else + edit_insert_file (edit, b); + break; + case EXT_SCRIPT_DISCARD: + break; + } + +/* if ((block_file = fopen (b, "w"))) + fclose (block_file); */ + + return; +} + +#define EXT_SCRIPT_H 17 +#define EXT_SCRIPT_W 51 + +void edit_escript_dialog (WEdit * edit) +{ + static char *type_str[] = /* no translate :-( */ + {N_("Basic"), N_("Pipe"), NULL}; + static char *sout_str[] = /* no translate :-( */ + {N_("Replace"), N_("Insert"), N_("Discard"), NULL}; + + static char *script = 0, *filter = 0; + static int type = 0, sout = 0, ifnhi = 0, pad = 1; + + char *script_temp = script, *filter_temp = filter; + int type_temp = type, sout_temp = sout, ifnhi_temp = ifnhi, pad_temp = pad; + + QuickDialog Quick_input = + {EXT_SCRIPT_W, EXT_SCRIPT_H, -1, 0, N_(" External scripting "), + "[Input Line Keys]", 0}; + + QuickWidget quick_widgets[] = + { +/* 0*/ {quick_button, 6, 10, 15, EXT_SCRIPT_H, N_("&Cancel"), 0, B_CANCEL, 0, + 0, NULL}, +/* 1*/ {quick_button, 2, 10, 15, EXT_SCRIPT_H, N_("&OK"), 0, B_ENTER, 0, + 0, NULL}, +/* 2*/ {quick_radio, 4, EXT_SCRIPT_W, 11, EXT_SCRIPT_H, "", 3, sout, &sout, + sout_str, "extscript-dlg-stdout"}, +/* 3*/ {quick_label, 3, EXT_SCRIPT_W, 10, EXT_SCRIPT_H, N_("Output:"), 0, 0, 0, + 0, NULL}, +/* 4*/ {quick_checkbox, 3, EXT_SCRIPT_W, 8, EXT_SCRIPT_H, N_("Pad if column highliting"), 0, 0, 0, + 0, NULL}, +/* 5*/ {quick_checkbox, 3, EXT_SCRIPT_W, 7, EXT_SCRIPT_H, N_("Whole file if no highlited block"), 0, 0, 0, + 0, NULL}, +/* 6*/ {quick_input, 14, EXT_SCRIPT_W, 5, EXT_SCRIPT_H, "", EXT_SCRIPT_W - 17, 0, 0, + 0, "extscript-dlg-pipe"}, +/* 7*/ {quick_input, 14, EXT_SCRIPT_W, 4, EXT_SCRIPT_H, "", EXT_SCRIPT_W - 17, 0, 0, + 0, "extscript-dlg-script"}, +/* 8*/ {quick_radio, 4, EXT_SCRIPT_W, 4, EXT_SCRIPT_H, "", 2, type, &type, + type_str, "extscript-dlg-type"}, +/* 9*/ {quick_label, 3, EXT_SCRIPT_W, 3, EXT_SCRIPT_H, N_("Script:"), 0, 0, 0, + 0, NULL}, +/*10*/ {quick_label, 14, EXT_SCRIPT_W, 3, EXT_SCRIPT_H, N_("Name/command line"), 0, 0, 0, + 0, 0}, + {0}}; + + quick_widgets[2].result = &sout_temp; + quick_widgets[4].result = &pad_temp; + quick_widgets[5].result = &ifnhi_temp; + + quick_widgets[6].str_result = &filter_temp; + quick_widgets[6].text = filter_temp ? filter_temp : ""; + + quick_widgets[7].str_result = &script_temp; + quick_widgets[7].text = script_temp ? script_temp : ""; + + quick_widgets[8].result = &type_temp; + + Quick_input.widgets = quick_widgets; + + if (quick_dialog (&Quick_input) != B_CANCEL) { + if (script) + g_free (script); + if (filter) + g_free (filter); + type = type_temp; + sout = sout_temp; + ifnhi = ifnhi_temp; + pad = pad_temp; + script = script_temp; + filter = filter_temp; + edit_ext_script_cmd (edit, (type == EXT_SCRIPT_BASIC ? script : filter), + type, sout, ifnhi, pad); + } +} + /*******************/ /* Word Completion */ /*******************/ @@ -2617,7 +2877,7 @@ static int edit_find_word_start (WEdit *edit, long *word_start, int *word_len) { int i, c, last; - + /* return if at begin of file */ if (edit->curs1 <= 0) return 0; @@ -2625,14 +2885,14 @@ c = (unsigned char) edit_get_byte (edit, edit->curs1 - 1); /* return if not at end or in word */ if (isspace (c) || !(isalnum (c) || c == '_')) - return 0; + return 0; /* search start of word to be completed */ for (i = 2;; i++) { /* return if at begin of file */ - if (edit->curs1 - i < 0) + if (edit->curs1 - i < 0) return 0; - + last = c; c = (unsigned char) edit_get_byte (edit, edit->curs1 - i); diff -urx '*.gz' mc-4.6.0-pre2/edit/edit.orig/editcmddef.h mc-4.6.0-pre2/edit/editcmddef.h --- mc-4.6.0-pre2/edit/edit.orig/editcmddef.h 2004-06-07 21:36:34.000000000 +0100 +++ mc-4.6.0-pre2/edit/editcmddef.h 2004-08-03 22:02:41.000000000 +0100 @@ -98,6 +98,7 @@ #define CK_ExtCmd 424 #define CK_User_Menu 425 +#define CK_Escript 427 /* application control */ #define CK_Save_Desktop 451 #define CK_New_Window 452 diff -urx '*.gz' mc-4.6.0-pre2/edit/edit.orig/editkeys.c mc-4.6.0-pre2/edit/editkeys.c --- mc-4.6.0-pre2/edit/edit.orig/editkeys.c 2004-06-17 00:51:27.000000000 +0100 +++ mc-4.6.0-pre2/edit/editkeys.c 2004-08-03 22:02:41.000000000 +0100 @@ -36,6 +36,7 @@ static const long cooledit_key_map[] = { ALT ('b'), CK_Match_Bracket, ALT ('m'), CK_Mail, + ALT ('x'), CK_Escript, XCTRL ('f'), CK_Save_Block, XCTRL ('n'), CK_New, XCTRL ('p'), CK_Pipe_Block (1), /* spell check */ diff -urx '*.gz' mc-4.6.0-pre2/edit/edit.orig/editmenu.c mc-4.6.0-pre2/edit/editmenu.c --- mc-4.6.0-pre2/edit/edit.orig/editmenu.c 2004-06-07 21:36:35.000000000 +0100 +++ mc-4.6.0-pre2/edit/editmenu.c 2004-08-03 22:02:41.000000000 +0100 @@ -272,12 +272,19 @@ { edit_options_dialog (); } + static void menu_user_menu_cmd (void) { menu_key (KEY_F (11)); } +static void +menu_escript_cmd (void) +{ + menu_cmd (CK_Escript); +} + static menu_entry FileMenu[] = { {' ', N_("&Open file..."), 'O', menu_load_cmd}, @@ -363,6 +370,7 @@ {' ', N_("Sor&t... M-t"), 'T', menu_sort_cmd}, {' ', N_("Paste o&utput of... M-u"), 'U', menu_ext_cmd}, {' ', N_("E&xternal Formatter F19"), 'C', menu_c_form_cmd}, + {' ', N_("Exte&rnal Scripting M-x"), 'X', menu_escript_cmd}, {' ', N_("&Mail... "), 'M', menu_mail_cmd} }; @@ -387,6 +395,7 @@ {' ', N_("Sor&t... M-t"), 'T', menu_sort_cmd}, {' ', N_("Paste o&utput of... M-u"), 'U', menu_ext_cmd}, {' ', N_("E&xternal Formatter F19"), 'C', menu_c_form_cmd}, + {' ', N_("Exte&rnal Scripting M-x"), 'X', menu_escript_cmd}, {' ', N_("&Mail... "), 'M', menu_mail_cmd} }; From gjansman1 at hotmail.com Tue Aug 10 07:29:12 2004 From: gjansman1 at hotmail.com (G Jansman) Date: Tue, 10 Aug 2004 09:29:12 +0200 Subject: extfs: .cab and .ace Message-ID: A long time ago I've created an ucab script to support .cab (and uimg for support of FAT disk images). They can still be found in the mail archive. Since an improved cabextract version appeared I've improved the ucab script. Extracting files has become much faster. Both ucab versions don't support changing the archive (adding files etc.) or InstallShield .cab files. Anyhow, here's the new ucab script. How to install the script can be found in my first mail and in $(mcdir)/extfs/README. Anyhow, here's the ucab script: #! /bin/sh # # Written by Guus Jansman # # This is a parser for Cabinet archives in Midnight Commander. You need # the GPL cabextract program (version >= 1.0) written by Stuart Caie. # Limitations: # - Archives can only be viewed # - Problems with empty directories as created by MsCab (non-standard cab files) UNCAB=cabextract mccabfs_list () { $UNCAB -lq "$1" 2>/dev/null | gawk -v uuid=${UID-0} ' BEGIN { flag=0; date="JanFebMarAprMayJunJulAugSepOctNovDec" } /^-------/ { flag++; if (flag > 1) exit 0; next } /^$/ { next } { if (flag == 0) next perm="-rw-r--r--" uid=uuid gid=0 line=substr($0, index($0, "|")+2) day=substr(line, 1, 2) month=substr(date, (substr(line, 4, 2)-1)*3+1, 3) year=substr(line, 7, 4) hour=substr(line, 12, 2) minute=substr(line, 15, 2) size=$1 if (substr(size, length(size)) == "|") size=substr(size, 1, length(size)-1) name=substr(line, 23) gsub(/\\/, "/", name) if (substr(name, length(name)) == "/") { name=substr(name, 1, length(name)-1) perm="drwxr-xr-x" } printf "%s 1 %-8d %-8d %8d %3s %02d %04d %02d:%02d %s\n", perm, uid, gid, size, month, day, year, hour, minute, name }' } mccabfs_copyout () { $UNCAB -F "$2" -p "$1" > "$3" 2>/dev/null } umask 077 cmd="$1" shift case "$cmd" in list) mccabfs_list "$@" ;; copyout) mccabfs_copyout "$@" ;; *) exit 1 ;; esac exit 0 >From: wwp >To: mc-devel at gnome.org >Subject: extfs: .cab and .ace >Date: Sun, 1 Aug 2004 16:15:19 +0200 > >Hi folks, > > >does anyone know about patches for .cab and .ace support to MC's extfs >(maybe >thru' cabextract and unace)? > > >Regards, > >-- >wwp >_______________________________________________ >Mc-devel mailing list >http://mail.gnome.org/mailman/listinfo/mc-devel _________________________________________________________________ Ook een gouden buddy worden in Messenger? Go for gold! http://mobile.msn.com/?lc=nl-nl From subscript at free.fr Tue Aug 10 16:13:19 2004 From: subscript at free.fr (wwp) Date: Tue, 10 Aug 2004 18:13:19 +0200 Subject: extfs: .cab and .ace In-Reply-To: References: Message-ID: <20040810181319.6167e789@localhost> Hello G Jansman, On Tue, 10 Aug 2004 09:29:12 +0200 "G Jansman" wrote: > A long time ago I've created an ucab script to support .cab (and uimg for > support of FAT disk images). They can still be found in the mail archive. > > Since an improved cabextract version appeared I've improved the ucab script. > Extracting files has become much faster. Both ucab versions don't support > changing the archive (adding files etc.) or InstallShield .cab files. > > Anyhow, here's the new ucab script. How to install the script can be found > in my first mail and in $(mcdir)/extfs/README. > > Anyhow, here's the ucab script: [snip] Many thanks for this! Regards, -- wwp From pavelsh at mail.ru Thu Aug 12 06:15:10 2004 From: pavelsh at mail.ru (Pavel S. Shirshov) Date: Thu, 12 Aug 2004 12:15:10 +0600 Subject: Is command 'mktemp' portable? Message-ID: <429374955.20040812121510@mail.ru> Hello, I'm having patches for fix problem with security of mc. It's good candidate for commiting into CVS tree now. But, this patches was used mktemp, e.g. MAN=%{Enter name of man:} - TMPFILE=/tmp/mcview.$MAN.$$ + TMPFILE=`mktemp ${MC_TMPDIR:-/tmp}/mcview.$MAN.$$` || exit 1 man -Pcat $MAN >$TMPFILE mcview $TMPFILE rm -f $TMPFILE Is command 'mktemp' portable? P.S.: Sorry for bad English. -- Best regards, Pavel mailto:pavelsh at mail.ru From subscript at free.fr Thu Aug 12 08:54:55 2004 From: subscript at free.fr (wwp) Date: Thu, 12 Aug 2004 10:54:55 +0200 Subject: howto ignore /tmp/mc-/mc commands in history Message-ID: <20040812105455.7903caf3@localhost> Hi folks, using mc-4.6.1-pre1, I noticed that each time I run a command over a file thru the user menu, I will get a command more in console history like: /tmp/mc-wwp/mcusrzeAkyb I could not find a way to get rid of such itchy thing, but it reminds me the HISTCONTROL=ignoreboth I've set in my user profile, which doesn't help for this. Any hint? Regards, -- wwp From ossi at kde.org Thu Aug 12 09:39:56 2004 From: ossi at kde.org (Oswald Buddenhagen) Date: Thu, 12 Aug 2004 11:39:56 +0200 Subject: howto ignore /tmp/mc-/mc commands in history In-Reply-To: <20040812105455.7903caf3@localhost> References: <20040812105455.7903caf3@localhost> Message-ID: <20040812093956.GB8986@ugly.local> On Thu, Aug 12, 2004 at 10:54:55AM +0200, wwp wrote: > using mc-4.6.1-pre1, I noticed that each time I run a command over a file thru > the user menu, I will get a command more in console history like: > /tmp/mc-wwp/mcusrzeAkyb > > I could not find a way to get rid of such itchy thing, but it reminds me the > HISTCONTROL=ignoreboth I've set in my user profile, which doesn't help for > this. > > Any hint? > yes, i reported this bug about a year ago. the internal commands should be prefixed with a space. unfortunately, pavel seems to have disappeared, so mc development is again non-existing these days. i'd like to hear a statement from pavel, be it "i'm back in two months" or even a call for a new maintainer. -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. From leonard at den.ottolander.nl Thu Aug 12 14:38:30 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 12 Aug 2004 16:38:30 +0200 Subject: Is command 'mktemp' portable? In-Reply-To: <429374955.20040812121510@mail.ru> References: <429374955.20040812121510@mail.ru> Message-ID: <1092321509.4752.6.camel@athlon.localdomain> Hello Pavel, On Thu, 2004-08-12 at 08:15, Pavel S. Shirshov wrote: > I'm having patches for fix problem with security of mc. > It's good candidate for commiting into CVS tree now. Yes it is. I've gathered and compared the patches used by SUSE and Red Hat/Fedora and put them up here: http://www.ottolander.nl/mc-patches/proposed/ The mktemp patches are there. Sadly Pavel Roskin hasn't taken a look at these in over a month, although these patches have been used by multitudes of users of both distributions for months or even years. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pth at suse.de Thu Aug 12 15:08:40 2004 From: pth at suse.de (Philipp Thomas) Date: Thu, 12 Aug 2004 17:08:40 +0200 Subject: Is command 'mktemp' portable? In-Reply-To: <1092321509.4752.6.camel@athlon.localdomain> References: <429374955.20040812121510@mail.ru> <1092321509.4752.6.camel@athlon.localdomain> Message-ID: <20040812150840.GA7220@paradies.suse.de> * Leonard den Ottolander (leonard at den.ottolander.nl) [20040812 16:38]: > multitudes of users of both distributions for months or even years. In case of SUSE it's definitely years. Philipp -- Philipp Thomas SUSE LINUX AG, Maxfeldstr. 5, D-90409 Nuremberg, Germany From leonard at den.ottolander.nl Fri Aug 13 12:32:58 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 13 Aug 2004 14:32:58 +0200 Subject: Is command 'mktemp' portable? In-Reply-To: <449215486.20040813095508@mail.ru> References: <429374955.20040812121510@mail.ru> <1092321509.4752.6.camel@athlon.localdomain> <449215486.20040813095508@mail.ru> Message-ID: <1092400377.4750.13.camel@athlon.localdomain> Hello Pavel, On Thu, 2004-08-12 at 08:15, Pavel S. Shirshov wrote: > LdO> Yes it is. I've gathered and compared the patches used by SUSE and Red > LdO> Hat/Fedora and put them up here: > LdO> http://www.ottolander.nl/mc-patches/proposed/ > > LdO> The mktemp patches are there. Sadly Pavel Roskin hasn't taken a look at > LdO> these in over a month, although these patches have been used by > LdO> multitudes of users of both distributions for months or even years. > > Linux RedHat, Suse, etc it's not all kind of Unices in World. I do believe mktemp to be portable (I don't think the original fix comes from SUSE or Red Hat). All I was saying is that this particular solution has been proposed and used before (and quite possible on other Unices as well). Also I was drawing your attention to a set of patches by two Linux vendors that have been used for a long time and which I believe should be merged. Please have a look. Any incompatibilities that you find I am happy to address. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Fri Aug 13 22:21:36 2004 From: roland.illig at gmx.de (Roland Illig) Date: Sat, 14 Aug 2004 00:21:36 +0200 Subject: [patch] diff.syntax Message-ID: <411D3EF0.2020509@gmx.de> Hi, I'm new to mc development and I want to commit a bugfix patch for diff.syntax. It fixes wrong coloring for diffs of diffs. Any comments on this? Roland -------------- next part -------------- A non-text attachment was scrubbed... Name: diff.syntax-1.7-bugfix.patch Type: text/x-patch Size: 793 bytes Desc: not available URL: From roland.illig at gmx.de Sat Aug 14 11:56:41 2004 From: roland.illig at gmx.de (Roland Illig) Date: Sat, 14 Aug 2004 13:56:41 +0200 Subject: code cleanup suggestions Message-ID: <411DFDF9.9010302@gmx.de> Hi, I'd like to add a line #define NULL_QuickWidget { 0, 0, 0, 0, 0, NULL, 0, 0, NULL, NULL, NULL } into src/wtools.h after the typedef of QuickWidget. Having that done, one can write NULL_QuickWidget instead of {0} in a QuickWidget array. This will eliminate some GCC warnings about unspecified initializers. (I use gcc -W -Wall -W(many more)). I think it also looks nicer. Any comments? If not, I'll commit it soon. Roland From avpak at ccphys.nsu.ru Sun Aug 15 17:50:25 2004 From: avpak at ccphys.nsu.ru (avpak at ccphys.nsu.ru) Date: Mon, 16 Aug 2004 00:50:25 +0700 (NOVST) Subject: recode patch Message-ID: Here is a recode patch for mc cvs version. Features(avaliable with --enable-charset): Each panel has a codepage(selected by Ctrl-t when panel is active). Filenames are translated from panel codepage to display codepage before displaying on the screen. When copy or move operation occurs there is a possibility to translate filenames from one codepage to another (avaliable on copy/move dialog, initial codepages are taken from panel codepages) When move a directory within one filesystem only this directory name is translated but no traslation for it's files and subdirectories. translation of new directory name(F7) to active panel codepage. FTP default codepage(avaliable from Options->Display Bits). When user goes to ftp link panel codepage changes to FTP default codepage. When user goes back from ftp link panel codepage is restored. panel codepages, default ftp codepage are stored in ini-file and loaded when mc starts. when display codepage is not defined mc tries get it from $LANG. mc restores directory when user goes back from ftp link. Fixed "Other 8 bit" when uses presses esc in codepage selection dialog. Anton Pak -------------- next part -------------- A non-text attachment was scrubbed... Name: mc-cvs.recode.patch-20040816.gz Type: application/octet-stream Size: 11137 bytes Desc: URL: From roland.illig at gmx.de Sun Aug 15 23:40:49 2004 From: roland.illig at gmx.de (Roland Illig) Date: Mon, 16 Aug 2004 01:40:49 +0200 Subject: CVS commits and ChangeLog Message-ID: <411FF481.3010603@gmx.de> Hi, I have seen that you there is much redundancy in the CVS log. You mostly added your commit message to ChangeLog and commited it, too. I don't quite understand why this is better than just letting cvs2cl run over the CVS repository once for every release. To make it short: Why do you write your log message thrice, once in the ChangeLog file, once in the ChangeLog log and once in the real-file log? Do you want me to do the same? Roland From roland.illig at gmx.de Mon Aug 16 10:28:46 2004 From: roland.illig at gmx.de (Roland Illig) Date: Mon, 16 Aug 2004 12:28:46 +0200 Subject: fixing memory leaks Message-ID: <41208C5E.50104@gmx.de> Hi, I'm currently trying to find and fix some memory leaks. That's rather difficult because there are many function taking a "char *" argument when a "const char *" would have sufficed. Additionally, it is not always clear where the memory should be freed. To enlighten this situation, I would like to annotate the code with small comments that tell the reader about what a function does with its arguments. Here they are: /*in*/ the function does not modify the object /*free*/ the function will free the object /*out*/ the function expects an uninitialized object /*inout*/ the function expects an initialized object and modifies it /*new*/ the function returns a newly allocated object declaration example: /*new*/ char *g_strdup(/*in*/ const char *); function call example: char *s = /*new*/ g_strdup("foo"); I know that this will make the code more unreadable, but we could do a bug hunting phase where we introduce such comments, then a code cleanup phase where we have implicit notations: 1. every const argument is implicitly of type /*in*/. 2. functions generally do not free their arguments, so /*free*/ keeps on existing. 3. /*out*/ keeps on existing, as it is rarely used. 4. every non-const argument is implicitly of type /*inout*/. 5. /*new*/ keeps on existing, as it should be rarely used. What are you thinking about this? Roland From leonard at den.ottolander.nl Mon Aug 16 14:42:51 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Mon, 16 Aug 2004 16:42:51 +0200 Subject: [patch] diff.syntax In-Reply-To: <411D3EF0.2020509@gmx.de> References: <411D3EF0.2020509@gmx.de> Message-ID: <1092667354.4750.48.camel@athlon.localdomain> Hello Roland, On Sat, 2004-08-14 at 00:21, Roland Illig wrote: > > I'm new to mc development and I want to commit a bugfix patch for > diff.syntax. It fixes wrong coloring for diffs of diffs. > +context linestart ---\s \n brightmagenta > +context linestart \+\+\+\s \n brightmagenta These fixes seem obvious. > +context linestart \*\*\*\s \n brightmagenta When is this causing problems? Diff -c's of diff -c's? Haven't seen this yet. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Mon Aug 16 14:46:24 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Mon, 16 Aug 2004 16:46:24 +0200 Subject: code cleanup suggestions In-Reply-To: <411DFDF9.9010302@gmx.de> References: <411DFDF9.9010302@gmx.de> Message-ID: <1092667584.4750.51.camel@athlon.localdomain> Hello Roland, On Sat, 2004-08-14 at 13:56, Roland Illig wrote: > Any comments? If not, I'll commit it soon. Do you have commit access to CVS? In your last mail you said you are new to mc development... Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Mon Aug 16 15:01:47 2004 From: roland.illig at gmx.de (Roland Illig) Date: Mon, 16 Aug 2004 17:01:47 +0200 Subject: code cleanup suggestions In-Reply-To: <1092667584.4750.51.camel@athlon.localdomain> References: <411DFDF9.9010302@gmx.de> <1092667584.4750.51.camel@athlon.localdomain> Message-ID: <4120CC5B.1050000@gmx.de> Leonard den Ottolander wrote: > Hello Roland, > > On Sat, 2004-08-14 at 13:56, Roland Illig wrote: > >>Any comments? If not, I'll commit it soon. > > > Do you have commit access to CVS? In your last mail you said you are new > to mc development... Yes I have. But Pavel adviced me to first discuss it here. But last night, I felt like coding a bit. You can see the result in the CVS repository. Please have a look at it. I didn't introduce new bugs intentionally, but maybe I have overlooked something. Roland From leonard at den.ottolander.nl Mon Aug 16 17:30:45 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Mon, 16 Aug 2004 19:30:45 +0200 Subject: code cleanup suggestions In-Reply-To: <4120CC5B.1050000@gmx.de> References: <411DFDF9.9010302@gmx.de> <1092667584.4750.51.camel@athlon.localdomain> <4120CC5B.1050000@gmx.de> Message-ID: <1092677445.4750.5.camel@athlon.localdomain> Hello Roland, On Mon, 2004-08-16 at 17:01, Roland Illig wrote: > > Do you have commit access to CVS? In your last mail you said you are new > > to mc development... > > Yes I have. Good. So Pavel is not alone (anymore(?)). Since Pavel seems very busy maybe you could be so kind to have a look at the SUSE and Red Hat patches that I proposed for inclusion in CVS (http://www.ottolander.nl/mc-patches/proposed/)? These patches have been used by SUSE and/or Red Hat/Fedora for ages and some are security related. Lower in the tree you can find the original patches to compare. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Tue Aug 17 11:44:41 2004 From: pavelsh at mail.ru (Pavel S. Shirshov) Date: Tue, 17 Aug 2004 17:44:41 +0600 Subject: code cleanup suggestions In-Reply-To: <1092677445.4750.5.camel@athlon.localdomain> References: <411DFDF9.9010302@gmx.de> <1092667584.4750.51.camel@athlon.localdomain> <4120CC5B.1050000@gmx.de> <1092677445.4750.5.camel@athlon.localdomain> Message-ID: <1039626350.20040817174441@mail.ru> Hello Leonard, Monday, August 16, 2004, 11:30:45 PM, you wrote: >> > Do you have commit access to CVS? In your last mail you said you are new >> > to mc development... >> >> Yes I have. LdO> Good. So Pavel is not alone (anymore(?)). Since Pavel seems very busy LdO> maybe you could be so kind to have a look at the SUSE and Red Hat LdO> patches that I proposed for inclusion in CVS LdO> (http://www.ottolander.nl/mc-patches/proposed/)? These patches have been LdO> used by SUSE and/or Red Hat/Fedora for ages and some are security LdO> related. Lower in the tree you can find the original patches to compare. Current version is release candidate. Hence, security patches, and patches for ToDo list is needed. -- Best regards, Pavel mailto:pavelsh at mail.ru From pavelsh at mail.ru Tue Aug 17 11:50:48 2004 From: pavelsh at mail.ru (Pavel S. Shirshov) Date: Tue, 17 Aug 2004 17:50:48 +0600 Subject: security fix for temp files in cvs now Message-ID: <18210192520.20040817175048@mail.ru> Hello, I was commited security fixes based on patches from Leonard den Ottolander. -- Best regards, Pavel mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Tue Aug 17 14:22:31 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 17 Aug 2004 16:22:31 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <18210192520.20040817175048@mail.ru> References: <18210192520.20040817175048@mail.ru> Message-ID: <1092752550.4746.3.camel@athlon.localdomain> Hi Pavel, (Yet another developer with CVS access?) On Tue, 2004-08-17 at 13:50, Pavel S. Shirshov wrote: > I was commited security fixes based on patches from Leonard den > Ottolander. Thanks. Could you please tell which of the patches you applied? Just the tempfile patches? How about the other proposed patches? This way I can keep track of what to keep in proposed and what not. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Tue Aug 17 18:57:16 2004 From: pavelsh at mail.ru (pavelsh) Date: Wed, 18 Aug 2004 00:57:16 +0600 Subject: security fix for temp files in cvs now In-Reply-To: <1092752550.4746.3.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> Message-ID: <723105939.20040818005716@mail.ru> Hello Leonard, Tuesday, August 17, 2004, 8:22:31 PM, you wrote: LdO> (Yet another developer with CVS access?) Yes. LdO> Thanks. Could you please tell which of the patches you applied? LdO> Just the tempfile patches? Yes. LdO> How about the other proposed patches? I was commited mc-4.6.0-getpwuid.patch now. I should to see remain, tonight. LdO> This way I can keep track of what to keep in proposed and what not. Ok. Thx for your work! -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Tue Aug 17 20:28:30 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 17 Aug 2004 22:28:30 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <1189380635.20040818020712@mail.ru> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> Message-ID: <1092774510.4746.6.camel@athlon.localdomain> Hi Pavel, On Tue, 2004-08-17 at 22:07, pavelsh wrote: > stderr.patch commited Thanks for keeping me updated like this! If there are patches you do *not* want to commit please also let me (us) know. TIA. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 20:32:55 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 17 Aug 2004 22:32:55 +0200 Subject: mc-4.6.0-asmsyntax.patch In-Reply-To: <903872321.20040818023333@mail.ru> References: <903872321.20040818023333@mail.ru> Message-ID: <1092774774.4746.9.camel@athlon.localdomain> Hi Pavel, On Tue, 2004-08-17 at 22:33, pavelsh wrote: > mc-4.6.0-asmsyntax.patch is very incomplete for 80x86 asm keywords. Haven't really checked it thoroughly, just like to add more syntax files (if they are somewhat useful of course). Let's just leave this for later then (maybe). Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 20:57:50 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 17 Aug 2004 22:57:50 +0200 Subject: mc-CVE-CAN-2003-1023.patch In-Reply-To: <155898744.20040818022955@mail.ru> References: <155898744.20040818022955@mail.ru> Message-ID: <1092776270.4746.21.camel@athlon.localdomain> Hi Pavel, On Tue, 2004-08-17 at 22:29, pavelsh wrote: > mc-CVE-CAN-2003-1023.patch is obsolete in mc-4.6.1-pre > This error was fixed 2003-10-16 Pavel Roskin My mistake. Seems I actually forget to check this against CVS. If I am correct I did check the other proposed patches, although I might have missed some parts here and there. It got somewhat confusing on occasions, splitting the huge patch blobs from both Red Hat and SUSE, comparing them against one another and against CVS. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 21:46:13 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 17 Aug 2004 23:46:13 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <9210557191.20040818025714@mail.ru> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> Message-ID: <1092779172.4746.70.camel@athlon.localdomain> Hi Pavel, On Tue, 2004-08-17 at 22:57, pavelsh wrote: > *Not* > edit-replace.patch - UTF-8 unsupported now > mc-4.6.0-slang.patch - Likewise Ok. We should then discuss Vladimir's extensions to the UTF-8 patches then. He proposed them a while ago and they seem to work quite well, but Pavel Roskin had some reservations, that he did not explain however. > mc-4.6.0-absoluterm.patch - replace rm with /bin/rm Why? I guess to avoid failure when PATH isn't set (correctly). Jakub, do we need this included? > mc-4.6.0-asmsyntax.patch - it's need for completed So you said. Ok. Vladimir, maybe you can contact the original author? I once applied for SUSE bugzilla access, but never got a reply, so I'll leave it to you to check the original author ;) . > mc-4.6.0-pre3-nocpio.patch - itsn't good for main branch I thought it is, as it saves some time when digging into an rpm when you only want to look at the meta info. Also it unclutters the initial view, and actually the cpio part is redundant. When you show the cpio content directly you are mixing the contents of the cpio with the meta info, which I find unclean. And if you do there is no need to actually show the "file" CONTENTS.cpio which can then be entered again. Quite likely two identical temp files are created in that process. I know it is vendor specific patch, but I like this behaviour. Please reconsider. Note that there need to be made a few more fixes to vfs/extfs/rpm. For example, most $1's in mcrpmfs_copyout are replaced by $f's, but not on the $RPM2CPIO line (#165). But maybe the extra quotes (") aren't necessary there? Also, there is the temp file issue again (TMPDIR=/tmp/mctmpdir.$$) which needs fixing if it is kept. See also https://savannah.gnu.org/bugs/?func=detailitem&item_id=4271 in relation to this file. And what about the redundancy in vfs/extfs/trpm? Temporary file that needs to be removed from CVS? Or is it still in use? > mc-CVE-CAN-2003-1023.patch - obsoleted I've seen it. My mistake for proposing it :) . There are some hunks in the suse91 and fc1 branches of http://www.ottolander.nl/mc-patches/ that I still might put up for proposal. > *May be* > mc-4.6.0-large_syntax.patch - Take now for review Jakub fixed the patch from https://savannah.gnu.org/patch/?func=detailitem&item_id=1628 . I believe the original patch had some pointer arithmetic issues. We have been using it for months with the php syntax file on Fedora Core. Note that without this fix large syntax files are useless. > mc-php.syntax - May you update this file? > PHP 5 is released. I'll see if the script still works against the docs, and whether the docs are updated. Do we need different syntax files for different versions? > *Already* > mc-4.6.0-getpwuid.patch > > stderr.patch > > tempfile.patch Ok. Thanks for your effort. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 21:58:20 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 17 Aug 2004 23:58:20 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <1092779172.4746.70.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> Message-ID: <1092779900.4746.80.camel@athlon.localdomain> Hi, On Tue, 2004-08-17 at 23:46, Leonard den Ottolander wrote: > > mc-php.syntax - May you update this file? > > PHP 5 is released. > > I'll see if the script still works against the docs, and whether the > docs are updated. Do we need different syntax files for different > versions? Note that the script I used to update this syntax file only extracts function calls from the docs. I added these to the original syntax file (apart from maybe some small other fixes). I haven't been following php development closely, so I might not be the best person to implement the *syntax* changes. I will however see if I can extract a new function list. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 22:29:38 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 00:29:38 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <1092779172.4746.70.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> Message-ID: <1092781777.6246.2.camel@athlon.localdomain> On Tue, 2004-08-17 at 23:46, Leonard den Ottolander wrote: > See also https://savannah.gnu.org/bugs/?func=detailitem&item_id=4271 in > relation to this file. Note that the bug reporter is probably half correct. INFO/OBSOLETES is probably not always empty, just few packages use it, and I don't have any around that I know of they do, so I can't check right now. However, he seems to be correct on the INFO/LICENSE issue. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 22:34:55 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 00:34:55 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <09951226.20040818042747@mail.ru> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092779900.4746.80.camel@athlon.localdomain> <09951226.20040818042747@mail.ru> Message-ID: <1092782094.6246.7.camel@athlon.localdomain> Hi Pavel, On Wed, 2004-08-18 at 00:27, pavelsh wrote: > Your script was return output below > DOMDocument->saveHTML > DOMElement->getAttributeNS > It's all ok? Is class name, method name, property name different > words? I am glad it still outputs something ;) . As you can see my patch was rather ad hoc against the then current documentation. Some things have obviously changed. I'll have a look this or next week. Looks fixable. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 17 22:49:47 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 00:49:47 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <1092781777.6246.2.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092781777.6246.2.camel@athlon.localdomain> Message-ID: <1092782986.6246.16.camel@athlon.localdomain> On Wed, 2004-08-18 at 00:29, Leonard den Ottolander wrote: > > See also https://savannah.gnu.org/bugs/?func=detailitem&item_id=4271 in > > relation to this file. > > Note that the bug reporter is probably half correct. INFO/OBSOLETES is > probably not always empty, just few packages use it, and I don't have > any around that I know of they do, so I can't check right now. > > However, he seems to be correct on the INFO/LICENSE issue. Duh. Topsy turvy. It's been too long. INFO/OBSOLETES is missing and needs to be added. That part is correct. However, although I don't oppose also adding INFO/LICENSE dropping INFO/COPYRIGHT is not a good idea. Don't apply that patch ad verbatim. Also note "$1" -> "$f". Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Tue Aug 17 23:10:10 2004 From: pavelsh at mail.ru (pavelsh) Date: Wed, 18 Aug 2004 05:10:10 +0600 Subject: security fix for temp files in cvs now In-Reply-To: <1092779172.4746.70.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> Message-ID: <14410233326.20040818051010@mail.ru> Hello Leonard, Wednesday, August 18, 2004, 3:46:13 AM, you wrote: >> *Not* >> edit-replace.patch - UTF-8 unsupported now >> mc-4.6.0-slang.patch - Likewise LdO> Ok. We should then discuss Vladimir's extensions to the UTF-8 patches LdO> then. He proposed them a while ago and they seem to work quite well, but LdO> Pavel Roskin had some reservations, that he did not explain however. Support for multibyte in the 4.7 branch (according TODO). >> *May be* >> mc-4.6.0-large_syntax.patch - Take now for review LdO> Jakub fixed the patch from LdO> https://savannah.gnu.org/patch/?func=detailitem&item_id=1628 . I believe LdO> the original patch had some pointer arithmetic issues. We have been LdO> using it for months with the php syntax file on Fedora Core. Note that LdO> without this fix large syntax files are useless. I'll commit it after tests. >> mc-php.syntax - May you update this file? >> PHP 5 is released. LdO> I'll see if the script still works against the docs, and whether the LdO> docs are updated. Do we need different syntax files for different LdO> versions? Problem is to distinguish src files from one version php to another. -- Best regards, pavelsh mailto:pavelsh at mail.ru From pavelsh at mail.ru Tue Aug 17 23:15:05 2004 From: pavelsh at mail.ru (pavelsh) Date: Wed, 18 Aug 2004 05:15:05 +0600 Subject: security fix for temp files in cvs now In-Reply-To: <1092782986.6246.16.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092781777.6246.2.camel@athlon.localdomain> <1092782986.6246.16.camel@athlon.localdomain> Message-ID: <613824307.20040818051505@mail.ru> Hello Leonard, Wednesday, August 18, 2004, 4:49:47 AM, you wrote: >> > See also >> https://savannah.gnu.org/bugs/?func=detailitem&item_id=4271 in >> > relation to this file. >> >> Note that the bug reporter is probably half correct. INFO/OBSOLETES is >> probably not always empty, just few packages use it, and I don't have >> any around that I know of they do, so I can't check right now. >> >> However, he seems to be correct on the INFO/LICENSE issue. LdO> Duh. Topsy turvy. It's been too long. INFO/OBSOLETES is missing and LdO> needs to be added. That part is correct. However, although I don't LdO> oppose also adding INFO/LICENSE dropping INFO/COPYRIGHT is not a good LdO> idea. Don't apply that patch ad verbatim. Also note "$1" -> "$f". Please, make single patch with all diff. And i'll review it. Ok? -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Tue Aug 17 23:20:53 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 01:20:53 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <14410233326.20040818051010@mail.ru> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <14410233326.20040818051010@mail.ru> Message-ID: <1092784853.6246.37.camel@athlon.localdomain> Hello Pavel, On Wed, 2004-08-18 at 01:10, pavelsh wrote: > Support for multibyte in the 4.7 branch (according TODO). I guess SUSE and Red Hat/Fedora could pretest these some more. > >> mc-php.syntax - May you update this file? Thinking a bit about "DOMDocument->saveHTML". Although "->" is a separate syntax element you might want to use that whole string as an identifier to avoid highlighting of wrong class/method combinations... > Problem is to distinguish src files from one version php to > another. Yeah, that's an RFE: View->Highlight menu option to override the syntax filed used, like in gedit. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Wed Aug 18 08:28:40 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Wed, 18 Aug 2004 12:28:40 +0400 Subject: trivial compile fix Message-ID: <200408181228.40521.dmi_a@qnx.org.ru> Hi, mc from CVS fails to compile complainig on type incompatability in command.[ch], declaration of command_insert () diff: --- mc/src/command.h.orig 2004-08-18 12:23:20 +0400 +++ mc/src/command.h 2004-08-18 12:23:53 +0400 @@ -5,6 +5,6 @@ WInput *command_new (int y, int x, int len); void do_cd_command (char *cmd); -void command_insert (WInput * in, char *text, int insert_extra_space); +void command_insert (WInput * in, const char *text, int insert_extra_space); #endif /* __COMMAND_H */ WBR Dmitry From roland.illig at gmx.de Wed Aug 18 08:58:48 2004 From: roland.illig at gmx.de (Roland Illig) Date: Wed, 18 Aug 2004 10:58:48 +0200 Subject: trivial compile fix In-Reply-To: <200408181228.40521.dmi_a@qnx.org.ru> References: <200408181228.40521.dmi_a@qnx.org.ru> Message-ID: <41231A48.7090606@gmx.de> Dmitry Alexeyev wrote: > Hi, > > mc from CVS fails to compile complainig on type incompatability in > command.[ch], declaration of command_insert () Sorry that I forgot to commit this change. I fixed it. Roland From jakub at redhat.com Tue Aug 17 19:56:23 2004 From: jakub at redhat.com (Jakub Jelinek) Date: Tue, 17 Aug 2004 21:56:23 +0200 Subject: security fix for temp files in cvs now In-Reply-To: <1092779172.4746.70.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> Message-ID: <20040817195623.GO30497@sunsite.ms.mff.cuni.cz> On Tue, Aug 17, 2004 at 11:46:13PM +0200, Leonard den Ottolander wrote: > > mc-4.6.0-pre3-nocpio.patch - itsn't good for main branch > > I thought it is, as it saves some time when digging into an rpm when you > only want to look at the meta info. Also it unclutters the initial view, > and actually the cpio part is redundant. When you show the cpio content > directly you are mixing the contents of the cpio with the meta info, > which I find unclean. And if you do there is no need to actually show > the "file" CONTENTS.cpio which can then be entered again. Quite likely > two identical temp files are created in that process. I know it is > vendor specific patch, but I like this behaviour. Please reconsider. The reason for this patch is mainly that without that patch people tend to copy the files in rpm's main directory and not CONTENTS.cpio out of the rpm. And that is horribly slow, especially for large rpms (as mc for each single file rpm2cpio's the whole rpm (which means ungzipping or unbzip2ing the whole payload) and then runs cpio again on the whole unpacked file. When CONTENTS.cpio is entered, it is unpacked just once and given that there is an internal cpio filesystem, even that unpacked file doesn't have to be searched through all the time. Jakub From leonard at den.ottolander.nl Wed Aug 18 13:18:03 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 15:18:03 +0200 Subject: fixing memory leaks In-Reply-To: <41208C5E.50104@gmx.de> References: <41208C5E.50104@gmx.de> Message-ID: <1092835082.4777.44.camel@athlon.localdomain> Hi Roland, On Mon, 2004-08-16 at 12:28, Roland Illig wrote: > I'm currently trying to find and fix some memory leaks. That's rather > difficult because there are many function taking a "char *" argument > when a "const char *" would have sufficed. Additionally, it is not > always clear where the memory should be freed. > declaration example: > /*new*/ char *g_strdup(/*in*/ const char *); > > function call example: > char *s = /*new*/ g_strdup("foo"); Although I appreciate the effort towards code cleanup I am not sure if it's a good idea to commit such comments to CVS. Partially because the code might look very cluttered (maybe I am overestimating this), and more importantly the fact that it might heavily interfere with existing but uncommitted patches. How much of the freeing and mallocing is done spanning multiple source files? This kind of auditing could be done using local copies with the comments added. Just keep a list of what has been audited. If you want to do this in CVS maybe it's a good idea to wait until bugzilla has been cleaned up a bit more before introducing such comments into the code. By the way, do you want to add those temporarily or permanently? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Wed Aug 18 13:29:52 2004 From: roland.illig at gmx.de (Roland Illig) Date: Wed, 18 Aug 2004 15:29:52 +0200 Subject: fixing memory leaks In-Reply-To: <1092835082.4777.44.camel@athlon.localdomain> References: <41208C5E.50104@gmx.de> <1092835082.4777.44.camel@athlon.localdomain> Message-ID: <412359D0.6090509@gmx.de> Leonard den Ottolander wrote: > Hi Roland, > > On Mon, 2004-08-16 at 12:28, Roland Illig wrote: > >>I'm currently trying to find and fix some memory leaks. That's rather >>difficult because there are many function taking a "char *" argument >>when a "const char *" would have sufficed. Additionally, it is not >>always clear where the memory should be freed. > > >>declaration example: >>/*new*/ char *g_strdup(/*in*/ const char *); >> >>function call example: >>char *s = /*new*/ g_strdup("foo"); > > > Although I appreciate the effort towards code cleanup I am not sure if > it's a good idea to commit such comments to CVS. Partially because the > code might look very cluttered (maybe I am overestimating this), and > more importantly the fact that it might heavily interfere with existing > but uncommitted patches. > > How much of the freeing and mallocing is done spanning multiple source > files? This kind of auditing could be done using local copies with the > comments added. Just keep a list of what has been audited. > > If you want to do this in CVS maybe it's a good idea to wait until > bugzilla has been cleaned up a bit more before introducing such comments > into the code. By the way, do you want to add those temporarily or > permanently? They should be only temporary. As I wrote in the second part of my mail, after having documented the whole source, the most common usages should have no extra comments. In most times, a const char * means the same as a possible /*in*/ char *, and so the /*in*/ can be replaced by the language keyword const. The other case is passing a char **. I currently do not know what is more frequent: an /*inout*/ parameter or a simple /*out*/ parameter. I guess it's the /*out*/ parameter, so we could agree that a missing comment means "can be uninitialized and will receive a value". Perhaps we can also replace the /*free*/ comment by function names that contain the word "free". Same goes for the /*new*/ comment. So the only comment that would appear after these substitutions would be either /*out*/ or /*inout*/. And that's not too much, I think. Roland From leonard at den.ottolander.nl Wed Aug 18 13:45:26 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 15:45:26 +0200 Subject: INFO/OBSOLETES, -/LICENSE & -/COPYRIGHT In-Reply-To: <1092782986.6246.16.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092781777.6246.2.camel@athlon.localdomain> <1092782986.6246.16.camel@athlon.localdomain> Message-ID: <1092836726.4777.51.camel@athlon.localdomain> Hi, On Wed, 2004-08-18 at 00:49, Leonard den Ottolander wrote: > > > See also https://savannah.gnu.org/bugs/?func=detailitem&item_id=4271 in > > However, he seems to be correct on the INFO/LICENSE issue. > > Duh. Topsy turvy. It's been too long. INFO/OBSOLETES is missing and > needs to be added. That part is correct. However, although I don't > oppose also adding INFO/LICENSE dropping INFO/COPYRIGHT is not a good > idea. Don't apply that patch ad verbatim. Also note "$1" -> "$f". Looking once more it seems "License:" is indeed the tag most used. Seems the bug report is correct after all. This means both INFO/OBSOLETES and INFO/LICENSE should be introduced in the extfs rpm file. The question then remains if INFO/COPYRIGHT should be kept around or dropped. When I have an answer to that question I'll put a patch in "proposed". Leonard. -- mount -t life -o ro /dev/dna /genetic/research From ossi at kde.org Wed Aug 18 14:13:15 2004 From: ossi at kde.org (Oswald Buddenhagen) Date: Wed, 18 Aug 2004 16:13:15 +0200 Subject: fixing memory leaks In-Reply-To: <41208C5E.50104@gmx.de> References: <41208C5E.50104@gmx.de> Message-ID: <20040818141315.GA942@ugly.local> On Mon, Aug 16, 2004 at 12:28:46PM +0200, Roland Illig wrote: > That's rather difficult because there are many function taking a "char > *" argument when a "const char *" would have sufficed. > how about fixing this first? > What are you thinking about this? > you could just use valgrind ... of course there is the audit vs. testing tradeoff, but let's not go overboard with formal correctness. -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. From leonard at den.ottolander.nl Wed Aug 18 14:31:48 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 16:31:48 +0200 Subject: deb vfs security issue (CAN-2004-0494) Message-ID: <1092839508.4777.57.camel@athlon.localdomain> Hi, I noticed this in Red Hat's bugzilla just now: http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=127973 . Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Wed Aug 18 15:17:44 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 17:17:44 +0200 Subject: PHP syntax [was: Re: Re[6]: security fix for temp files in cvs now] In-Reply-To: <1092784853.6246.37.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <14410233326.20040818051010@mail.ru> <1092784853.6246.37.camel@athlon.localdomain> Message-ID: <1092842264.4777.71.camel@athlon.localdomain> Hi Pavel, On Wed, 2004-08-18 at 01:20, Leonard den Ottolander wrote: > On Wed, 2004-08-18 at 01:10, pavelsh wrote: > > > >> mc-php.syntax - May you update this file? > > Thinking a bit about "DOMDocument->saveHTML". Although "->" is a > separate syntax element you might want to use that whole string as an > identifier to avoid highlighting of wrong class/method combinations... Looking at this I stumbled upon a "[deprecated]" tag in the syntax file that slipped through the script's filters. However that tag is no longer available in the docs so you probably haven't noticed it ;) . Removed the syntax file from proposed for now until it is cleaned up. It's still available under the fc1 branch. I think keeping the whole strings including the "->"'s is a good idea, as it only highlights correct class/method combinations. Sadly in some cases a separate class description seems to be failing from the docs, which makes it necessary to add those by hand. For example in the current php syntax file there is no entry for DOMDocument. There is one for hw_api_attribute though. For a moment I thought the keywords should be sorted "largest match first" but this seems to be unnecessary. Thus it doesn't matter hw_api_attribute->values comes after hw_api_attribute->value. I also believe it is probably a good idea to add a separate keyword entry for "->". Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Wed Aug 18 15:23:18 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 17:23:18 +0200 Subject: rpm vfs [was: Re: Re[6]: security fix for temp files in cvs now] In-Reply-To: <613824307.20040818051505@mail.ru> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092781777.6246.2.camel@athlon.localdomain> <1092782986.6246.16.camel@athlon.localdomain> <613824307.20040818051505@mail.ru> Message-ID: <1092842598.4777.76.camel@athlon.localdomain> Hi Pavel, On Wed, 2004-08-18 at 01:15, pavelsh wrote: > Please, make single patch with all diff. And i'll review it. > Ok? Cleaned up the patch tree at http://www.ottolander.nl/mc-patches/ according to the latest developments. Added the extfs_rpm.patch to the proposed branch. It adds INFO/OBSOLETES and INFO/LICENSE, but leaves INFO/COPYRIGHT alone (should probably be kept around while rpm supports the tag). Also included in that patch is a minor fix concerning rpm upgrade (-iUvh replaced with -Uvh). Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Wed Aug 18 17:28:23 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 19:28:23 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092839508.4777.57.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> Message-ID: <1092850103.4777.83.camel@athlon.localdomain> Hi, On Wed, 2004-08-18 at 16:31, Leonard den Ottolander wrote: > I noticed this in Red Hat's bugzilla just now: > http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=127973 . Attached is a patch that escapes all dangerous characters for function arguments. More specifically, everything not in A-Z, a-z, 0-9, _, /, ., - and +. Could somebody on a system with dpkg installed verify that things still work correctly after applying this patch? Thanks. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deb.diff Type: text/x-patch Size: 1021 bytes Desc: not available URL: From leonard at den.ottolander.nl Wed Aug 18 17:42:55 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 19:42:55 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <20040818152209.GS30497@sunsite.ms.mff.cuni.cz> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <20040818152209.GS30497@sunsite.ms.mff.cuni.cz> Message-ID: <1092850975.4777.87.camel@athlon.localdomain> Hi Jakub, > There are many other scripts which need similar treatment. > grep -l bin/perl /usr/share/mc/extfs/* | xargs grep open > shows a lot of (potential) problems, at least in a, apt, debd, mailfs, > patchfs, rpms and uzip. I thought so :) . But I'd rather first hear this is a correct fix before applying it to all of the above. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Wed Aug 18 18:54:29 2004 From: roland.illig at gmx.de (Roland Illig) Date: Wed, 18 Aug 2004 20:54:29 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092850103.4777.83.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> Message-ID: <4123A5E5.5010501@gmx.de> Leonard den Ottolander wrote: > Hi, > > On Wed, 2004-08-18 at 16:31, Leonard den Ottolander wrote: > >>I noticed this in Red Hat's bugzilla just now: >>http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=127973 . > > Could somebody on a system with dpkg installed verify that things still > work correctly after applying this patch? Thanks. I successfully opened a .deb package with the patched midnight commander. It works fine. My approach to this problem has been to write a safe_system(...) subroutine for Perl that is not as complicated as IPC::Open3. ;) Roland From leonard at den.ottolander.nl Wed Aug 18 19:05:24 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 21:05:24 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <4123A5E5.5010501@gmx.de> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <4123A5E5.5010501@gmx.de> Message-ID: <1092855924.4777.98.camel@athlon.localdomain> On Wed, 2004-08-18 at 20:54, Roland Illig wrote: > I successfully opened a .deb package with the patched midnight > commander. It works fine. No problems with copying files from the .deb either? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Wed Aug 18 19:15:25 2004 From: roland.illig at gmx.de (Roland Illig) Date: Wed, 18 Aug 2004 21:15:25 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092855924.4777.98.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <4123A5E5.5010501@gmx.de> <1092855924.4777.98.camel@athlon.localdomain> Message-ID: <4123AACD.7080208@gmx.de> Leonard den Ottolander wrote: > On Wed, 2004-08-18 at 20:54, Roland Illig wrote: > >>I successfully opened a .deb package with the patched midnight >>commander. It works fine. > > > No problems with copying files from the .deb either? Works, too. From leonard at den.ottolander.nl Wed Aug 18 19:27:30 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 21:27:30 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <20040818152209.GS30497@sunsite.ms.mff.cuni.cz> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <20040818152209.GS30497@sunsite.ms.mff.cuni.cz> Message-ID: <1092857250.4777.123.camel@athlon.localdomain> Hi Jakub, On Wed, 2004-08-18 at 17:22, Jakub Jelinek wrote: > There are many other scripts which need similar treatment. > grep -l bin/perl /usr/share/mc/extfs/* | xargs grep open > shows a lot of (potential) problems, at least in a, apt, debd, mailfs, > patchfs, rpms and uzip. a looks vulnerable. This script uses $disk:/$path. Hm. Very dozy. What files is it used for anyway? apt uses the result of a find. Probably vulnerable. Also uses the output of an "apt-cache dumpavail". Maybe somebody could enlighten me on this command, but I think it could use escaping anyway. And an unchecked $file. Bad script! Bad! deba also vulnerable. debd idem. dpkg idem. mailfs idem. patchfs idem. rpms uses an unchecked $ARGS[3]. Looks vulnerable. And last but not least: uzip uses map(quotemeta, ). quotemeta is even a bit more restrictive than the substitution I use (only leave A-Z, a-z, 0-9 and _, ie \w alone). So this is the only script *not* vulnerable. Maybe somebody could explain the use of these deb. and dpkg files are for? And maybe somebody with on Debian could check if these are indeed vulnerable like "deb"? And what does rpms do? As a side note: What about trpm? Is that still in use? Patches are in the making. I assume these should be against the source tree, ie the .in files instead of against the compiled versions as my previous patch is. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Wed Aug 18 19:33:28 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 21:33:28 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <20040818164036.GT30497@sunsite.ms.mff.cuni.cz> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <20040818164036.GT30497@sunsite.ms.mff.cuni.cz> Message-ID: <1092857607.4777.125.camel@athlon.localdomain> On Wed, 2004-08-18 at 18:40, Jakub Jelinek wrote: > Also, isn't mcdebfs_copyout's destfile not used just in system () > (where it should be escaped), but also in > if ( open(FILEOUT,">$destfile") ) > (where I'd say it should not be escaped)? Why not there? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From andrew at email.zp.ua Wed Aug 18 19:35:23 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Wed, 18 Aug 2004 22:35:23 +0300 (EEST) Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092850975.4777.87.camel@athlon.localdomain> Message-ID: <200408181935.i7IJZNCq037163@email.zp.ua> > Hi Jakub, > > > There are many other scripts which need similar treatment. > > grep -l bin/perl /usr/share/mc/extfs/* | xargs grep open > > shows a lot of (potential) problems, at least in a, apt, debd, mailfs, > > patchfs, rpms and uzip. patchfs and uzip is ok ;-) -- Regards, Andrew V. Samoilov. From leonard at den.ottolander.nl Wed Aug 18 21:00:25 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 23:00:25 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <200408181935.i7IJZNCq037163@email.zp.ua> References: <200408181935.i7IJZNCq037163@email.zp.ua> Message-ID: <1092862825.4777.228.camel@athlon.localdomain> Hi Andrew, On Wed, 2004-08-18 at 21:35, Andrew V. Samoilov wrote: > patchfs and uzip is ok ;-) I see. copyin is passed unchecked parameters, but those are quotemeta'd with myin. This seems to be the case in most opens, except one: copyout. Are you sure 'open 0, "> $out";' is fine? Spaces in file names seem not to be handled correctly by patchfs... Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Wed Aug 18 21:24:36 2004 From: pavelsh at mail.ru (pavelsh) Date: Thu, 19 Aug 2004 03:24:36 +0600 Subject: rpm vfs [was: Re: Re[6]: security fix for temp files in cvs now] In-Reply-To: <1092842598.4777.76.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092781777.6246.2.camel@athlon.localdomain> <1092782986.6246.16.camel@athlon.localdomain> <613824307.20040818051505@mail.ru> <1092842598.4777.76.camel@athlon.localdomain> Message-ID: <1929795893.20040819032436@mail.ru> Hello Leonard, Wednesday, August 18, 2004, 9:23:18 PM, you wrote: LdO> On Wed, 2004-08-18 at 01:15, pavelsh wrote: >> Please, make single patch with all diff. And i'll review it. >> Ok? LdO> Cleaned up the patch tree at http://www.ottolander.nl/mc-patches/ LdO> according to the latest developments. Added the extfs_rpm.patch to the LdO> proposed branch. It adds INFO/OBSOLETES and INFO/LICENSE, but leaves LdO> INFO/COPYRIGHT alone (should probably be kept around while rpm supports LdO> the tag). LdO> Also included in that patch is a minor fix concerning rpm upgrade (-iUvh LdO> replaced with -Uvh). Commited. -- Best regards, pavelsh mailto:pavelsh at mail.ru From jakub at redhat.com Wed Aug 18 15:22:09 2004 From: jakub at redhat.com (Jakub Jelinek) Date: Wed, 18 Aug 2004 17:22:09 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092850103.4777.83.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> Message-ID: <20040818152209.GS30497@sunsite.ms.mff.cuni.cz> On Wed, Aug 18, 2004 at 07:28:23PM +0200, Leonard den Ottolander wrote: > Hi, > > On Wed, 2004-08-18 at 16:31, Leonard den Ottolander wrote: > > I noticed this in Red Hat's bugzilla just now: > > http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=127973 . > > Attached is a patch that escapes all dangerous characters for function > arguments. More specifically, everything not in A-Z, a-z, 0-9, _, /, ., > - and +. > > Could somebody on a system with dpkg installed verify that things still > work correctly after applying this patch? Thanks. There are many other scripts which need similar treatment. grep -l bin/perl /usr/share/mc/extfs/* | xargs grep open shows a lot of (potential) problems, at least in a, apt, debd, mailfs, patchfs, rpms and uzip. Jakub From jakub at redhat.com Wed Aug 18 16:40:36 2004 From: jakub at redhat.com (Jakub Jelinek) Date: Wed, 18 Aug 2004 18:40:36 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092850103.4777.83.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> Message-ID: <20040818164036.GT30497@sunsite.ms.mff.cuni.cz> On Wed, Aug 18, 2004 at 07:28:23PM +0200, Leonard den Ottolander wrote: > Hi, > > On Wed, 2004-08-18 at 16:31, Leonard den Ottolander wrote: > > I noticed this in Red Hat's bugzilla just now: > > http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=127973 . > > Attached is a patch that escapes all dangerous characters for function > arguments. More specifically, everything not in A-Z, a-z, 0-9, _, /, ., > - and +. > > Could somebody on a system with dpkg installed verify that things still > work correctly after applying this patch? Thanks. Also, isn't mcdebfs_copyout's destfile not used just in system () (where it should be escaped), but also in if ( open(FILEOUT,">$destfile") ) (where I'd say it should not be escaped)? Jakub From leonard at den.ottolander.nl Wed Aug 18 21:44:53 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 18 Aug 2004 23:44:53 +0200 Subject: rpm vfs [was: Re: Re[6]: security fix for temp files in cvs now] In-Reply-To: <1929795893.20040819032436@mail.ru> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <1092781777.6246.2.camel@athlon.localdomain> <1092782986.6246.16.camel@athlon.localdomain> <613824307.20040818051505@mail.ru> <1092842598.4777.76.camel@athlon.localdomain> <1929795893.20040819032436@mail.ru> Message-ID: <1092865493.4777.232.camel@athlon.localdomain> Hello Pavel, On Wed, 2004-08-18 at 23:24, pavelsh wrote: > Wednesday, August 18, 2004, 9:23:18 PM, you wrote: > LdO> On Wed, 2004-08-18 at 01:15, pavelsh wrote: > >> Please, make single patch with all diff. And i'll review it. > >> Ok? > > LdO> Added the extfs_rpm.patch to the > LdO> proposed branch. It adds INFO/OBSOLETES and INFO/LICENSE, but leaves > LdO> INFO/COPYRIGHT alone (should probably be kept around while rpm supports > LdO> the tag). > Commited. Ok. Thanks. Do you have bugzilla rights? If so, please close https://savannah.gnu.org/bugs/?func=detailitem&item_id=4271 as resolved. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 01:56:06 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 03:56:06 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092850103.4777.83.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> Message-ID: <1092880566.4777.252.camel@athlon.localdomain> Hi, On Wed, 2004-08-18 at 19:28, Leonard den Ottolander wrote: > > http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=127973 . > > Attached is a patch that escapes all dangerous characters for function > arguments. More specifically, everything not in A-Z, a-z, 0-9, _, /, ., > - and +. The attached file is a much nicer solution. I believe the original escaping in mcdebfs_run is redundant as the involved parameters are passed to mcdebfs_copyout where they will be escaped. Is my use of map correct? Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deb.in.diff Type: text/x-patch Size: 864 bytes Desc: not available URL: From leonard at den.ottolander.nl Thu Aug 19 07:48:53 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 09:48:53 +0200 Subject: extfs: rpm vs. trpm and rpms.ini Message-ID: <1092901732.4751.12.camel@athlon.localdomain> Hi, I would like to know if the changes that have been made to rpm should also be applied to trpm. The only difference between the script seems to be the fact that rpm queries uninstalled packages (-qp) whereas trpm queries installed packages (-q). I would assume the "whitespace" fix is also relevant to trpm. Why not make these two files identical copies (apart from the query format)? Or are there other issues for which the files should differ? Regarding rpms.ini: What's the use of the "print @trpms;" on line 31? That seems to be an uninitialized variable. Also the comments on lines 56-62 seem totally irrelevant. Can it be removed? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From andrew at email.zp.ua Thu Aug 19 07:51:48 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Thu, 19 Aug 2004 10:51:48 +0300 (EEST) Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092862471.4777.220.camel@athlon.localdomain> Message-ID: <200408190751.i7J7pmCx054779@email.zp.ua> Hi Leonard, > On Wed, 2004-08-18 at 21:35, Andrew V. Samoilov wrote: > > patchfs and uzip is ok ;-) > > I see. copyin is passed unchecked parameters, but those are quotemeta'd > with myin. This seems to be the case in most opens, except one: copyout. > Are you sure 'open 0, "> $out";' is fine? Well `open O, '>', $out` is more right and secure here and in apt.in, rpms.in . Patch attached. Can you commit this one? vfs/ChangeLog: patchfs -- Regards, Andrew V. Samoilov. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: patchfs.in.patch URL: From leonard at den.ottolander.nl Thu Aug 19 07:54:52 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 09:54:52 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <200408190742.i7J7gJKU054464@email.zp.ua> References: <200408190742.i7J7gJKU054464@email.zp.ua> Message-ID: <1092902092.4751.18.camel@athlon.localdomain> Hi Andrew, On Thu, 2004-08-19 at 09:42, Andrew V. Samoilov wrote: > > I see. copyin is passed unchecked parameters, but those are quotemeta'd > > with myin. This seems to be the case in most opens, except one: copyout. > > Are you sure 'open 0, "> $out";' is fine? > > Well `open O, '>', $out` is more right and secure here. > Patch attached. Can you commit this one? Please wait with committing this. As I am going through many of the files in vfs/extfs it's probably better to wait for a comprehensive patch that I intend to make. There are many more occurrences of the above open syntax. Why is the latter form more correct? If it is I am happy to change all occurrences of the old form where I find them. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 08:03:26 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 10:03:26 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092902092.4751.18.camel@athlon.localdomain> References: <200408190742.i7J7gJKU054464@email.zp.ua> <1092902092.4751.18.camel@athlon.localdomain> Message-ID: <1092902606.4751.23.camel@athlon.localdomain> Hi, On Thu, 2004-08-19 at 09:54, Leonard den Ottolander wrote: > Please wait with committing this. As I am going through many of the > files in vfs/extfs it's probably better to wait for a comprehensive > patch that I intend to make. There are many more occurrences of the > above open syntax. Never mind the above. As I had no intention on touching that file anyway (it seems correct wrt escaping parameters) the changes could just as well be committed. Currently looking at rpm/trpm/rpms.in and apt.in/deba.in/debd.in/dpkg.in. Seems to be a lot of redundancy, especially in the latter 4. By the way, apart from patchfs and uzip also mailfs seems to be correct wrt escaping parameters. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From andrew at email.zp.ua Thu Aug 19 08:14:29 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Thu, 19 Aug 2004 11:14:29 +0300 (EEST) Subject: CVS commits and ChangeLog In-Reply-To: <411FF481.3010603@gmx.de> Message-ID: <200408190814.i7J8ETWt055373@email.zp.ua> > Hi, > > I have seen that you there is much redundancy in the CVS log. You mostly > added your commit message to ChangeLog and commited it, too. I don't > quite understand why this is better than just letting cvs2cl run over > the CVS repository once for every release. > > To make it short: Why do you write your log message thrice, once in the > ChangeLog file, once in the ChangeLog log and once in the real-file log? > > Do you want me to do the same? Yes, life is much more fine with ChangeLog entries ;-) -- Regards, Andrew V. Samoilov. From nadvornik at suse.cz Thu Aug 19 08:20:23 2004 From: nadvornik at suse.cz (Vladimir Nadvornik) Date: Thu, 19 Aug 2004 10:20:23 +0200 Subject: extfs: rpm vs. trpm and rpms.ini In-Reply-To: <1092901732.4751.12.camel@athlon.localdomain> References: <1092901732.4751.12.camel@athlon.localdomain> Message-ID: <200408191020.23747.nadvornik@suse.cz> On Thursday 19 August 2004 09:48, Leonard den Ottolander wrote: > Hi, > > I would like to know if the changes that have been made to rpm should > also be applied to trpm. The only difference between the script seems to > be the fact that rpm queries uninstalled packages (-qp) whereas trpm > queries installed packages (-q). > There is also a difference in copyout: trpm takes the file directly from filesystem and there is no cpio handling. > I would assume the "whitespace" fix is also relevant to trpm. Why not > make these two files identical copies (apart from the query format)? Or > are there other issues for which the files should differ? > I propose merge the files, make trpm symlink to rpm and switch the behavior according to $0 -- Vladimir Nadvornik developer --------------------------------------------------------------------- SuSE CR, s.r.o. e-mail: nadvornik at suse.cz Drahobejlova 27 tel:+420 2 9654 2373 190 00 Praha 9 fax:+420 2 9654 2374 Ceska republika http://www.suse.cz From leonard at den.ottolander.nl Thu Aug 19 08:54:41 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 10:54:41 +0200 Subject: extfs: rpm vs. trpm and rpms.ini In-Reply-To: <200408191020.23747.nadvornik@suse.cz> References: <1092901732.4751.12.camel@athlon.localdomain> <200408191020.23747.nadvornik@suse.cz> Message-ID: <1092905681.4751.30.camel@athlon.localdomain> On Thu, 2004-08-19 at 10:20, Vladimir Nadvornik wrote: > There is also a difference in copyout: trpm takes the file directly from > filesystem and there is no cpio handling. Yeah, I noticed. > I propose merge the files, make trpm symlink to rpm and switch > the behavior according to $0 Thought of something like that. Also a parameter $RPMQ could be added representing either "rpm -q" or "rpm -qp" to avoid redundancy. Before attempting this I first want to propose the attached changes. extfs_rpm2.diff holds another small fix from rpm, extfs_trpm.diff is a synchronization of trpm with rpm. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: extfs_rpm2.diff Type: text/x-patch Size: 460 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: extfs_trpm.diff Type: text/x-patch Size: 9626 bytes Desc: not available URL: From leonard at den.ottolander.nl Thu Aug 19 09:10:02 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 11:10:02 +0200 Subject: deba.in.diff Message-ID: <1092906602.4751.35.camel@athlon.localdomain> Hi, Since ls() is not used in deba.in and ls() is the only function calling bt(), ft() and fm() all four functions are redundant here. Also 3 occurrences of shell invocations that I escape using map(s/([^\w\/.+-])/\\$1/g, @_). Note that this escapes all parameters, although there is only one vulnerable in each function. Instead of the used regular expression "quotemeta" could be used instead. Another possibility is only to escape the vulnerable parameter. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deba.in.diff Type: text/x-patch Size: 3122 bytes Desc: not available URL: From roland.illig at gmx.de Thu Aug 19 09:18:44 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 11:18:44 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092880566.4777.252.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <1092880566.4777.252.camel@athlon.localdomain> Message-ID: <41247074.5030602@gmx.de> Leonard den Ottolander wrote: > The attached file is a much nicer solution. I believe the original > escaping in mcdebfs_run is redundant as the involved parameters are > passed to mcdebfs_copyout where they will be escaped. Is my use of map > correct? > > + local($archivename) = map(s/([^\w\/.+-])/\\$1/g, @_); I would rather write my ($archivename) = map { s/([^\w\/.+-])/\\$1/g } @_; map(expr, list) yields the result of the s/// operator, whereas map(block, list) executes the block with $_ aliased to the list item. Roland From andrew at email.zp.ua Thu Aug 19 09:37:39 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Thu, 19 Aug 2004 12:37:39 +0300 (EEST) Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092902092.4751.18.camel@athlon.localdomain> Message-ID: <200408190937.i7J9bddg057831@email.zp.ua> > Hi Andrew, > > On Thu, 2004-08-19 at 09:42, Andrew V. Samoilov wrote: > > > I see. copyin is passed unchecked parameters, but those are quotemeta'd > > > with myin. This seems to be the case in most opens, except one: copyout. > > > Are you sure 'open 0, "> $out";' is fine? > > > > Well `open O, '>', $out` is more right and secure here. > > Patch attached. Can you commit this one? > > Please wait with committing this. As I am going through many of the > files in vfs/extfs it's probably better to wait for a comprehensive > patch that I intend to make. There are many more occurrences of the > above open syntax. > > Why is the latter form more correct? If it is I am happy to change all > occurrences of the old form where I find them. man perlfunc Use 3-argument form to open a file with arbitrary weird characters in it, Unfortunatelly 3-argument form is not portable. Solaris8 is shiped with perl 5.005 and this perl does not understand such open calls ;-( It seems there is no way to create file with trailing spaces by open now. We can use sysopen() for such files but this function is also new for perl5 and so is not portable too. I have no access to cvs now, so commit second part of pathchfs.in path, please. -- Regards, Andrew V. Samoilov. From andrew at email.zp.ua Thu Aug 19 10:08:26 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Thu, 19 Aug 2004 13:08:26 +0300 (EEST) Subject: [smbfs.c]: Fix recently introduced memory leak Message-ID: <200408191008.i7JA8Qvt058827@email.zp.ua> Hi Roland! This patch should be applied some years ago, but I didn't abuse by huge patches... -- Regards, Andrew V. Samoilov. -------------- next part -------------- A non-text attachment was scrubbed... Name: smbfs.c.patch.gz Type: application/x-gzip Size: 1604 bytes Desc: not available URL: From roland.illig at gmx.de Thu Aug 19 10:25:31 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 12:25:31 +0200 Subject: C Syntax file Message-ID: <4124801B.5050204@gmx.de> Hi, may I mark "evil C functions" like gets in a wonderful color combination of brightred on brightgreen? It really looks good. ;) in c.syntax: keyword whole gets brightred brightgreen Roland From jakub at redhat.com Thu Aug 19 08:30:57 2004 From: jakub at redhat.com (Jakub Jelinek) Date: Thu, 19 Aug 2004 10:30:57 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092857607.4777.125.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <20040818164036.GT30497@sunsite.ms.mff.cuni.cz> <1092857607.4777.125.camel@athlon.localdomain> Message-ID: <20040819083057.GV30497@sunsite.ms.mff.cuni.cz> On Wed, Aug 18, 2004 at 09:33:28PM +0200, Leonard den Ottolander wrote: > On Wed, 2004-08-18 at 18:40, Jakub Jelinek wrote: > > Also, isn't mcdebfs_copyout's destfile not used just in system () > > (where it should be escaped), but also in > > if ( open(FILEOUT,">$destfile") ) > > (where I'd say it should not be escaped)? > > Why not there? I don't even pretend I know perl, but given: foo1: #!/usr/bin/perl local($destfile)=$ARGV[0]; if ( open(FILEOUT,">$destfile") ) { print FILEOUT "bar"; close FILEOUT; } system("echo foo >> $destfile"); foo2: !/usr/bin/perl local($destfile)=map(s/([^\w\/.+-])/\\$1/g, $ARGV[0]); if ( open(FILEOUT,">$destfile") ) { print FILEOUT "bar"; close FILEOUT; } system("echo foo >> $destfile"); foo3: #!/usr/bin/perl local($destfile)=quotemeta($ARGV[0]); if ( open(FILEOUT,">$destfile") ) { print FILEOUT "bar"; close FILEOUT; } system("echo foo >> $destfile"); I get two files created by the first and third when running ./fooN 'x`date`y' For foo1: xThu Aug 19 12:44:37 CEST 2004y (containing foo) x`date`y (containing bar) For foo2: 2 (containing barfoo) For foo3: x`date`y (containing foo) x\`date\`y (containing bar) Jakub From roland.illig at gmx.de Thu Aug 19 12:07:16 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 14:07:16 +0200 Subject: [smbfs.c]: Fix recently introduced memory leak In-Reply-To: <200408191008.i7JA8Qvt058827@email.zp.ua> References: <200408191008.i7JA8Qvt058827@email.zp.ua> Message-ID: <412497F4.8030701@gmx.de> Andrew V. Samoilov wrote: > Hi Roland! > > This patch should be applied some years ago, but I didn't abuse by huge patches... > Hello Andrew, I have applied a slight variant of your patch. Your patch didn't fix a memory leak (to my knowledge), but I discovered one while checking and applying your patch. But most of the patch has been committed. Roland From andrew at email.zp.ua Thu Aug 19 13:42:03 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Thu, 19 Aug 2004 16:42:03 +0300 (EEST) Subject: [smbfs.c]: Fix recently introduced memory leak In-Reply-To: <412497F4.8030701@gmx.de> Message-ID: <200408191342.i7JDg35H066194@email.zp.ua> Hi, Roland! > I have applied a slight variant of your patch. Your patch didn't fix a > memory leak (to my knowledge), but I discovered one while checking and What about previous smbfs_get_remote_stat(), smbfs_mkdir() and smbfs_rmdir()? > applying your patch. But most of the patch has been committed. Thanks. BTW, these lines are from HACKING: Please use the same indentation as other developers. To indent a block, select in the internal editor and use Shift-F9 to call the external indent. For historic reasons, GNU Midnight Commander used formatting that is not default for GNU Indent. Please put following text to your ~/.indent.pro file to make GNU Indent follow the style used in GNU Midnight Commander: -kr -i4 -pcs -psl --ignore-newlines It's OK to indent the whole function if you edit it. However, please refrain from it if you are posting your patch for review. In this case you would save time of other developers if you only include significant changes. The developer applying your patch can format the code for you. P.S. Please use () after function name in ChangeLog(s) entries and begin this one with * filename (function): Description . -- Regards, Andrew V. Samoilov. From leonard at den.ottolander.nl Thu Aug 19 14:38:24 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 16:38:24 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <200408190937.i7J9bddg057831@email.zp.ua> References: <200408190937.i7J9bddg057831@email.zp.ua> Message-ID: <1092926303.4751.11.camel@athlon.localdomain> On Thu, 2004-08-19 at 11:37, Andrew V. Samoilov wrote: > Unfortunatelly 3-argument form is not portable. > Solaris8 is shiped with perl 5.005 and this perl does not understand such open calls ;-( It seems there is no way to create file with trailing spaces by open now. We can use sysopen() for such files but this function is also new for perl5 and so is not portable too. So we shouldn't use that then... > I have no access to cvs now, so commit second part of pathchfs.in path, please. I agree with the elsifs, but I do not have CVS commit access. Pavel Shirshov or Roland Iilig should be reading it - ah, the latter already committed. Not the "'>', $file" fix I hope. If so, please revert. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 14:46:12 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 16:46:12 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <41247074.5030602@gmx.de> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <1092880566.4777.252.camel@athlon.localdomain> <41247074.5030602@gmx.de> Message-ID: <1092926772.4751.17.camel@athlon.localdomain> Hi Roland, On Thu, 2004-08-19 at 11:18, Roland Illig wrote: > > + local($archivename) = map(s/([^\w\/.+-])/\\$1/g, @_); > > I would rather write > > my ($archivename) = map { s/([^\w\/.+-])/\\$1/g } @_; That would render the attached deb.in.diff. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deb.in.diff Type: text/x-patch Size: 868 bytes Desc: not available URL: From leonard at den.ottolander.nl Thu Aug 19 14:53:47 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 16:53:47 +0200 Subject: deba.in.diff In-Reply-To: <1092906602.4751.35.camel@athlon.localdomain> References: <1092906602.4751.35.camel@athlon.localdomain> Message-ID: <1092927227.4751.24.camel@athlon.localdomain> On Thu, 2004-08-19 at 11:10, Leonard den Ottolander wrote: > Also 3 occurrences of shell invocations that I escape using > map(s/([^\w\/.+-])/\\$1/g, @_). Changed according to suggestion from Roland. Also reverted the open '>', $destfile to "> $destfile". I see I left a quote in the original patch :( . Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deba.in.diff.gz Type: application/x-gzip Size: 819 bytes Desc: not available URL: From roland.illig at gmx.de Thu Aug 19 15:27:47 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 17:27:47 +0200 Subject: [smbfs.c]: Fix recently introduced memory leak In-Reply-To: <200408191342.i7JDg35H066194@email.zp.ua> References: <200408191342.i7JDg35H066194@email.zp.ua> Message-ID: <4124C6F3.3000404@gmx.de> Andrew V. Samoilov wrote: > Hi, Roland! > > >>I have applied a slight variant of your patch. Your patch didn't fix a >>memory leak (to my knowledge), but I discovered one while checking and > > > What about previous smbfs_get_remote_stat(), smbfs_mkdir() and smbfs_rmdir()? smbfs_get_remote_stat(): char *mypath = g_strdup(path); ... smbfs_convert_path (&mypath, FALSE); /* in-place editing */ ... if (condition) { ... free (mypath); return; } else { ... free (mypath); return; } That's the part that matters to the mypath variable. I can't see any memory leak here. Ah, well, it's the smbfs_convert_path() function that had leaked. Sorry, sorry, sorry. Now I see it. > BTW, these lines are from HACKING: > [...] > changes. The developer applying your patch can format the code for you. Sorry, I had forgotten to remember this line while committing. I'll make up for it. > P.S. Please use () after function name in ChangeLog(s) entries and begin this one with > * filename (function): Description . I'll do. Roland From roland.illig at gmx.de Thu Aug 19 15:35:32 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 17:35:32 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092857250.4777.123.camel@athlon.localdomain> References: <1092839508.4777.57.camel@athlon.localdomain> <1092850103.4777.83.camel@athlon.localdomain> <20040818152209.GS30497@sunsite.ms.mff.cuni.cz> <1092857250.4777.123.camel@athlon.localdomain> Message-ID: <4124C8C4.1020808@gmx.de> Leonard den Ottolander wrote: > Hi Jakub, > > On Wed, 2004-08-18 at 17:22, Jakub Jelinek wrote: > >>There are many other scripts which need similar treatment. >>grep -l bin/perl /usr/share/mc/extfs/* | xargs grep open >>shows a lot of (potential) problems, at least in a, apt, debd, mailfs, >>patchfs, rpms and uzip. > > > a looks vulnerable. This script uses $disk:/$path. Hm. Very dozy. What > files is it used for anyway? For files on floppies. The mtools provide the good old DOS commands (dir, copy, move) for UNIX as (mdir, mcopy, mmove). > apt uses the result of a find. Probably vulnerable. Also uses the output > of an "apt-cache dumpavail". Maybe somebody could enlighten me on this > command, but I think it could use escaping anyway. And an unchecked > $file. Bad script! Bad! apt-cache dumpavail does: for i in ${all-available-packages}; do print_complete_package_info $i done The first package info looks like: Package: 3dchess Priority: optional Section: games Installed-Size: 152 Maintainer: Stephen Stafford Architecture: i386 Version: 0.8.1-11 Depends: libc6 (>= 2.3.2.ds1-4), xaw3dg (>= 1.5+E-1), xlibs (>> 4.1.0) Filename: pool/main/3/3dchess/3dchess_0.8.1-11_i386.deb Size: 33116 MD5sum: 7248665d99d529342a5cd050a9128ff6 Description: 3D chess for X11 3 dimensional Chess game for X11R6. There are three boards, stacked vertically; 96 pieces of which most are the traditional chess pieces with just a couple of additions; 26 possible directions in which to move. The AI isn't wonderful, but provides a challenging enough game to all but the most highly skilled players. Roland From roland.illig at gmx.de Thu Aug 19 15:43:24 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 17:43:24 +0200 Subject: deb vfs security issue (CAN-2004-0494) In-Reply-To: <1092926303.4751.11.camel@athlon.localdomain> References: <200408190937.i7J9bddg057831@email.zp.ua> <1092926303.4751.11.camel@athlon.localdomain> Message-ID: <4124CA9C.9070908@gmx.de> Leonard den Ottolander wrote: >>I have no access to cvs now, so commit second part of pathchfs.in path, please. Done. Roland From roland.illig at gmx.de Thu Aug 19 15:53:54 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 17:53:54 +0200 Subject: deba.in.diff In-Reply-To: <1092927227.4751.24.camel@athlon.localdomain> References: <1092906602.4751.35.camel@athlon.localdomain> <1092927227.4751.24.camel@athlon.localdomain> Message-ID: <4124CD12.6020601@gmx.de> Leonard den Ottolander wrote: > Changed according to suggestion from Roland. > > Also reverted the open '>', $destfile to "> $destfile". I see I left a > quote in the original patch :( . > > Leonard. Committed. Roland From leonard at den.ottolander.nl Thu Aug 19 16:08:45 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 18:08:45 +0200 Subject: deba.in.diff In-Reply-To: <4124CD12.6020601@gmx.de> References: <1092906602.4751.35.camel@athlon.localdomain> <1092927227.4751.24.camel@athlon.localdomain> <4124CD12.6020601@gmx.de> Message-ID: <1092931724.4751.50.camel@athlon.localdomain> Hi Roland, On Thu, 2004-08-19 at 17:53, Roland Illig wrote: > Leonard den Ottolander wrote: > > Changed according to suggestion from Roland. > > > > Also reverted the open '>', $destfile to "> $destfile". I see I left a > > quote in the original patch :( . > > > > Leonard. > > Committed. Hm. This code was more in "still needs discussion" phase. The removal of the redundant code is fine, but I seriously start to doubt if the use of map is correct... I have rewritten deb.in.diff to create quoted equivalents of the variables where necessary, and use the plain ones if not. I am afraid the same should be done for deba.in.diff. Sorry if I am wasting your time. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From makovick at kmlinux.fjfi.cvut.cz Thu Aug 19 16:14:19 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Thu, 19 Aug 2004 18:14:19 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change Message-ID: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> Hello, this bug is very easily reproducible. Just open any text file, move one line down and press PageUp. The viewer locks up because of unsigned int rollover. The attached patch should fix it. Regards, -- Jindrich Makovicka From leonard at den.ottolander.nl Thu Aug 19 16:19:35 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 18:19:35 +0200 Subject: deb.in.diff, another try Message-ID: <1092932374.4751.63.camel@athlon.localdomain> Hi, Reading Jakub's analysis I seriously started to doubt my use of map, and the fact that I quote where it is not necessary. I thought the double quoting was harmless, but it appears not to be. In this version I use both escaped and none escaped parameters as necessary. I hope I got it right this time... Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deb.in.diff Type: text/x-patch Size: 2652 bytes Desc: not available URL: From makovick at kmlinux.fjfi.cvut.cz Thu Aug 19 16:24:40 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Thu, 19 Aug 2004 18:24:40 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> Message-ID: <4124D448.3050501@kmlinux.fjfi.cvut.cz> Oops... I shouldn't use alpha version of Mozilla... here it goes. --- view.c.orig 2004-08-17 17:51:30.000000000 +0200 +++ view.c 2004-08-19 18:10:55.000000000 +0200 @@ -1296,7 +1296,7 @@ line = 1; else line = 0; - for (q = p = current - 1; p >= view->first; p--) + for (q = p = current - 1;; p--) { if (get_byte (view, p) == '\n' || p == view->first) { pm = p > view->first ? p + 1 : view->first; if (!view->wrap_mode) { @@ -1315,6 +1315,9 @@ q = p + 1; } } + if (p <= view->first) break; + } + } return p > view->first ? p : view->first; } -- Jindrich Makovicka From leonard at den.ottolander.nl Thu Aug 19 16:34:16 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 18:34:16 +0200 Subject: And deb.in.diff another try Message-ID: <1092933255.4751.66.camel@athlon.localdomain> Hi, This is against current CVS. Removed the maps and use q-parameters where escaping is needed. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: deba.in.diff Type: text/x-patch Size: 2909 bytes Desc: not available URL: From roland.illig at gmx.de Thu Aug 19 16:44:44 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 18:44:44 +0200 Subject: shell quoting unified Message-ID: <4124D8FC.20205@gmx.de> Hi, I've written a Perl function shell_quote that works fine with my current Perl and a Perl-5.005_003 that I digged out of the depth of the internet. sub shell_quote { my (@result); foreach my $i (@_) { my $tmp = $i; $tmp =~ s/([^\w\/.+-])/\\$1/g; push(@result, $tmp); } return wantarray ? @result : $result[0]; } You can use it like this: sub foofs_copyin { my ($archive, $filename) = shell_quote(@_); $filename = shell_quote($filename); # double-quoted } The advantage over the current version is the name. It is easier readable and it tells the reader our intention. I'd like it to replace the current regexps. (Sorry, I'm studying computer science with the special topic "Software Development Methodology", so I'm kind of strict concerning these things.) Roland From roland.illig at gmx.de Thu Aug 19 16:45:56 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 18:45:56 +0200 Subject: deba.in.diff In-Reply-To: <1092931724.4751.50.camel@athlon.localdomain> References: <1092906602.4751.35.camel@athlon.localdomain> <1092927227.4751.24.camel@athlon.localdomain> <4124CD12.6020601@gmx.de> <1092931724.4751.50.camel@athlon.localdomain> Message-ID: <4124D944.8020102@gmx.de> Leonard den Ottolander wrote: > I have rewritten deb.in.diff to create quoted equivalents of the > variables where necessary, and use the plain ones if not. I am afraid > the same should be done for deba.in.diff. Sorry if I am wasting your > time. You aren't. I have just done the same. ;) (See the new thread.) Roland From leonard at den.ottolander.nl Thu Aug 19 16:49:55 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 18:49:55 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124D448.3050501@kmlinux.fjfi.cvut.cz> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> Message-ID: <1092934194.4751.72.camel@athlon.localdomain> Hi Jindrich, On Thu, 2004-08-19 at 18:24, Jindrich Makovicka wrote: > --- view.c.orig 2004-08-17 17:51:30.000000000 +0200 > +++ view.c 2004-08-19 18:10:55.000000000 +0200 > @@ -1296,7 +1296,7 @@ > line = 1; > else > line = 0; > - for (q = p = current - 1; p >= view->first; p--) > + for (q = p = current - 1;; p--) { > if (get_byte (view, p) == '\n' || p == view->first) { > pm = p > view->first ? p + 1 : view->first; > if (!view->wrap_mode) { > @@ -1315,6 +1315,9 @@ > q = p + 1; > } > } > + if (p <= view->first) break; > + } > + > } > return p > view->first ? p : view->first; > } I am not able to reproduce this, but I would say the last hunk is unnecessary. No need to have a break statement at the end of the loop when the for loop can do it's own comparision. All this seems to indicate is: - for (q = p = current - 1; p >= view->first; p--) + for (q = p = current - 1; p > view->first; p--) Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 16:53:02 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 18:53:02 +0200 Subject: shell quoting unified In-Reply-To: <4124D8FC.20205@gmx.de> References: <4124D8FC.20205@gmx.de> Message-ID: <1092934382.4751.76.camel@athlon.localdomain> Hi Roland, On Thu, 2004-08-19 at 18:44, Roland Illig wrote: > sub foofs_copyin > { > my ($archive, $filename) = shell_quote(@_); > $filename = shell_quote($filename); # double-quoted > } Please do not use this. One of the issues I did not address before is the fact that in the case of open(FILEOUT, "> $destfile") the variable should *not* be quoted. Hence my last patch to create two variables where necessary: One plain and one quoted. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 17:00:32 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 19:00:32 +0200 Subject: shell quoting unified In-Reply-To: <1092934382.4751.76.camel@athlon.localdomain> References: <4124D8FC.20205@gmx.de> <1092934382.4751.76.camel@athlon.localdomain> Message-ID: <1092934832.4751.81.camel@athlon.localdomain> Hi, On Thu, 2004-08-19 at 18:53, Leonard den Ottolander wrote: > On Thu, 2004-08-19 at 18:44, Roland Illig wrote: > > sub foofs_copyin > > { > > my ($archive, $filename) = shell_quote(@_); > > $filename = shell_quote($filename); # double-quoted > > } > > Please do not use this. One of the issues I did not address before is > the fact that in the case of open(FILEOUT, "> $destfile") the variable > should *not* be quoted. What of course I meant to say is don't use the my ($archive, $filename) = shell_quote(@_); construct where it is not appropriate. You could of course use it where all parameters need to be quoted. And the second form is alright. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Thu Aug 19 17:07:35 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 19:07:35 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124D448.3050501@kmlinux.fjfi.cvut.cz> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> Message-ID: <4124DE57.2040903@gmx.de> Jindrich Makovicka wrote: > Oops... I shouldn't use alpha version of Mozilla... here it goes. > > --- view.c.orig 2004-08-17 17:51:30.000000000 +0200 > +++ view.c 2004-08-19 18:10:55.000000000 +0200 > @@ -1296,7 +1296,7 @@ > line = 1; > else > line = 0; > - for (q = p = current - 1; p >= view->first; p--) > + for (q = p = current - 1;; p--) { > if (get_byte (view, p) == '\n' || p == view->first) { > pm = p > view->first ? p + 1 : view->first; > if (!view->wrap_mode) { > @@ -1315,6 +1315,9 @@ > q = p + 1; > } > } > + if (p <= view->first) break; > + } > + > } > return p > view->first ? p : view->first; > } > > Sorry that I intruduced this bug. I'm afraid fixing this bug will not be that easy. If you apply your patch and create a file whose first lines are empty, you will jump from line 3 upto line 1 when continuously pressing . I'll work on this. Roland From roland.illig at gmx.de Thu Aug 19 17:13:48 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 19:13:48 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124DE57.2040903@gmx.de> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> <4124DE57.2040903@gmx.de> Message-ID: <4124DFCC.8010604@gmx.de> Roland Illig wrote: > Jindrich Makovicka wrote: > >> Oops... I shouldn't use alpha version of Mozilla... here it goes. >> >> --- view.c.orig 2004-08-17 17:51:30.000000000 +0200 >> +++ view.c 2004-08-19 18:10:55.000000000 +0200 >> @@ -1296,7 +1296,7 @@ >> line = 1; >> else >> line = 0; >> - for (q = p = current - 1; p >= view->first; p--) >> + for (q = p = current - 1;; p--) { >> if (get_byte (view, p) == '\n' || p == view->first) { >> pm = p > view->first ? p + 1 : view->first; >> if (!view->wrap_mode) { >> @@ -1315,6 +1315,9 @@ >> q = p + 1; >> } >> } >> + if (p <= view->first) break; >> + } >> + } >> return p > view->first ? p : view->first; >> } >> >> > > Sorry that I intruduced this bug. I'm afraid fixing this bug will not be > that easy. If you apply your patch and create a file whose first lines > are empty, you will jump from line 3 upto line 1 when continuously > pressing . Sorry for not checking thoroughly. And sorry for always saying sorry. ;) I just discovered that the "two-lines-up" bug already existed in mc-4.6.0. I'll nevertheless try to fix it. And I will commit the change as suggested by Leonard, as it looks simpler and almost does the same. Roland From makovick at kmlinux.fjfi.cvut.cz Thu Aug 19 17:19:34 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Thu, 19 Aug 2004 19:19:34 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124DFCC.8010604@gmx.de> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> <4124DE57.2040903@gmx.de> <4124DFCC.8010604@gmx.de> Message-ID: <4124E126.8010307@kmlinux.fjfi.cvut.cz> Roland Illig wrote: > I just discovered that the "two-lines-up" bug already existed in > mc-4.6.0. I'll nevertheless try to fix it. And I will commit the change > as suggested by Leonard, as it looks simpler and almost does the same. Before you commit it, there is a similar bug in the hex viewer branch :) @@ -1276,8 +1346,10 @@ return current; if (view->hex_mode) { - p = current - lines * view->bytes_per_line; - p = (p < view->first) ? view->first : p; + if (current < lines * view->bytes_per_line + view->first) + p = view->first; + else + p = current - lines * view->bytes_per_line; if (lines == 1) { q = view->edit_cursor - view->bytes_per_line; view->edit_cursor = (q < view->first) ? view->edit_cursor : q; -- Jindrich Makovicka From roland.illig at gmx.de Thu Aug 19 17:32:43 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 19:32:43 +0200 Subject: And deb.in.diff another try In-Reply-To: <1092933255.4751.66.camel@athlon.localdomain> References: <1092933255.4751.66.camel@athlon.localdomain> Message-ID: <4124E43B.60207@gmx.de> Leonard den Ottolander wrote: > Hi, > > This is against current CVS. Removed the maps and use q-parameters where > escaping is needed. Looks nice. Committed. Roland From roland.illig at gmx.de Thu Aug 19 17:38:11 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 19 Aug 2004 19:38:11 +0200 Subject: deb.in.diff, another try In-Reply-To: <1092932374.4751.63.camel@athlon.localdomain> References: <1092932374.4751.63.camel@athlon.localdomain> Message-ID: <4124E583.9040707@gmx.de> Leonard den Ottolander wrote: > Hi, > > Reading Jakub's analysis I seriously started to doubt my use of map, and > the fact that I quote where it is not necessary. I thought the double > quoting was harmless, but it appears not to be. > > In this version I use both escaped and none escaped parameters as > necessary. I hope I got it right this time... Looks nice, too. Committed. Roland From leonard at den.ottolander.nl Thu Aug 19 17:48:20 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 19:48:20 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124E00F.1040408@kmlinux.fjfi.cvut.cz> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> <1092934194.4751.72.camel@athlon.localdomain> <4124E00F.1040408@kmlinux.fjfi.cvut.cz> Message-ID: <1092937700.4751.87.camel@athlon.localdomain> Hi Jindrich, On Thu, 2004-08-19 at 19:14, Jindrich Makovicka wrote: > Leonard den Ottolander wrote: > > - for (q = p = current - 1; p >= view->first; p--) > > + for (q = p = current - 1; p > view->first; p--) > > This isn't equivalent, your modification misses the pass with > p==view->first. That's why there is the test at the end of the loop, but > before the decrement. I guess you are correct, although it seems odd to use a for loop when having a test at the end. Maybe it should be replaced with a do/while? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 17:51:11 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 19 Aug 2004 19:51:11 +0200 Subject: deb.in.diff, another try In-Reply-To: <4124E583.9040707@gmx.de> References: <1092932374.4751.63.camel@athlon.localdomain> <4124E583.9040707@gmx.de> Message-ID: <1092937871.4751.91.camel@athlon.localdomain> Hi Roland, On Thu, 2004-08-19 at 19:38, Roland Illig wrote: > > In this version I use both escaped and none escaped parameters as > > necessary. I hope I got it right this time... > > Looks nice, too. Committed. Good. Now I can make a definite(?) patch for use with Fedora Core. I will fix the other deb* files accordingly. But not tonight :) . Leonard. -- mount -t life -o ro /dev/dna /genetic/research From andrew at email.zp.ua Thu Aug 19 18:47:25 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Thu, 19 Aug 2004 21:47:25 +0300 (EEST) Subject: smbfs_open(): Implement O_APPEND via smbfs_lseek() Message-ID: <200408191847.i7JIlPpr075434@email.zp.ua> Hello, any complains/side effects? -- Regards, Andrew V. Samoilov. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: smbfs-O_APPEND.patch URL: From pavelsh at mail.ru Thu Aug 19 20:11:53 2004 From: pavelsh at mail.ru (pavelsh) Date: Fri, 20 Aug 2004 02:11:53 +0600 Subject: Please In-Reply-To: <4124E583.9040707@gmx.de> References: <1092932374.4751.63.camel@athlon.localdomain> <4124E583.9040707@gmx.de> Message-ID: <1492944065.20040820021153@mail.ru> Hello Roland, Please, break down your string in ChangeLog file. Long string is being readed terrible. -- Best regards, pavelsh mailto:pavelsh at mail.ru From proski at gnu.org Thu Aug 19 20:47:41 2004 From: proski at gnu.org (Pavel Roskin) Date: Thu, 19 Aug 2004 16:47:41 -0400 (EDT) Subject: Maintenance question Message-ID: Hello! This e-mail is long overdue, but I hoped things would change. I cannot be a single maintainer of GNU Midnight Commander anymore. Not only my time is limited due to my job, but I'm also emotionally exhausted because of my personal life (I'm a "NonBP" now, use Google for details). Also, I don't find hacking userspace software as exciting and it was for me in the past, although the experience was very important for me. I set high standards for what I expect to be in the future versions of GNU Midnight Commander, but it looks like I'll not be able to meet those standards in a reasonable time. If somebody is seriously considering sharing the burden of maintaining GNU Midnight Commander, please write me and I'll assist you in getting access to the website, mailing lists and administrative privileges on savannah.gnu.org. I'd like Miguel de Icaza to participate in choosing the co-maintainers. Miguel wrote this program and he entrusted me with maintaining it. I intend to continue to contribute to the project as much as I can. I think it's primarily the sole responsibility for the project that is weighing on me now. I cannot read 30 e-mails per day knowing (or rather fearing) that I'll have to reply 10 of them. Sorry for everything I did wrong. Sorry for e-mails I didn't have a chance to answer and for patches I didn't have a chance to apply. -- Regards, Pavel Roskin From leonard at den.ottolander.nl Thu Aug 19 23:46:37 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 01:46:37 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <4124D448.3050501@kmlinux.fjfi.cvut.cz> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> Message-ID: <1092959193.4747.27.camel@athlon.localdomain> Hi Jindrich, On Thu, 2004-08-19 at 18:24, Jindrich Makovicka wrote: > - for (q = p = current - 1; p >= view->first; p--) > + for (q = p = current - 1;; p--) { --- view.c.000 2004-08-17 22:57:08.000000000 +0200 +++ view.c 2004-08-20 01:41:23.000000000 +0200 @@ -1296,7 +1296,10 @@ move_backward2 (WView *view, offset_type line = 1; else line = 0; - for (q = p = current - 1; p >= view->first; p--) + q = (p = current) - 1; + if (p > view->first) + do { + p--; if (get_byte (view, p) == '\n' || p == view->first) { pm = p > view->first ? p + 1 : view->first; if (!view->wrap_mode) { @@ -1315,6 +1318,7 @@ move_backward2 (WView *view, offset_type q = p + 1; } } + } while (p > view->first); } return p > view->first ? p : view->first; } Is the above patch functionally identical to what you try to achieve? Your patch assumes the initial test if (p >[=] view->first) is no longer necessary. In that case it can be removed from this patch. The break you use assumes we don't want a decrement of p on exit, so I didn't add it. Correct? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 19 23:59:43 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 01:59:43 +0200 Subject: [patch] viewer lockup due to recent signed->unsigned change In-Reply-To: <1092959193.4747.27.camel@athlon.localdomain> References: <4124D1DB.80809@kmlinux.fjfi.cvut.cz> <4124D448.3050501@kmlinux.fjfi.cvut.cz> <1092959193.4747.27.camel@athlon.localdomain> Message-ID: <1092959983.4747.32.camel@athlon.localdomain> On Fri, 2004-08-20 at 01:46, Leonard den Ottolander wrote: > + if (p > view->first) If you do want the initial "if" you can of course make this a while loop instead. > + do { > + } while (p > view->first); Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Fri Aug 20 07:30:43 2004 From: roland.illig at gmx.de (Roland Illig) Date: Fri, 20 Aug 2004 09:30:43 +0200 Subject: I'm away / viewer patch Message-ID: <4125A8A3.7050806@gmx.de> Hi, I'll be away from the computer for about one month. Don't expect me to do anything in this time. Concerning the patches for the viewer: I had left out the discussion about the patch and have committed one myself. It's completely rewritten and it works for me. The only part that still uses the old code is for text mode with line wrapping enabled. The other modes should be easier understandible than before. Especially I did not use the ?: operator. ;) I had especially focused on only doing non-overflowing arithmetics. Instead of a -= b; if (a_too_small) { correct(a); } I used if (a > b) { a -= b; } else { a = other_value; } Using this style it's more unlikely to get into an endless loop as before. Roland From vlad at lrsehosting.com Thu Aug 19 20:59:14 2004 From: vlad at lrsehosting.com (William Scott Lockwood III) Date: Thu, 19 Aug 2004 15:59:14 -0500 (CDT) Subject: Maintenance question In-Reply-To: References: Message-ID: <11861.65.208.227.246.1092949154.squirrel@www.lrsehosting.com> Best of luck, and thanks for all your hard work! I hope you still stick around in some capacity. -- Regards, Scott ---==== [ Pavel Roskin ] ===--- > Hello! > > This e-mail is long overdue, but I hoped things would change. I cannot be > a single maintainer of GNU Midnight Commander anymore. Not only my time > is limited due to my job, but I'm also emotionally exhausted because of my > personal life (I'm a "NonBP" now, use Google for details). Also, I don't > find hacking userspace software as exciting and it was for me in the past, > although the experience was very important for me. > > I set high standards for what I expect to be in the future versions of GNU > Midnight Commander, but it looks like I'll not be able to meet those > standards in a reasonable time. > > If somebody is seriously considering sharing the burden of maintaining GNU > Midnight Commander, please write me and I'll assist you in getting access > to the website, mailing lists and administrative privileges on > savannah.gnu.org. > > I'd like Miguel de Icaza to participate in choosing the co-maintainers. > Miguel wrote this program and he entrusted me with maintaining it. > > I intend to continue to contribute to the project as much as I can. I > think it's primarily the sole responsibility for the project that is > weighing on me now. I cannot read 30 e-mails per day knowing (or rather > fearing) that I'll have to reply 10 of them. > > Sorry for everything I did wrong. Sorry for e-mails I didn't have a > chance to answer and for patches I didn't have a chance to apply. > > -- > Regards, > Pavel Roskin > _______________________________________________ > Mc mailing list > http://mail.gnome.org/mailman/listinfo/mc > From leonard at den.ottolander.nl Fri Aug 20 14:57:24 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 16:57:24 +0200 Subject: Maintenance question In-Reply-To: References: Message-ID: <1093013843.4746.58.camel@athlon.localdomain> Hello Pavel, On Thu, 2004-08-19 at 22:47, Pavel Roskin wrote: > I cannot be a single maintainer of GNU Midnight Commander anymore. > If somebody is seriously considering sharing the burden of maintaining GNU > Midnight Commander, please write me and I'll assist you in getting access > to the website, mailing lists and administrative privileges on > savannah.gnu.org. Before handing over (some of) the responsibilities for the project maybe you could present us with an overview of the current state of affairs. People might be more willing to help out if they know what is expected of them. F.e: What are the tasks/responsibilities that need to be fulfilled? I can think of mailing list administration, bugzilla administration, patch review, development. With regard to administrative tasks, are there any OS related tasks that need to be done (software installation)? Current access rights: Lately two people have popped up that appear to have CVS commit rights: Pavel Shirshov and Roland Illig. Are there other people with CVS commit rights? What have been the criteria to grant these people commit access? Please give us an overview of all CVS committers and people with bugzilla rights. Code of conduct: Are there any explicit or implicit rules on how people should exercise their rights? F.e., I believe the tasks of coding and committing should be kept somewhat separated. At least people with CVS access should not commit their own patches without them being reviewed by others. Infrastructure: The various midnight commander related pages seem to be scattered around a bit: http://directory.fsf.org/midnightcommander.html http://www.ibiblio.org/mc/ http://savannah.gnu.org/projects/mc mc & mc-devel at gnome.org Are there any more? Current state of development: Which developers are currently working on which parts of the http://www.ibiblio.org/mc/TODO list? Who are the current developers? Please add any omissions in the above overview. > I intend to continue to contribute to the project as much as I can. I > think it's primarily the sole responsibility for the project that is > weighing on me now. That would be nice. Some occasional input on various matters would be much appreciated. As I have let you know before I am willing to share *some* of the above tasks. I wouldn't mind administering the mailing lists (I have some experience with mailman). Also I want to do some work on bugzilla (following up on bug reports, reviewing patches and proposing solutions to developers/committers). Maybe some coding as well. I hope you can spend a little time discussing these matters and informing us of your experiences maintaining the project (suggestions, pitfalls). Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Fri Aug 20 15:36:18 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 17:36:18 +0200 Subject: I'm away / viewer patch In-Reply-To: <4125A8A3.7050806@gmx.de> References: <4125A8A3.7050806@gmx.de> Message-ID: <1093016178.4746.66.camel@athlon.localdomain> Hello Roland, On Fri, 2004-08-20 at 09:30, Roland Illig wrote: > Concerning the patches for the viewer: I had left out the discussion > about the patch and have committed one myself. It's completely rewritten > and it works for me. When you get back we might want to discuss a code of conduct. I don't think it is a good idea that CVS committers commit their own code without it being reviewed by others. If people start committing their own code without having it reviewed things can become pretty messy very soon. Note that this criticism is not meant to be taken personally. It's just that I believe the responsibilities of coding and committing should be somewhat separated. New code should be reviewed by at least two other people. Nothing wrong with coders having commit access, but to avoid arbitrary decisions it would be good to keep the tasks separated, and thus not commit own code (without review). Code might "work for you", but that doesn't mean you've investigated all possible pitfalls. That's why there's peer review. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Fri Aug 20 16:42:04 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 18:42:04 +0200 Subject: debd.in.diff (escaping system/open calls) Message-ID: <1093020123.4746.69.camel@athlon.localdomain> Hi, Ok, debd.in checked for missing escapes on calls to system and opens that spawn a shell. See attached debd.in.diff. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: debd.in.diff Type: text/x-patch Size: 5654 bytes Desc: not available URL: From leonard at den.ottolander.nl Fri Aug 20 16:55:07 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 18:55:07 +0200 Subject: dpkg.in.diff (escaping system/open calls) Message-ID: <1093020907.4746.73.camel@athlon.localdomain> Hi, Ok, dpkg.in checked for missing escapes on calls to system and opens that spawn a shell. See attached dpkg.in.diff. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: dpkg.in.diff Type: text/x-patch Size: 2670 bytes Desc: not available URL: From leonard at den.ottolander.nl Fri Aug 20 17:13:01 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 19:13:01 +0200 Subject: apt.in.diff (escaping system/open calls) Message-ID: <1093021980.4746.87.camel@athlon.localdomain> Hi, Ok, apt.in checked for missing escapes on calls to system and opens that spawn a shell. See attached apt.in.diff. Some questions: line 135: if ( open(PIPEIN, "find /var/cache/apt/archives -type f |") ) { What about found files in that directory? Should the output of find also be escaped? idem line 190: open STAT, "apt-cache dumpavail |" Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: apt.in.diff Type: text/x-patch Size: 3536 bytes Desc: not available URL: From leonard at den.ottolander.nl Fri Aug 20 17:38:43 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 20 Aug 2004 19:38:43 +0200 Subject: a.in.diff (escaping system/open calls) Message-ID: <1093023523.4746.91.camel@athlon.localdomain> Hi, Ok, a.in checked for missing escapes on calls to system and opens that spawn a shell. See attached a.in.diff. One question: What's the "1;" at the end of the file? Can it be removed? Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: a.in.diff Type: text/x-patch Size: 2062 bytes Desc: not available URL: From leonard at den.ottolander.nl Sat Aug 21 02:46:37 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 04:46:37 +0200 Subject: Comprehensive patch escaping system/open calls in vfs/extfs Message-ID: <1093056396.4746.125.camel@athlon.localdomain> Hi, This is a comprehensive patch that escapes parameters to system and open calls that spawn a shell. The patch affects a.in, apt.in, deba.in (parts already committed), debd.in, deb.in (parts committed) and dpkg.in. Afaict parameters in mailfs.in, patchfs.in and uzip.in are properly quoted. It's a compilation of the previous patches with a few fixes. Also replaced the regular expression with a subroutine "quote". Please check for errors and omissions. Leonard. -- mount -t life -o ro /dev/dna /genetic/research -------------- next part -------------- A non-text attachment was scrubbed... Name: extfs-quoted-cvs.patch.gz Type: application/x-gzip Size: 3616 bytes Desc: not available URL: From dmi_a at qnx.org.ru Sat Aug 21 12:21:28 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Sat, 21 Aug 2004 16:21:28 +0400 Subject: Can't open any RPM file Message-ID: <200408211621.28886.dmi_a@qnx.org.ru> Hi, Just updates from CVS and built mc. I now can't browse RPM packages. I tried to restore old /usr/local/share/extfs/rpm* files, but that doesn't solve the problem. mc gives "inconsistent extfs acrhive message" on all rpms. Checked about 20 packages, all of them are okay, some of them are installed. Pressing F3 on RPM files shows it's contents correclty. How can I reach rpm's error messages for debugging this? System: ALT Linux Sisyphus (20040813) Kernel: Linux dmi.home 2.4.26-std-up-alt5 #1 Mon Jun 28 21:59:23 MSD 2004 i686 unknown unknown GNU/Linux Locale: Locale [dmi at dmi mc]$ locale LANG=ru_RU.KOI8-R LC_CTYPE="ru_RU.KOI8-R" LC_NUMERIC="ru_RU.KOI8-R" LC_TIME="ru_RU.KOI8-R" LC_COLLATE="ru_RU.KOI8-R" LC_MONETARY="ru_RU.KOI8-R" LC_MESSAGES="ru_RU.KOI8-R" LC_PAPER="ru_RU.KOI8-R" LC_NAME="ru_RU.KOI8-R" LC_ADDRESS="ru_RU.KOI8-R" LC_TELEPHONE="ru_RU.KOI8-R" LC_MEASUREMENT="ru_RU.KOI8-R" LC_IDENTIFICATION="ru_RU.KOI8-R" LC_ALL= RPM: [dmi at dmi mc]$ rpm -q rpm rpm-4.0.4-alt40 mv CVS checkout from August, 21. WBR Dmitry From dmi_a at qnx.org.ru Sat Aug 21 13:55:48 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Sat, 21 Aug 2004 17:55:48 +0400 Subject: Can't open any RPM file In-Reply-To: <4130346.20040821190701@mail.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <4130346.20040821190701@mail.ru> Message-ID: <200408211755.48005.dmi_a@qnx.org.ru> > DA> Just updates from CVS and built mc. I now can't browse RPM > packages. I DA> tried to restore old /usr/local/share/extfs/rpm* > files, but that DA> doesn't solve the problem. mc gives "inconsistent > extfs acrhive DA> message" on all rpms. Checked about 20 packages, > all of them are okay, DA> some of them are installed. > > Please, send me one package or give me url. > I can't reproduce error with my environment. Pavel, I can't open _any_ package. Just rolled back to an old version (browsing rpm packages is very essintial mc's dfeauture for me), all these packages opens fine. How can I turn obn vfs/extfs debugging? WBR Dmitry From leonard at den.ottolander.nl Sat Aug 21 14:34:14 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 16:34:14 +0200 Subject: Can't open any RPM file In-Reply-To: <200408211621.28886.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> Message-ID: <1093098853.4765.10.camel@athlon.localdomain> Hi Dmitry, On Sat, 2004-08-21 at 14:21, Dmitry Alexeyev wrote: > I now can't browse RPM packages. > mc gives "inconsistent extfs acrhive message" on all rpms. Does it give any more warnings than that? What are the permissions on the file vfs/extfs/rpm? Executable for everyone? I am using that same file and I am experiencing no problems. Might there be issues with the latest modifications to extfs.c? Are you having problems with other archives as well? If so revert that file to before August 14 and see if that changes anything... Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Sat Aug 21 15:10:03 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Sat, 21 Aug 2004 19:10:03 +0400 Subject: Can't open any RPM file In-Reply-To: <1093098853.4765.10.camel@athlon.localdomain> References: <200408211621.28886.dmi_a@qnx.org.ru> <1093098853.4765.10.camel@athlon.localdomain> Message-ID: <200408211910.03063.dmi_a@qnx.org.ru> On Saturday 21 August 2004 18:34, Leonard den Ottolander wrote: > Hi Dmitry, > > On Sat, 2004-08-21 at 14:21, Dmitry Alexeyev wrote: > > I now can't browse RPM packages. > > > > mc gives "inconsistent extfs acrhive message" on all rpms. > > Does it give any more warnings than that? > Nope > What are the permissions on the file vfs/extfs/rpm? Executable for > everyone? I am using that same file and I am experiencing no > problems. > [dmi at dmi mc]$ LC_TIME=C ls -la /usr/local/share/mc/extfs/rpm* -rwxr-xr-x 1 root root 8464 Aug 21 19:04 /usr/local/share/mc/extfs/rpm -rwxr-xr-x 1 root root 1509 Aug 21 19:04 /usr/local/share/mc/extfs/rpms I just copied extfs/rpm file from mc 4.6.0 to this location and opened rpm successfully. > Might there be issues with the latest modifications to extfs.c? Probably no, since 4.6.0's rpm file works. > Are > you having problems with other archives as well? I tried only gz/bz/tgz and such, no problems. Do you want me to download some debs or similar and test? > If so revert that > file to before August 14 and see if that changes anything... > > Leonard. I have just located problem file (rpm) and will try to comment out some changes. Will keep you informed. Thanks Dmitry From leonard at den.ottolander.nl Sat Aug 21 15:10:41 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 17:10:41 +0200 Subject: extfs.c extfs_internal_stat: g_free(path2)? Message-ID: <1093101041.4765.33.camel@athlon.localdomain> Hi, Shouldn't path2 be g_freed in extfs_internal_stat? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Sat Aug 21 15:18:39 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Sat, 21 Aug 2004 19:18:39 +0400 Subject: Can't open any RPM file In-Reply-To: <200408211910.03063.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <1093098853.4765.10.camel@athlon.localdomain> <200408211910.03063.dmi_a@qnx.org.ru> Message-ID: <200408211918.39204.dmi_a@qnx.org.ru> [..] > I have just located problem file (rpm) and will try to comment out > some changes. Will keep you informed. > I changed line f="\"$1\"" in mcrpmfs_list () to f=$1 and was able to browse the package. Some problems with escaping again? WBR Dmitry From leonard at den.ottolander.nl Sat Aug 21 15:18:18 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 17:18:18 +0200 Subject: Can't open any RPM file In-Reply-To: <200408211910.03063.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <1093098853.4765.10.camel@athlon.localdomain> <200408211910.03063.dmi_a@qnx.org.ru> Message-ID: <1093101498.4765.36.camel@athlon.localdomain> Hi Dmitry, On Sat, 2004-08-21 at 17:10, Dmitry Alexeyev wrote: > I just copied extfs/rpm file from mc 4.6.0 to this location and opened > rpm successfully. I don't get it. In your previous mail you wrote: I tried to restore old /usr/local/share/extfs/rpm* files, but that doesn't solve the problem. So which is it? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Sat Aug 21 15:50:58 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 17:50:58 +0200 Subject: Can't open any RPM file In-Reply-To: <200408211918.39204.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <1093098853.4765.10.camel@athlon.localdomain> <200408211910.03063.dmi_a@qnx.org.ru> <200408211918.39204.dmi_a@qnx.org.ru> Message-ID: <1093103458.4765.42.camel@athlon.localdomain> Hi Dmitry, On Sat, 2004-08-21 at 17:18, Dmitry Alexeyev wrote: > I changed line > f="\"$1\"" > in mcrpmfs_list () to > f=$1 > > and was able to browse the package. > > Some problems with escaping again? Nothing has changed wrt this since January 23th. The only changes made to rpm lately are the adding of INFO/OBSOLETES and INFO/LICENSE. What version of bash are you using? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Sat Aug 21 15:58:38 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Sat, 21 Aug 2004 19:58:38 +0400 Subject: Can't open any RPM file In-Reply-To: <1093101498.4765.36.camel@athlon.localdomain> References: <200408211621.28886.dmi_a@qnx.org.ru> <200408211910.03063.dmi_a@qnx.org.ru> <1093101498.4765.36.camel@athlon.localdomain> Message-ID: <200408211958.38564.dmi_a@qnx.org.ru> > On Sat, 2004-08-21 at 17:10, Dmitry Alexeyev wrote: > > I just copied extfs/rpm file from mc 4.6.0 to this location and > > opened rpm successfully. > > I don't get it. In your previous mail you wrote: > > I tried to restore old /usr/local/share/extfs/rpm* files, but that > doesn't solve the problem. > > So which is it? Hi Leonard, Sorry it just wasn't old enough :) (but did differ in size, so confused me) It was from CVS dump made couple of days ago, I just didn't find anything older when writing first message. I have downloaded mc rpm after sending this to the list and then tested old extfs/rpm file. Sorry Dmitry From leonard at den.ottolander.nl Sat Aug 21 16:14:54 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 18:14:54 +0200 Subject: Missing changelog entries Message-ID: <1093104894.4765.45.camel@athlon.localdomain> Hello, Who is the one that touched a lot of files in vfs? Why are there no entries in the Changelog that describe the changes made? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Sat Aug 21 16:21:16 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 18:21:16 +0200 Subject: Can't open any RPM file In-Reply-To: <200408211958.38564.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <200408211910.03063.dmi_a@qnx.org.ru> <1093101498.4765.36.camel@athlon.localdomain> <200408211958.38564.dmi_a@qnx.org.ru> Message-ID: <1093105276.4765.51.camel@athlon.localdomain> Hi Dmitry, On Sat, 2004-08-21 at 17:58, Dmitry Alexeyev wrote: > Sorry it just wasn't old enough :) (but did differ in size, so confused > me) Yeah, those are the INFO/OBSOLETES and INFO/LICENSE changes. The escaping was introduced January 23, but I don't have problems with the extfs/rpm file. Running 4.6.0 + patches though, not latest CVS. Somebody heavily touched vfs on August 17, 20:57 GMT. No entries in the Changelog. Very bad. Not sure if it is related, but it could well be. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Sat Aug 21 16:28:03 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 18:28:03 +0200 Subject: Can't open any RPM file In-Reply-To: <200408211958.38564.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <200408211910.03063.dmi_a@qnx.org.ru> <1093101498.4765.36.camel@athlon.localdomain> <200408211958.38564.dmi_a@qnx.org.ru> Message-ID: <1093105683.4765.53.camel@athlon.localdomain> Hi Dmitry, Could you please check if you can use the new rpm file with an older version of mc? TIA. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Sat Aug 21 16:39:07 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Sat, 21 Aug 2004 20:39:07 +0400 Subject: Can't open any RPM file In-Reply-To: <1093105683.4765.53.camel@athlon.localdomain> References: <200408211621.28886.dmi_a@qnx.org.ru> <200408211958.38564.dmi_a@qnx.org.ru> <1093105683.4765.53.camel@athlon.localdomain> Message-ID: <200408212039.07621.dmi_a@qnx.org.ru> Hi Leonard, > > Could you please check if you can use the new rpm file with an older > version of mc? TIA. Nope, same error. Unfortunatelly Savannah CVS is fown at the moment, so I can't request cvs log/diff agains rpm, but IIRC there were some escaping related changes. Will see in some 3 hours. WBR Dmitry From jakub at redhat.com Sat Aug 21 15:02:42 2004 From: jakub at redhat.com (Jakub Jelinek) Date: Sat, 21 Aug 2004 17:02:42 +0200 Subject: Comprehensive patch escaping system/open calls in vfs/extfs In-Reply-To: <1093056396.4746.125.camel@athlon.localdomain> References: <1093056396.4746.125.camel@athlon.localdomain> Message-ID: <20040821150242.GY30497@sunsite.ms.mff.cuni.cz> On Sat, Aug 21, 2004 at 04:46:37AM +0200, Leonard den Ottolander wrote: > Hi, > > This is a comprehensive patch that escapes parameters to system and open > calls that spawn a shell. The patch affects a.in, apt.in, deba.in (parts > already committed), debd.in, deb.in (parts committed) and dpkg.in. > > Afaict parameters in mailfs.in, patchfs.in and uzip.in are properly > quoted. > > It's a compilation of the previous patches with a few fixes. Also > replaced the regular expression with a subroutine "quote". > > Please check for errors and omissions. You missed 3 places in a.in: --- extfs/a.in 2004-08-21 13:45:50.000000000 +0200 +++ extfs/a.in 2004-08-21 19:06:15.458358276 +0200 @@ -36,17 +36,20 @@ SWITCH: for ( $ARGV[0] ) { /mkdir/ && do { shift; shift; exit 1 if scalar(@ARGV) != 1; - system("$mmd $qdisk:/$ARGV[0] >/dev/null"); + $qname = quote($ARGV[0]); + system("$mmd $qdisk:/$qname >/dev/null"); exit 0; }; /rmdir/ && do { shift; shift; exit 1 if scalar(@ARGV) != 1; - system("$mrd $qdisk:/$ARGV[0] >/dev/null"); + $qname = quote($ARGV[0]); + system("$mrd $qdisk:/$qname >/dev/null"); exit 0; }; /rm/ && do { shift; shift; exit 1 if scalar(@ARGV) != 1; - system("$mdel $qdisk:/$ARGV[0] >/dev/null"); + $qname = quote($ARGV[0]); + system("$mdel $qdisk:/$qname >/dev/null"); exit 0; }; /copyout/ && do { shift; shift; Jakub From leonard at den.ottolander.nl Sat Aug 21 17:48:23 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 19:48:23 +0200 Subject: Can't open any RPM file In-Reply-To: <200408212106.25809.dmi_a@qnx.org.ru> References: <200408211621.28886.dmi_a@qnx.org.ru> <200408212039.07621.dmi_a@qnx.org.ru> <1093107062.4765.55.camel@athlon.localdomain> <200408212106.25809.dmi_a@qnx.org.ru> Message-ID: <1093110503.4765.60.camel@athlon.localdomain> Hi Dmitry, On Sat, 2004-08-21 at 19:06, Dmitry Alexeyev wrote: > [dmi at dmi dmi]$ rpm -q bash > bash-2.05b-alt6 Ok. I thought maybe some issues with bash 3. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Sat Aug 21 17:52:02 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sat, 21 Aug 2004 19:52:02 +0200 Subject: Comprehensive patch escaping system/open calls in vfs/extfs In-Reply-To: <20040821153201.GZ30497@sunsite.ms.mff.cuni.cz> References: <1093056396.4746.125.camel@athlon.localdomain> <20040821150242.GY30497@sunsite.ms.mff.cuni.cz> <1093109825.4765.58.camel@athlon.localdomain> <20040821153201.GZ30497@sunsite.ms.mff.cuni.cz> Message-ID: <1093110721.4765.64.camel@athlon.localdomain> On Sat, 2004-08-21 at 17:32, Jakub Jelinek wrote: > > Nope. The value of $ARGV[0] is already known because of the SWITCH. > > There are two shifts in between. You are correct. I had overseen those, so I had skipped them on purpose. That's why I responded too hastily, but double checking noticed you are correct. Scuze me and thanks for checking. Did you verify the other files as well? No other omissions? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Sun Aug 22 01:14:07 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 03:14:07 +0200 Subject: autogen.sh error: AS_HELP_STRING Message-ID: <1093137247.6868.4.camel@athlon.localdomain> Hello, When running autogen.sh it fails with the following error: configure:872: error: possibly undefined macro: AS_HELP_STRING If this token and others are legitimate, please use m4_pattern_allow. See the Autoconf documentation. Just a typo. Please apply: --- ri-gcc-warnings.m4.000 2004-08-17 11:41:19.000000000 +0200 +++ ri-gcc-warnings.m4 2004-08-22 03:09:28.000000000 +0200 @@ -4,7 +4,7 @@ dnl Adjust the CPPFLAGS, CFLAGS and CXXF AC_DEFUN([ri_GCC_WARNINGS], [AC_ARG_ENABLE([gcc-warnings], - [AS_HELP_STRING([--enable-gcc-warnings[[[[=error]]]]], + [AC_HELP_STRING([--enable-gcc-warnings[[[[=error]]]]], [enable additional gcc warning flags])], [ if test x"$GCC" = x"yes"; then Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Sun Aug 22 08:10:41 2004 From: pavelsh at mail.ru (pavelsh) Date: Sun, 22 Aug 2004 14:10:41 +0600 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <1093137247.6868.4.camel@athlon.localdomain> References: <1093137247.6868.4.camel@athlon.localdomain> Message-ID: <1721166974.20040822141041@mail.ru> Hello Leonard, Sunday, August 22, 2004, 7:14:07 AM, you wrote: LdO> When running autogen.sh it fails with the following error: LdO> configure:872: error: possibly undefined macro: AS_HELP_STRING LdO> If this token and others are legitimate, please use LdO> m4_pattern_allow. LdO> See the Autoconf documentation. LdO> Just a typo. Please apply: LdO> --- ri-gcc-warnings.m4.000 2004-08-17 11:41:19.000000000 +0200 LdO> +++ ri-gcc-warnings.m4 2004-08-22 03:09:28.000000000 +0200 LdO> @@ -4,7 +4,7 @@ dnl Adjust the CPPFLAGS, CFLAGS and CXXF LdO> AC_DEFUN([ri_GCC_WARNINGS], LdO> [AC_ARG_ENABLE([gcc-warnings], LdO> - [AS_HELP_STRING([--enable-gcc-warnings[[[[=error]]]]], LdO> + [AC_HELP_STRING([--enable-gcc-warnings[[[[=error]]]]], LdO> [enable additional gcc warning flags])], LdO> [ LdO> if test x"$GCC" = x"yes"; then Fixed. Thanks. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Sun Aug 22 10:00:51 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 12:00:51 +0200 Subject: More proposed patches Message-ID: <1093168850.4747.24.camel@athlon.localdomain> Hello Pavel, I've done some work on the patch tree at http://www.ottolander.nl/mc-patches/ . Please have another look at the proposed patches. There are a few that you looked at already, but you should look at them again. And there is a patch to complete the fix for CAN-2004-0494. can-2004-0494-p2.patch: Name says it all. Part two for the CAN-2004-0494 fix. Part 1 (under proposed-accepted) has already been committed (in phases) by Roland. This should fix all missing quoting of parameters to system/open calls in the vfs/extfs tree. edit-replace.patch: I think you misunderstood this patch. It is *not* UTF-8 related. It just happens to touch the same code. This fixes buffer overflows in the find/replace code in editcmd.c. mc-4.6.0-absoluterm.patch: Absolute path to /bin/rm in mc wrapper scripts. You've taken no decision on this yet. mc-4.6.0-large_syntax.patch: Fixes crash when using large syntax files. Have you already tested this? mc-4.6.0-pre3-nocpio.patch: No view of contents in initial rpm view. Copying files from the main view is very expensive (see Jakub Jelinek's email on this subject. If you do not apply this patch you should fix the temp file creation! Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Sun Aug 22 10:26:15 2004 From: pavelsh at mail.ru (pavelsh) Date: Sun, 22 Aug 2004 16:26:15 +0600 Subject: More proposed patches In-Reply-To: <1093168850.4747.24.camel@athlon.localdomain> References: <1093168850.4747.24.camel@athlon.localdomain> Message-ID: <693286704.20040822162615@mail.ru> Hello Leonard, Sunday, August 22, 2004, 4:00:51 PM, you wrote: LdO> can-2004-0494-p2.patch: Name says it all. Part two for the CAN-2004-0494 LdO> fix. Part 1 (under proposed-accepted) has already been committed (in LdO> phases) by Roland. This should fix all missing quoting of parameters to LdO> system/open calls in the vfs/extfs tree. LdO> edit-replace.patch: I think you misunderstood this patch. It is *not* LdO> UTF-8 related. It just happens to touch the same code. This fixes buffer LdO> overflows in the find/replace code in editcmd.c. Ok. I should review it. LdO> mc-4.6.0-absoluterm.patch: Absolute path to /bin/rm in mc wrapper LdO> scripts. You've taken no decision on this yet. It is security related? Why? LdO> mc-4.6.0-large_syntax.patch: Fixes crash when using large syntax files. LdO> Have you already tested this? No. I havn't enough time for it. This crash was fixed by to deny loading large syntax files now. But i should review this patch today or tomorrow. LdO> mc-4.6.0-pre3-nocpio.patch: No view of contents in initial rpm view. LdO> Copying files from the main view is very expensive (see Jakub Jelinek's LdO> email on this subject. If you do not apply this patch you should fix the LdO> temp file creation! Ok. I should test and commit this patch soon. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Sun Aug 22 10:27:02 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 12:27:02 +0200 Subject: More proposed patches In-Reply-To: <693286704.20040822162615@mail.ru> References: <1093168850.4747.24.camel@athlon.localdomain> <693286704.20040822162615@mail.ru> Message-ID: <1093170421.4747.52.camel@athlon.localdomain> Hi Pavel, On Sun, 2004-08-22 at 12:26, pavelsh wrote: > LdO> mc-4.6.0-absoluterm.patch: Absolute path to /bin/rm in mc wrapper > LdO> scripts. You've taken no decision on this yet. > > It is security related? Why? No. It just fixes the wrapper scripts if no PATH is set in the environment. Thanks for looking at the other patches. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From ossi at kde.org Sun Aug 22 12:34:32 2004 From: ossi at kde.org (Oswald Buddenhagen) Date: Sun, 22 Aug 2004 14:34:32 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <1093137247.6868.4.camel@athlon.localdomain> References: <1093137247.6868.4.camel@athlon.localdomain> Message-ID: <20040822123432.GA14774@ugly.local> On Sun, Aug 22, 2004 at 03:14:07AM +0200, Leonard den Ottolander wrote: > Just a typo. Please apply: > this is not a typo, but non-deprecated ac 2.58(?) syntax ... -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. From leonard at den.ottolander.nl Sun Aug 22 12:45:09 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 14:45:09 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <20040822123432.GA14774@ugly.local> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> Message-ID: <1093178709.4747.65.camel@athlon.localdomain> Hello Oswald, On Sun, 2004-08-22 at 14:34, Oswald Buddenhagen wrote: > On Sun, Aug 22, 2004 at 03:14:07AM +0200, Leonard den Ottolander wrote: > > Just a typo. Please apply: > > > this is not a typo, but non-deprecated ac 2.58(?) syntax ... Non-deprecated? You mean as in new? How do I avoid this error with autoconf-2.57? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From andrew at email.zp.ua Sun Aug 22 12:53:19 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Sun, 22 Aug 2004 15:53:19 +0300 (EEST) Subject: extfs.c extfs_internal_stat: g_free(path2)? In-Reply-To: <1093101041.4765.33.camel@athlon.localdomain> Message-ID: <200408221253.i7MCrJCN079297@email.zp.ua> Hi, Leonard! > > Shouldn't path2 be g_freed in extfs_internal_stat? You are right. Patch attached. -- Regards, Andrew V. Samoilov. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: extfs.c.patch URL: From leonard at den.ottolander.nl Sun Aug 22 12:54:33 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 14:54:33 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <20040822123432.GA14774@ugly.local> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> Message-ID: <1093179272.4747.68.camel@athlon.localdomain> On Sun, 2004-08-22 at 14:34, Oswald Buddenhagen wrote: > On Sun, Aug 22, 2004 at 03:14:07AM +0200, Leonard den Ottolander wrote: > > Just a typo. Please apply: > > > this is not a typo, but non-deprecated ac 2.58(?) syntax ... In which case the following change to m4.syntax is appropriate: --- m4.syntax.000 2002-09-29 22:22:25.000000000 +0200 +++ m4.syntax 2004-08-22 14:52:25.000000000 +0200 @@ -59,6 +59,7 @@ context default # Autoconf and Automake macros keyword whole AC_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan keyword whole AM_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan + keyword whole AS_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan #========================= # Variables Leonard. -- mount -t life -o ro /dev/dna /genetic/research From ossi at kde.org Sun Aug 22 12:58:17 2004 From: ossi at kde.org (Oswald Buddenhagen) Date: Sun, 22 Aug 2004 14:58:17 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <1093179272.4747.68.camel@athlon.localdomain> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> <1093179272.4747.68.camel@athlon.localdomain> Message-ID: <20040822125817.GA17218@ugly.local> On Sun, Aug 22, 2004 at 02:54:33PM +0200, Leonard den Ottolander wrote: > > this is not a typo, but non-deprecated ac 2.58(?) syntax ... > > In which case the following change to m4.syntax is appropriate: > > --- m4.syntax.000 2002-09-29 22:22:25.000000000 +0200 > +++ m4.syntax 2004-08-22 14:52:25.000000000 +0200 > + keyword whole AS_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan > yup -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. From leonard at den.ottolander.nl Sun Aug 22 13:07:23 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 15:07:23 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <20040822125817.GA17218@ugly.local> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> <1093179272.4747.68.camel@athlon.localdomain> <20040822125817.GA17218@ugly.local> Message-ID: <1093180042.4747.71.camel@athlon.localdomain> Hi Oswald, On Sun, 2004-08-22 at 14:58, Oswald Buddenhagen wrote: > > In which case the following change to m4.syntax is appropriate: > > > > --- m4.syntax.000 2002-09-29 22:22:25.000000000 +0200 > > +++ m4.syntax 2004-08-22 14:52:25.000000000 +0200 > > + keyword whole AS_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan > > > yup Still leaves the question how to work around this when using autoconf-2.57. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From ossi at kde.org Sun Aug 22 13:07:33 2004 From: ossi at kde.org (Oswald Buddenhagen) Date: Sun, 22 Aug 2004 15:07:33 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <1093178709.4747.65.camel@athlon.localdomain> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> <1093178709.4747.65.camel@athlon.localdomain> Message-ID: <20040822130733.GB17218@ugly.local> On Sun, Aug 22, 2004 at 02:45:09PM +0200, Leonard den Ottolander wrote: > On Sun, 2004-08-22 at 14:34, Oswald Buddenhagen wrote: > > On Sun, Aug 22, 2004 at 03:14:07AM +0200, Leonard den Ottolander wrote: > > > Just a typo. Please apply: > > > > > this is not a typo, but non-deprecated ac 2.58(?) syntax ... > > Non-deprecated? You mean as in new? > yes, and more. > How do I avoid this error with autoconf-2.57? > the simple version is a hack like this: m4_ifdef([AS_HELP_STRING], , [m4_define([AS_HELP_STRING], m4_defn([AC_HELP_STRING]))]) alternatively try this code that fixes some bracket quoting bugs; i hope the gnus merge it soon. m4_undefine([AS_HELP_STRING]) m4_define([AS_HELP_STRING], [m4_pushdef([AS_Prefix], m4_default([$3], [ ]))dnl m4_pushdef([AS_Prefix_Format], [ %-]m4_eval(m4_len(AS_Prefix) + 1)[s ])dnl [ %-23s ] m4_text_wrap([$2], AS_Prefix, [m4_format(AS_Prefix_Format, [[[$1]]])])dnl m4_popdef([AS_Prefix_Format])dnl m4_popdef([AS_Prefix])dnl ]) -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. From leonard at den.ottolander.nl Sun Aug 22 13:17:58 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Sun, 22 Aug 2004 15:17:58 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <20040822130733.GB17218@ugly.local> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> <1093178709.4747.65.camel@athlon.localdomain> <20040822130733.GB17218@ugly.local> Message-ID: <1093180678.4747.78.camel@athlon.localdomain> Hi Oswald, (Duh. Can't get used to mailing lists that let you automatically reply to the poster...) On Sun, 2004-08-22 at 15:07, Oswald Buddenhagen wrote: > > How do I avoid this error with autoconf-2.57? > > > the simple version is a hack like this: > m4_ifdef([AS_HELP_STRING], , [m4_define([AS_HELP_STRING], m4_defn([AC_HELP_STRING]))]) > > alternatively try this code that fixes some bracket quoting bugs; i hope > the gnus merge it soon. > m4_undefine([AS_HELP_STRING]) > m4_define([AS_HELP_STRING], > [m4_pushdef([AS_Prefix], m4_default([$3], [ ]))dnl > m4_pushdef([AS_Prefix_Format], > [ %-]m4_eval(m4_len(AS_Prefix) + 1)[s ])dnl [ %-23s ] > m4_text_wrap([$2], AS_Prefix, [m4_format(AS_Prefix_Format, [[[$1]]])])dnl > m4_popdef([AS_Prefix_Format])dnl > m4_popdef([AS_Prefix])dnl > ]) Thanks for your answer. I'll have to look a little closer into this before I can really understand this, so please excuse my ignorant question: Would just replacing AS_HELP_STRING with AC_HELP_STRING cause any undesired side effects (I mean for a vendor build, not in CVS)? There's only that one occurence. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Sun Aug 22 13:29:15 2004 From: pavelsh at mail.ru (pavelsh) Date: Sun, 22 Aug 2004 19:29:15 +0600 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <1093179272.4747.68.camel@athlon.localdomain> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> <1093179272.4747.68.camel@athlon.localdomain> Message-ID: <6916217.20040822192915@mail.ru> Hello Leonard, Sunday, August 22, 2004, 6:54:33 PM, you wrote: LdO> On Sun, 2004-08-22 at 14:34, Oswald Buddenhagen wrote: >> On Sun, Aug 22, 2004 at 03:14:07AM +0200, Leonard den Ottolander wrote: >> > Just a typo. Please apply: >> > >> this is not a typo, but non-deprecated ac 2.58(?) syntax ... LdO> In which case the following change to m4.syntax is appropriate: LdO> --- m4.syntax.000 2002-09-29 22:22:25.000000000 +0200 LdO> +++ m4.syntax 2004-08-22 14:52:25.000000000 +0200 LdO> @@ -59,6 +59,7 @@ context default LdO> # Autoconf and Automake macros LdO> keyword whole AC_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan LdO> keyword whole AM_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan LdO> + keyword whole AS_\[ABCDEFGHIJKLMNOPQRSTUVWXYZ_\] cyan LdO> #========================= LdO> # Variables Applied. -- Best regards, pavelsh mailto:pavelsh at mail.ru From andrew at email.zp.ua Sun Aug 22 13:52:48 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Sun, 22 Aug 2004 16:52:48 +0300 (EEST) Subject: [patch]: undelfs.c warning fixes Message-ID: <200408221352.i7MDqm0f081110@email.zp.ua> Hi, -- Regards, Andrew V. Samoilov. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: undelfs.c.patch URL: From ossi at kde.org Sun Aug 22 14:21:41 2004 From: ossi at kde.org (Oswald Buddenhagen) Date: Sun, 22 Aug 2004 16:21:41 +0200 Subject: autogen.sh error: AS_HELP_STRING In-Reply-To: <1093180678.4747.78.camel@athlon.localdomain> References: <1093137247.6868.4.camel@athlon.localdomain> <20040822123432.GA14774@ugly.local> <1093178709.4747.65.camel@athlon.localdomain> <20040822130733.GB17218@ugly.local> <1093180678.4747.78.camel@athlon.localdomain> Message-ID: <20040822142141.GA20237@ugly.local> On Sun, Aug 22, 2004 at 03:17:58PM +0200, Leonard den Ottolander wrote: > Would just replacing AS_HELP_STRING with AC_HELP_STRING cause any > undesired side effects (I mean for a vendor build, not in CVS)? There's > only that one occurence. > no, it just causes the ugly deprecation warning. unfortunately many users are too stupid to recognize a warning as such and report these "errors" if some other problem occurs. -- Hi! I'm a .signature virus! Copy me into your ~/.signature, please! -- Chaos, panic, and disorder - my work here is done. From pavelsh at mail.ru Sun Aug 22 20:53:27 2004 From: pavelsh at mail.ru (pavelsh) Date: Mon, 23 Aug 2004 02:53:27 +0600 Subject: More proposed patches In-Reply-To: <1093168850.4747.24.camel@athlon.localdomain> References: <1093168850.4747.24.camel@athlon.localdomain> Message-ID: <346631394.20040823025327@mail.ru> Hello Leonard, Sunday, August 22, 2004, 4:00:51 PM, you wrote: LdO> mc-4.6.0-large_syntax.patch: Fixes crash when using large syntax files. LdO> Have you already tested this? Applied and committed. -- Best regards, pavelsh mailto:pavelsh at mail.ru From pavelsh at mail.ru Sun Aug 22 21:52:00 2004 From: pavelsh at mail.ru (pavelsh) Date: Mon, 23 Aug 2004 03:52:00 +0600 Subject: More proposed patches In-Reply-To: <1093168850.4747.24.camel@athlon.localdomain> References: <1093168850.4747.24.camel@athlon.localdomain> Message-ID: <103225107.20040823035200@mail.ru> Hello Leonard, Sunday, August 22, 2004, 4:00:51 PM, you wrote: LdO> mc-4.6.0-pre3-nocpio.patch: No view of contents in initial rpm view. LdO> Copying files from the main view is very expensive (see Jakub Jelinek's LdO> email on this subject. If you do not apply this patch you should fix the LdO> temp file creation! Applied. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Mon Aug 23 19:14:56 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Mon, 23 Aug 2004 21:14:56 +0200 Subject: --enable-charset Message-ID: <1093288496.4751.53.camel@athlon.localdomain> Hi, Just curious what the configure option --enable-charset is meant to achieve. Could anyone explain? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Mon Aug 23 22:06:50 2004 From: roland.illig at gmx.de (Roland Illig) Date: Tue, 24 Aug 2004 00:06:50 +0200 Subject: Missing changelog entries In-Reply-To: <1093104894.4765.45.camel@athlon.localdomain> References: <1093104894.4765.45.camel@athlon.localdomain> Message-ID: <412A6A7A.6080601@gmx.de> Leonard den Ottolander wrote: > Hello, > > Who is the one that touched a lot of files in vfs? Why are there no > entries in the Changelog that describe the changes made? It has been me. I changed many occurences of "char *" to "const char *", but I'm not sure what else I did. I will have a thorough look at it tomorrow and either complete the ChangeLog or revert the patches. Sorry, Roland From leonard at den.ottolander.nl Mon Aug 23 23:05:00 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 24 Aug 2004 01:05:00 +0200 Subject: Missing changelog entries In-Reply-To: <412A6A7A.6080601@gmx.de> References: <1093104894.4765.45.camel@athlon.localdomain> <412A6A7A.6080601@gmx.de> Message-ID: <1093302300.6236.9.camel@athlon.localdomain> Hi Roland, On Tue, 2004-08-24 at 00:06, Roland Illig wrote: > It has been me. I changed many occurences of "char *" to "const char *", > but I'm not sure what else I did. I will have a thorough look at it > tomorrow and either complete the ChangeLog or revert the patches. It seems you made a start with fixing some memory leaks, g_free-ing duplicated strings here and there. But you missed one or two in files you touched (see thread "extfs.c extfs_internal_stat: g_free(path2)?"). Not sure what else you did. Maybe it's a good idea to work with local copies/patches that you only commit once they are completely fixed/finished. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 24 00:15:16 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 24 Aug 2004 02:15:16 +0200 Subject: edit-replace patch fix Message-ID: <1093306516.6236.12.camel@athlon.localdomain> Hi, I added a little fix to the edit-replace patch (int n -> size_t n and strncpy -> memcpy). It should now patch cleanly against CVS. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Tue Aug 24 07:47:38 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Tue, 24 Aug 2004 11:47:38 +0400 Subject: --enable-charset In-Reply-To: <1093288496.4751.53.camel@athlon.localdomain> References: <1093288496.4751.53.camel@athlon.localdomain> Message-ID: <200408241147.38755.dmi_a@qnx.org.ru> Hi Leonard, > > Just curious what the configure option --enable-charset is meant to > achieve. Could anyone explain? It enables selection of charset in editor (and probably viewer) when pressing Ctrl+T. Very useful feauture for Russians ;] Bear in mind we have 6 cyrillic encodings, and Windows and Linux/UNIX use different ones. WBR Dmitry > > Leonard. From leonard at den.ottolander.nl Tue Aug 24 14:40:05 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 24 Aug 2004 16:40:05 +0200 Subject: Drawing errors Message-ID: <1093358404.4815.13.camel@athlon.localdomain> Hi, Currently running mv from CVS 2004-08-22, with a few patches that will be soon merged. I am *not* using any UTF-8 patches at the moment. I am seeing drawing issues. As soon as I start mc I notice the straight line under the files is not terminated correctly. Yesterday I also noticed it being continued on the next line, and after toggling I now see that the last line "re-begins" halfway through. Since I have been running a patched 4.6.0 until recently I am not sure when this bug was introduced. Any idea where I should look to fix this? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From makovick at kmlinux.fjfi.cvut.cz Tue Aug 24 14:59:16 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Tue, 24 Aug 2004 16:59:16 +0200 Subject: Drawing errors In-Reply-To: <1093358404.4815.13.camel@athlon.localdomain> References: <1093358404.4815.13.camel@athlon.localdomain> Message-ID: <412B57C4.5010807@kmlinux.fjfi.cvut.cz> Leonard den Ottolander wrote: > Hi, > > Currently running mv from CVS 2004-08-22, with a few patches that will > be soon merged. I am *not* using any UTF-8 patches at the moment. > > I am seeing drawing issues. As soon as I start mc I notice the straight > line under the files is not terminated correctly. Yesterday I also > noticed it being continued on the next line, and after toggling I now > see that the last line "re-begins" halfway through. > > Since I have been running a patched 4.6.0 until recently I am not sure > when this bug was introduced. Any idea where I should look to fix this? There's a off-by-one error in certain patched versions of SLang, at least some RetHat versions and Debian versions. In Debian, it has been finally fixed in latest unstable distribution. -- Jindrich Makovicka From roland.illig at gmx.de Tue Aug 24 15:20:04 2004 From: roland.illig at gmx.de (Roland Illig) Date: Tue, 24 Aug 2004 17:20:04 +0200 Subject: Bugfixes for vfs Message-ID: <412B5CA4.8040200@gmx.de> Hi, I just discovered some memory leaks that I have introduced myself. May I commit them? Roland -------------- next part -------------- A non-text attachment was scrubbed... Name: mc-vfs-memory-leaks.patch Type: text/x-patch Size: 1785 bytes Desc: not available URL: From leonard at den.ottolander.nl Tue Aug 24 15:35:07 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 24 Aug 2004 17:35:07 +0200 Subject: Drawing errors In-Reply-To: <412B57C4.5010807@kmlinux.fjfi.cvut.cz> References: <1093358404.4815.13.camel@athlon.localdomain> <412B57C4.5010807@kmlinux.fjfi.cvut.cz> Message-ID: <1093361707.4815.17.camel@athlon.localdomain> Hello Jindrich, On Tue, 2004-08-24 at 16:59, Jindrich Makovicka wrote: > Leonard den Ottolander wrote: > > I am seeing drawing issues. As soon as I start mc I notice the straight > > line under the files is not terminated correctly. Yesterday I also > > noticed it being continued on the next line, and after toggling I now > > see that the last line "re-begins" halfway through. > There's a off-by-one error in certain patched versions of SLang, at > least some RetHat versions and Debian versions. In Debian, it has been > finally fixed in latest unstable distribution. I was thinking slang might be involved. Could it cause these drawing errors? Can you provide more details? Where is this off by one located? Maybe you have a patch lying around? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Tue Aug 24 15:39:21 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 24 Aug 2004 17:39:21 +0200 Subject: Bugfixes for vfs In-Reply-To: <412B5CA4.8040200@gmx.de> References: <412B5CA4.8040200@gmx.de> Message-ID: <1093361960.4815.22.camel@athlon.localdomain> Hi Roland, On Tue, 2004-08-24 at 17:20, Roland Illig wrote: > I just discovered some memory leaks that I have introduced myself. May I > commit them? I already mentioned the fact that you missed one or two g_free's a couple of days ago. Please read my post from last night, and the one I referred to from Andrew Smile-Love :) . And yes, of course, please fix/commit. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From roland.illig at gmx.de Tue Aug 24 16:09:31 2004 From: roland.illig at gmx.de (Roland Illig) Date: Tue, 24 Aug 2004 18:09:31 +0200 Subject: Missing changelog entries In-Reply-To: <1093302300.6236.9.camel@athlon.localdomain> References: <1093104894.4765.45.camel@athlon.localdomain> <412A6A7A.6080601@gmx.de> <1093302300.6236.9.camel@athlon.localdomain> Message-ID: <412B683B.6040203@gmx.de> Leonard den Ottolander wrote: > Hi Roland, > It seems you made a start with fixing some memory leaks, g_free-ing > duplicated strings here and there. But you missed one or two in files > you touched (see thread "extfs.c extfs_internal_stat: g_free(path2)?"). > Not sure what else you did. Most of what I did was changing "char *" to "const char *" in many places, some of them requiring changes in the calling functions. I have committed the missing ChangeLog entries in /src, /edit and /vfs (dated 2004-08-16). Please have a look at them for details. I just commited the patch to the g_free issues that I found while reviewing my own changes. Roland From buliabyak at gmail.com Wed Aug 25 00:37:15 2004 From: buliabyak at gmail.com (bulia byak) Date: Tue, 24 Aug 2004 21:37:15 -0300 Subject: Maintenance question Message-ID: <3c78ff0304082417373a8500f3@mail.gmail.com> Leonard den Ottolander wrote: > Before handing over (some of) the responsibilities for the project maybe > you could present us with an overview of the current state of affairs. I support this request. I haven't been watching mc development closely in the last year, mostly because I had a feeling that it was stagnating. On the other hand, the rate of patches and even fork attempts clearly indicates that there's still a lot of interest in the program. So I think we should now attempt to aggressively revitalize the project. At the first stage, this means making the project more open and transparent. The current policies for patches, reviews, commits, and granting developer access should be clearly stated, discussed, amended if necessary, and written down somewhere on the web site. Personally, I hope that these policies can be relaxed somewhat. People who want to help must have an easy way to do so. For example, patches should be reviewed, but it's really frustrating to wait for this review forever. Besides, I think it's only core or security-related patches that really need pre-commit review; for UI tweaks, much better feedback can be received from users of the CVS version than from a single reviewer, and it's also much faster for other developers to do cvs update and test than to apply a patch. I also think that the project must be more liberal in granting developer access, because working via patches and not having up-to-the-minute CVS access is very frustrating. My own area of interest is the UI, and there's a lot of room for improvement here. My 2-years-old patch in the tracker ("hotkeys in hotlist") was never committed and has probably bitrotten by now. I also had a number of other UI improvements in my tree, but I never even submitted them as patches because that first patch was ignored. I really hope the situation will change and I will be able to contribute again. By the way, I am a developer for Inkscape (www.inkscape.org), and I can't help noticing the similarities. Inkscape was forked half a year ago from Sodipodi, because the developers were frustrated by not having CVS access, long periods of stagnation, ignored patches, and other unpleasant signs of a "too centralized" project. I don't want to sound like boasting, but Inkscape has been a success. It now has a MUCH more usable UI, lots of new features, important architectural changes, and surprisingly, it's also more stable (rarely crashes, unlike Sodipodi) despite having several times more developers with CVS access. The openness of Inkscape has played out really well. Of course mc is a different kind of project, but it would still benefit by borrowing something from Inkscape's approach IMHO. At the very least, we found that additional channels of communication and collaboration, such as Wiki and Jabber/IRC channel, are very useful in attracting new developers and helping them get familiar with the codebase. Adding a news list to the front page of the site would also be nice; developers usually first announce the important changes, new features etc on the list, and from there a person posts them on the web site periodically (see front page of www.inkscape.org). Looks like a small thing, but it's very important to for the project to look alive and well. And of course, it's important to make releases more often, so that a wider audience can test the new stuff. We could have a stable release and a series of testing releases, so that those who want stability would be able to stick with something that is well tested. I would appreciate everyone's thoughts on this. From roland.illig at gmx.de Wed Aug 25 02:42:30 2004 From: roland.illig at gmx.de (Roland Illig) Date: Wed, 25 Aug 2004 04:42:30 +0200 Subject: Maintenance question In-Reply-To: <3c78ff0304082417373a8500f3@mail.gmail.com> References: <3c78ff0304082417373a8500f3@mail.gmail.com> Message-ID: <412BFC96.7020408@gmx.de> bulia byak wrote: > [...] > Personally, I hope that these policies can be relaxed somewhat. People > who want to help must have an easy way to do so. For example, patches > should be reviewed, but it's really frustrating to wait for this > review forever. Besides, I think it's only core or security-related > patches that really need pre-commit review; for UI tweaks, much better > feedback can be received from users of the CVS version than from a > single reviewer, and it's also much faster for other developers to do > cvs update and test than to apply a patch. I also think that the > project must be more liberal in granting developer access, because > working via patches and not having up-to-the-minute CVS access is very > frustrating. That's also my point of view. Some other projects I know (pkgsrc-wip, MoinMoin, buildtool, and my own towers-of-hanoi) have a $project-cvs mailing list that gets a mail on every CVS commit. Especially for pkgsrc-wip, there are often comments from the reviewers that go to the $project-discuss mailing list. > [...] And of course, it's important to make releases more > often, so that a wider audience can test the new stuff. We could have > a stable release and a series of testing releases, so that those who > want stability would be able to stick with something that is well > tested. I also agree to this. I have been thinking about an mc-unstable CVS branch where everyone may commit freely. But I like the above idea (mc-cvs mailing list) much more. Roland From nadvornik at suse.cz Wed Aug 25 09:15:16 2004 From: nadvornik at suse.cz (Vladimir Nadvornik) Date: Wed, 25 Aug 2004 11:15:16 +0200 Subject: Drawing errors In-Reply-To: <1093361707.4815.17.camel@athlon.localdomain> References: <1093358404.4815.13.camel@athlon.localdomain> <412B57C4.5010807@kmlinux.fjfi.cvut.cz> <1093361707.4815.17.camel@athlon.localdomain> Message-ID: <200408251115.16433.nadvornik@suse.cz> On Tuesday 24 August 2004 17:35, Leonard den Ottolander wrote: > Hello Jindrich, > > On Tue, 2004-08-24 at 16:59, Jindrich Makovicka wrote: > > Leonard den Ottolander wrote: > > > I am seeing drawing issues. As soon as I start mc I notice the straight > > > line under the files is not terminated correctly. Yesterday I also > > > noticed it being continued on the next line, and after toggling I now > > > see that the last line "re-begins" halfway through. > > > > There's a off-by-one error in certain patched versions of SLang, at > > least some RetHat versions and Debian versions. In Debian, it has been > > finally fixed in latest unstable distribution. > > I was thinking slang might be involved. Could it cause these drawing > errors? Can you provide more details? Where is this off by one located? > Maybe you have a patch lying around? > It is probably this patch: # Bug fix http://bugs.debian.org/192025 # off-by-one error when drawing horizontal lines # diff -ruN slang-1.4.9-old/src/slsmg.c slang-1.4.9/src/slsmg.c --- slang-1.4.9-old/src/slsmg.c 2004-06-25 10:20:57.000000000 +0100 +++ slang-1.4.9/src/slsmg.c 2004-06-25 10:11:09.000000000 +0100 @@ -1709,7 +1709,7 @@ This_Color |= ALT_CHAR_FLAG; This_Col = cmin; - SLsmg_draw_object(This_Row, This_Col, ch); + // SLsmg_draw_object(This_Row, This_Col, ch); while (n-- > 0) { SLsmg_draw_object(This_Row, This_Col, ch); -- Vladimir Nadvornik developer --------------------------------------------------------------------- SuSE CR, s.r.o. e-mail: nadvornik at suse.cz Drahobejlova 27 tel:+420 2 9654 2373 190 00 Praha 9 fax:+420 2 9654 2374 Ceska republika http://www.suse.cz From leonard at den.ottolander.nl Wed Aug 25 13:09:26 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 25 Aug 2004 15:09:26 +0200 Subject: Drawing errors In-Reply-To: <200408251115.16433.nadvornik@suse.cz> References: <1093358404.4815.13.camel@athlon.localdomain> <412B57C4.5010807@kmlinux.fjfi.cvut.cz> <1093361707.4815.17.camel@athlon.localdomain> <200408251115.16433.nadvornik@suse.cz> Message-ID: <1093439365.4788.6.camel@athlon.localdomain> Hi Vladimir, On Wed, 2004-08-25 at 11:15, Vladimir Nadvornik wrote: > > I was thinking slang might be involved. Could it cause these drawing > > errors? Can you provide more details? Where is this off by one located? > > Maybe you have a patch lying around? > > > > It is probably this patch: > > # Bug fix http://bugs.debian.org/192025 > # off-by-one error when drawing horizontal lines > # > diff -ruN slang-1.4.9-old/src/slsmg.c slang-1.4.9/src/slsmg.c > --- slang-1.4.9-old/src/slsmg.c 2004-06-25 10:20:57.000000000 +0100 > +++ slang-1.4.9/src/slsmg.c 2004-06-25 10:11:09.000000000 +0100 > @@ -1709,7 +1709,7 @@ > This_Color |= ALT_CHAR_FLAG; > This_Col = cmin; > > - SLsmg_draw_object(This_Row, This_Col, ch); > + // SLsmg_draw_object(This_Row, This_Col, ch); > while (n-- > 0) > { > SLsmg_draw_object(This_Row, This_Col, ch); Indeed it is. Figured it out with some help from Jindrich. Proposed it for inclusion in Fedora/Red Hat. Is SUSE already using that patch? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Wed Aug 25 13:46:48 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 25 Aug 2004 15:46:48 +0200 Subject: Maintenance question In-Reply-To: <3c78ff0304082417373a8500f3@mail.gmail.com> References: <3c78ff0304082417373a8500f3@mail.gmail.com> Message-ID: <1093441607.4788.35.camel@athlon.localdomain> Hello Bulia, On Wed, 2004-08-25 at 02:37, bulia byak wrote: > Personally, I hope that these policies can be relaxed somewhat. People > who want to help must have an easy way to do so. For example, patches > should be reviewed, but it's really frustrating to wait for this > review forever. Besides, I think it's only core or security-related > patches that really need pre-commit review; for UI tweaks, much better > feedback can be received from users of the CVS version than from a > single reviewer, and it's also much faster for other developers to do > cvs update and test than to apply a patch. I already figured that trying to regulate things too much is probably counter productive at this stage of revitalizing the project. Otoh I would prefer to see a little reluctance towards committing (incomplete/untested) patches too quickly. Give yourself a week to check your own work before committing. This is why I propose to put up patches for review, also by people with commit access. It's still possible to commit them after a while if no-one responded. Waiting with committing until a piece of work is complete also makes the Changelogs more useful to work with. Instead of having to read through dozens of Changelog entries to get an impression of what changed it's easier to have a single (large) Changelog entry describing a whole patch. This is especially important if people need to touch the same part of the tree. > I also think that the > project must be more liberal in granting developer access, because > working via patches and not having up-to-the-minute CVS access is very > frustrating. Personally I do not have a big problem with working with patches, assuming they will be reviewed and hopefully committed in a reasonable time. Frustration only creeps in if they are not reviewed within a reasonable time, or not at all. Roland and Pavel Shirshov have been very responsive in this respect, although I wonder if they can keep up the pace if they remain with only the two of them ;) . If there are enough people with commit access (say 10) people submitting patches can be sure theirs will be reviewed in a reasonable time. Or they can let them be peer reviewed and present them as such, thus taking away some of the burden of the reviewers/committers. > My own area of interest is the UI, and there's a lot of room for > improvement here. My 2-years-old patch in the tracker ("hotkeys in > hotlist") was never committed and has probably bitrotten by now. I > also had a number of other UI improvements in my tree, but I never > even submitted them as patches because that first patch was ignored. I > really hope the situation will change and I will be able to contribute > again. It is my intention to do some work on Savannah Bugzilla. Maybe I should ask Pavel Roskin directly, but actually I am waiting for a reply to my mail that started this thread. > By the way, I am a developer for Inkscape (www.inkscape.org), and I > can't help noticing the similarities. Inkscape was forked half a year > ago from Sodipodi, because the developers were frustrated by not > having CVS access, long periods of stagnation, ignored patches, and > other unpleasant signs of a "too centralized" project. I don't want to > sound like boasting, but Inkscape has been a success. It now has a > MUCH more usable UI, lots of new features, important architectural > changes, and surprisingly, it's also more stable (rarely crashes, > unlike Sodipodi) despite having several times more developers with CVS > access. The openness of Inkscape has played out really well. Of course > mc is a different kind of project, but it would still benefit by > borrowing something from Inkscape's approach IMHO. > > At the very least, we found that additional channels of communication > and collaboration, such as Wiki and Jabber/IRC channel, are very > useful in attracting new developers and helping them get familiar with > the codebase. I reserved #mc-devel at Freenode for this just yesterday. Please come and join me there. It's rather boring only chatting to ChanServ ;-) . > Looks like a small thing, but it's very important to for the project to look > alive and well. Very true. > And of course, it's important to make releases more > often, so that a wider audience can test the new stuff. Sure. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Wed Aug 25 15:55:24 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Wed, 25 Aug 2004 17:55:24 +0200 Subject: PHP syntax In-Reply-To: <1092842264.4777.71.camel@athlon.localdomain> References: <18210192520.20040817175048@mail.ru> <1092752550.4746.3.camel@athlon.localdomain> <1189380635.20040818020712@mail.ru> <1092774510.4746.6.camel@athlon.localdomain> <9210557191.20040818025714@mail.ru> <1092779172.4746.70.camel@athlon.localdomain> <14410233326.20040818051010@mail.ru> <1092784853.6246.37.camel@athlon.localdomain> <1092842264.4777.71.camel@athlon.localdomain> Message-ID: <1093449323.4788.44.camel@athlon.localdomain> Hi, I wrote: > I think keeping the whole strings including the "->"'s is a good idea, > as it only highlights correct class/method combinations. Thought about this a bit more, and came to the conclusion that in code one doesn't use "class->method", only instances of classes. I've regenerated the function list from the PHP docs. Now my question is what I should do with the classes and methods. My current opinion is to just colour the classes as normal keywords, and use "->method" as an identifier for methods. This way you can see whether the method is actually a method (whole string including "->" coloured), or you are just using a keyword as a method. Of course the issue that methods that are not valid for the instantiated class are also highlighted remains, but that is because the syntax highlighting is not "smart". What do you think? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From buliabyak at gmail.com Wed Aug 25 19:59:10 2004 From: buliabyak at gmail.com (bulia byak) Date: Wed, 25 Aug 2004 16:59:10 -0300 Subject: Maintenance question In-Reply-To: <1093441607.4788.35.camel@athlon.localdomain> References: <3c78ff0304082417373a8500f3@mail.gmail.com> <1093441607.4788.35.camel@athlon.localdomain> Message-ID: <3c78ff0304082512593add2431@mail.gmail.com> On Wed, 25 Aug 2004 15:46:48 +0200, Leonard den Ottolander wrote: > I already figured that trying to regulate things too much is probably > counter productive at this stage of revitalizing the project. True! > would prefer to see a little reluctance towards committing > (incomplete/untested) patches too quickly. Give yourself a week to check > your own work before committing. This is why I propose to put up patches > for review, also by people with commit access. It's still possible to > commit them after a while if no-one responded. Yes, that sounds like an ideal policy: "silence is a sign of approval". If you have concerns, voice them, otherwise it gets committed by default. > Waiting with committing until a piece of work is complete also makes the > Changelogs more useful to work with. Instead of having to read through > dozens of Changelog entries to get an impression of what changed it's > easier to have a single (large) Changelog entry describing a whole > patch. This is especially important if people need to touch the same > part of the tree. This is true, but it is also true that smaller commits make it easier to find the exact point where something broke, by testing dated checkouts. The point of CVS is to make changes easily traceable; by lumping together many patches into one commit we lose this advantage. So we need some sort of a balance here. In Inkscape, our policy is simple: the CVS must compile and basically work at all times; other than that it's up to the developers how big or small or big are their commits. We also don't have the requirement that each commit must be commented in the changelog; you can do several small commits over several days, and then describe them in one descriptive changelog entry. And if anyone needs to see comments for each small commit, that's what CVS history is for. > Personally I do not have a big problem with working with patches, > assuming they will be reviewed and hopefully committed in a reasonable > time. Frustration only creeps in if they are not reviewed within a > reasonable time, or not at all. Roland and Pavel Shirshov have been very > responsive in this respect, although I wonder if they can keep up the > pace if they remain with only the two of them ;) . Yes, I think 2 or 3 developers with CVS access is certainly too low for a project like this. The best ratio of the committing developers to all active developers can be argued, but I think it's obviously too low now. By the way we could contact the fork projects (such as Arpi with his AMC) to see if they would be interested to merge their stuff back. With AMC in particular, there was some talk of merging in 2003 but it apparently didn't play out for some reason. From pavelsh at mail.ru Wed Aug 25 22:42:39 2004 From: pavelsh at mail.ru (pavelsh) Date: Thu, 26 Aug 2004 04:42:39 +0600 Subject: More proposed patches In-Reply-To: <1093168850.4747.24.camel@athlon.localdomain> References: <1093168850.4747.24.camel@athlon.localdomain> Message-ID: <1328981333.20040826044239@mail.ru> Hello Leonard, Sunday, August 22, 2004, 4:00:51 PM, you wrote: LdO> edit-replace.patch: I think you misunderstood this patch. It is *not* LdO> UTF-8 related. It just happens to touch the same code. This fixes buffer LdO> overflows in the find/replace code in editcmd.c. Applied. LdO> mc-4.6.0-absoluterm.patch: Absolute path to /bin/rm in mc wrapper LdO> scripts. You've taken no decision on this yet. This patch is os distribution related. -- Best regards, pavelsh mailto:pavelsh at mail.ru From roland.illig at gmx.de Wed Aug 25 23:57:19 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 26 Aug 2004 01:57:19 +0200 Subject: Maintenance question In-Reply-To: <3c78ff0304082512593add2431@mail.gmail.com> References: <3c78ff0304082417373a8500f3@mail.gmail.com> <1093441607.4788.35.camel@athlon.localdomain> <3c78ff0304082512593add2431@mail.gmail.com> Message-ID: <412D275F.5050205@gmx.de> bulia byak wrote: > Yes, I think 2 or 3 developers with CVS access is certainly too low > for a project like this. The best ratio of the committing developers > to all active developers can be argued, but I think it's obviously too > low now. See the URL below for a list of all 15 project members with CVS access. There aren't just two or three. https://savannah.gnu.org/project/memberlist.php?group=mc&detailed=0 Roland From makovick at kmlinux.fjfi.cvut.cz Thu Aug 26 08:02:56 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Thu, 26 Aug 2004 10:02:56 +0200 Subject: [patch] more sesponsive file search dialog Message-ID: <412D9930.3000604@kmlinux.fjfi.cvut.cz> Hello, Currently, during the file content search, no events are checked, which is very annoying when mc hits a big file and the UI freezes. This patch adds event checks to the code for searching the file contents. After applying, the Find File dialog works all the time during the search, so the searching can be stopped or suspended&resumed even in the middle of a file. For suspend&resume, the code is modified to make sure the search will continue where it previously ended. This is about third time I try to submit this patch and hope it could be finally included in some form. Regards, -- Jindrich Makovicka -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: find.c.diff URL: From makovick at kmlinux.fjfi.cvut.cz Thu Aug 26 08:22:59 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Thu, 26 Aug 2004 10:22:59 +0200 Subject: [patch] bugfix for strip_password Message-ID: <412D9DE3.9060600@kmlinux.fjfi.cvut.cz> Hi, In current mc, strip_password misbehaves for URLs like /#ftp:user:pass at hostname/path/path/path which makes mc display only the string in the form /#ftp:user at hostname in the title bar, instead of (IMHO correct) /#ftp:user at hostname/path/path/path The attached patch attempts to re-implement the function with the correct behavior. -- Jindrich Makovicka -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: util.c.diff URL: From roland.illig at gmx.de Thu Aug 26 09:17:45 2004 From: roland.illig at gmx.de (Roland Illig) Date: Thu, 26 Aug 2004 11:17:45 +0200 Subject: [patch]: undelfs.c warning fixes In-Reply-To: <200408221352.i7MDqm0f081110@email.zp.ua> References: <200408221352.i7MDqm0f081110@email.zp.ua> Message-ID: <412DAAB9.4030306@gmx.de> Andrew V. Samoilov wrote: > vfs/ChangeLog: > > * undelfs.c (undelfs_lstat): Adjust declaration. > (undelfs_stat): Make macro. > > --- mc/vfs/undelfs.c~ Tue Aug 17 12:53:13 2004 > +++ mc/vfs/undelfs.c Sun Aug 22 16:43:27 2004 Committed. Roland From leonard at den.ottolander.nl Thu Aug 26 09:25:09 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 26 Aug 2004 11:25:09 +0200 Subject: Bug in search backwards Message-ID: <1093512309.4786.8.camel@athlon.localdomain> Hi, Not sure who implemented the backward search, but there's a little bug in it. On find the first character is chopped of the sentence. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 26 16:00:37 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 26 Aug 2004 18:00:37 +0200 Subject: TODO 4.6.1 Message-ID: <1093536037.4786.103.camel@athlon.localdomain> Hi, Had some discussion with Pavel Shirshov on IRC today (Freenode #mc-devel). From a link he presented me I figured out that entry 3 ("Crash when opening tar archives") on the current TODO for 4.6.1 is already fixed in CVS. See http://mail.gnome.org/archives/mc-devel/2004-March/msg00003.html . If I am confusing issues please let me know, although I do not see a crash when viewing an archive with links in it's root. Regarding entry 2 ("separate color for hotkeys"), that's an easy fix. I chose for white on a black background (this seems the best pick from the default_colors - (bold) black on red might be nice but is not defined): --- mc/src/color.c.000 2004-08-16 05:15:17.000000000 +0200 +++ mc/src/color.c 2004-08-26 17:31:39.000000000 +0200 @@ -269,7 +269,7 @@ alarm_colors [0] = ERROR_COLOR; alarm_colors [1] = REVERSE_COLOR; - alarm_colors [2] = ERROR_COLOR; + alarm_colors [2] = MENU_SELECTED_COLOR; alarm_colors [3] = COLOR_HOT_NORMAL; } Entry 4 ("Tilde not expanded") should probably also be easy to fix, but I don't know where to look. Anyone? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From leonard at den.ottolander.nl Thu Aug 26 17:12:01 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Thu, 26 Aug 2004 19:12:01 +0200 Subject: extfs rpm file name quoting Message-ID: <1093540321.4786.112.camel@athlon.localdomain> Hi, The sed command used in mcrpmfs_list() should probably also be used in mcrpmfs_copyout(): --- rpm.000 2004-08-26 17:44:13.000000000 +0200 +++ rpm 2004-08-26 19:02:32.000000000 +0200 @@ -133,7 +133,7 @@ mcrpmfs_copyout () { - f="\"$1\"" + f="`echo "$1" | $SED "$SEDCMD"`" case "$2" in HEADER) $RPM -qip "$f" > "$3"; exit 0;; INSTALL) echo "# Run this to install this RPM package" > "$3"; exit 0;; Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Thu Aug 26 19:04:13 2004 From: pavelsh at mail.ru (pavelsh) Date: Fri, 27 Aug 2004 01:04:13 +0600 Subject: extfs rpm file name quoting In-Reply-To: <1093540321.4786.112.camel@athlon.localdomain> References: <1093540321.4786.112.camel@athlon.localdomain> Message-ID: <15272132.20040827010413@mail.ru> Hello Leonard, Thursday, August 26, 2004, 11:12:01 PM, you wrote: LdO> The sed command used in mcrpmfs_list() should probably also be used in LdO> mcrpmfs_copyout(): LdO> --- rpm.000 2004-08-26 17:44:13.000000000 +0200 LdO> +++ rpm 2004-08-26 19:02:32.000000000 +0200 LdO> @@ -133,7 +133,7 @@ LdO> mcrpmfs_copyout () LdO> { LdO> - f="\"$1\"" LdO> + f="`echo "$1" | $SED "$SEDCMD"`" LdO> case "$2" in LdO> HEADER) $RPM -qip "$f" > "$3"; exit 0;; LdO> INSTALL) echo "# Run this to install this RPM package" > "$3"; LdO> exit 0;; Fixed. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Fri Aug 27 08:49:27 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Fri, 27 Aug 2004 10:49:27 +0200 Subject: Can't open any RPM file In-Reply-To: <1093110503.4765.60.camel@athlon.localdomain> References: <200408211621.28886.dmi_a@qnx.org.ru> <200408212039.07621.dmi_a@qnx.org.ru> <1093107062.4765.55.camel@athlon.localdomain> <200408212106.25809.dmi_a@qnx.org.ru> <1093110503.4765.60.camel@athlon.localdomain> Message-ID: <1093596567.4786.13.camel@athlon.localdomain> Hi, I wrote: > On Sat, 2004-08-21 at 19:06, Dmitry Alexeyev wrote: > > [dmi at dmi dmi]$ rpm -q bash > > bash-2.05b-alt6 > > Ok. I thought maybe some issues with bash 3. Well, we figured out in an irc chat yesterday (Freenode #mc-devel) that this is probably a bash issue after all. It seems ALT uses a patched bash that adds some quoting to parameters itself. Seems like an ugly hack to me. How would it react f.e. to ls foo\ bar? No such file 'foo\ bar'? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From siver at sirius.ihep.su Fri Aug 27 15:20:37 2004 From: siver at sirius.ihep.su (Siver Andrey) Date: Fri, 27 Aug 2004 19:20:37 +0400 Subject: Question on developing Message-ID: <011101c48c49$6ab059e0$f3a1bec2@hermes> Dears mc-developers, I would like to become skilled mc-developer. But now I want to realize these features: 1) Make Alt-F1 and Alt-F2 to be as in Far program (and NC): make them to open sub-window to change current path position (for example, "to go" to another hard drive) [****]; 2) Make Ctrl-O to hide panel, but to save current state of the command line [**]; 3) Make comands history to be saved regardless of running or exiting of MC [**]; 4) Make Alt+ to spring on files which begin with (as in Far and NC) [*****]. Could you point me out the source files which should be changed (and anything about reality of that changes :)) ? Thanks a lot, Andrey S. -------------- next part -------------- An HTML attachment was scrubbed... URL: From buliabyak at gmail.com Fri Aug 27 15:49:34 2004 From: buliabyak at gmail.com (bulia byak) Date: Fri, 27 Aug 2004 12:49:34 -0300 Subject: Question on developing In-Reply-To: <011101c48c49$6ab059e0$f3a1bec2@hermes> References: <011101c48c49$6ab059e0$f3a1bec2@hermes> Message-ID: <3c78ff030408270849217dc9ca@mail.gmail.com> > 1) Make Alt-F1 and Alt-F2 to be as in Far program (and NC): make them to open sub-window to change current path position (for example, "to go" to another hard drive) [****]; There are no "hard drives" on Linux. There are places you go to often, and that's what hotlist is for (ctrl+\). I added hotkeys to this hotlist and made it appear on alt+f1/f2 in addition to ctrl+\, and this works very well for me. My hotlist contains entries for the CD mount dir, network locations, etc., in addition to my often-used project directories. I will update these (and other) patches for the current CVS as soon as I find a bit of time for this. > 2) Make Ctrl-O to hide panel, but to save current state of the command line [**]; More knowledgeable people will likely comment on this, but I think this requires embedding the shell into mc which is planned for 4.7 (according to TODO). I very much look forward to this, too. > 3) Make comands history to be saved regardless of running or exiting of MC [**]; It is saved, at least it works for me. > 4) Make Alt+ to spring on files which begin with (as in Far and NC) [*****]. This is done already, except that you have to press alt+s and then type your string without alt. I got used to this and find that it's even more convenient than the old way, if only because you can press alt+s while typing again and this will take you to the next match, and because typing without having to hold alt is easier on my fingers. Also this frees a whole lot of other alt+letter keys that can be put to better use :) From buliabyak at gmail.com Fri Aug 27 16:03:09 2004 From: buliabyak at gmail.com (bulia byak) Date: Fri, 27 Aug 2004 13:03:09 -0300 Subject: Question on developing In-Reply-To: <3c78ff030408270849217dc9ca@mail.gmail.com> References: <011101c48c49$6ab059e0$f3a1bec2@hermes> <3c78ff030408270849217dc9ca@mail.gmail.com> Message-ID: <3c78ff0304082709032e5dfcea@mail.gmail.com> By the way, these and other notable differences between mc and FAR/NC should be placed in the FAQ, to help new users coming from Windows. On Fri, 27 Aug 2004 12:49:34 -0300, bulia byak wrote: > > 1) Make Alt-F1 and Alt-F2 to be as in Far program (and NC): make them to open sub-window to change current path position (for example, "to go" to another hard drive) [****]; > > There are no "hard drives" on Linux. There are places you go to often, > and that's what hotlist is for (ctrl+\). I added hotkeys to this > hotlist and made it appear on alt+f1/f2 in addition to ctrl+\, and > this works very well for me. My hotlist contains entries for the CD > mount dir, network locations, etc., in addition to my often-used > project directories. I will update these (and other) patches for the > current CVS as soon as I find a bit of time for this. > > > 2) Make Ctrl-O to hide panel, but to save current state of the command line [**]; > > More knowledgeable people will likely comment on this, but I think > this requires embedding the shell into mc which is planned for 4.7 > (according to TODO). I very much look forward to this, too. > > > 3) Make comands history to be saved regardless of running or exiting of MC [**]; > > It is saved, at least it works for me. > > > 4) Make Alt+ to spring on files which begin with (as in Far and NC) [*****]. > > This is done already, except that you have to press alt+s and then > type your string without alt. I got used to this and find that it's > even more convenient than the old way, if only because you can press > alt+s while typing again and this will take you to the next match, and > because typing without having to hold alt is easier on my fingers. > Also this frees a whole lot of other alt+letter keys that can be put > to better use :) > From makovick at kmlinux.fjfi.cvut.cz Sat Aug 28 15:45:43 2004 From: makovick at kmlinux.fjfi.cvut.cz (Jindrich Makovicka) Date: Sat, 28 Aug 2004 17:45:43 +0200 Subject: Question on developing In-Reply-To: <3c78ff0304082709032e5dfcea@mail.gmail.com> References: <011101c48c49$6ab059e0$f3a1bec2@hermes> <3c78ff030408270849217dc9ca@mail.gmail.com> <3c78ff0304082709032e5dfcea@mail.gmail.com> Message-ID: <4130A8A7.6090807@kmlinux.fjfi.cvut.cz> bulia byak wrote: >>>3) Make comands history to be saved regardless of running or exiting of MC [**]; >> >>It is saved, at least it works for me. Well, sometimes it is saved, sometimes not at all. It is an issue with more instances of mc running - when exiting, mc just writes its own updated history to the configuration file, overwriting the previous contents, which can contain another new entries written by another instance of mc. I think mc should just append the new entries instead of overwriting whole history. Another approach would be saving the commands immediately into the history file and the instances would check the history modification time and updating their own history when necessary (Alt-H, Alt-P, Alt-N etc). It would probably need some locking of the history file to avoid race conditions. PS: Why there's not a Reply-to header set to the mailing list? -- Jindrich Makovicka From roland.illig at gmx.de Sun Aug 29 15:14:01 2004 From: roland.illig at gmx.de (Roland Illig) Date: Sun, 29 Aug 2004 17:14:01 +0200 Subject: Committing trivial changes Message-ID: <4131F2B9.20608@gmx.de> Hi, I'd like to commit some changes that convert many declarations of "char *" into "const char *". The latter is nicer because you can assume that the declared string will not be accidentally modified. This is especially important when passing string literals to functions or storing them in variables, as they may be placed in read-only memory. To distinguish read-write usage from read-only usage, strings which are not written to should be declared as "const char *". Lately, Pavel banned me from doing these changes because mc-4.6.1 will be released soon, but I would nevertheless like to commit these, as they improve the code quality and also help me find memory leaks. The changes will only affect the declarations of functions, not the code itself, as far as I know the C language. (And I know it quite well.) During this phase, I will not commit any changes to the code. May I? Roland From pavelsh at mail.ru Sun Aug 29 15:31:00 2004 From: pavelsh at mail.ru (Pavel S. Shirshov) Date: Sun, 29 Aug 2004 21:31:00 +0600 Subject: Committing trivial changes In-Reply-To: <4131F2B9.20608@gmx.de> References: <4131F2B9.20608@gmx.de> Message-ID: <864321660.20040829213100@mail.ru> Hello Roland, Sunday, August 29, 2004, 9:14:01 PM, you wrote: RI> May I? Ok. -- Best regards, Pavel mailto:pavelsh at mail.ru From siver at sirius.ihep.su Mon Aug 30 09:23:39 2004 From: siver at sirius.ihep.su (Siver Andrey) Date: Mon, 30 Aug 2004 13:23:39 +0400 Subject: Question on developing References: <011101c48c49$6ab059e0$f3a1bec2@hermes> <3c78ff030408270849217dc9ca@mail.gmail.com> Message-ID: <007f01c48e73$0bcaa670$f3a1bec2@hermes> Dear "bulia byak", Thank You for Your answer. ----- Original Message ----- From: "bulia byak" To: "Siver Andrey" Cc: Sent: Friday, August 27, 2004 7:49 PM Subject: Re: Question on developing > > 1) Make Alt-F1 and Alt-F2 to be as in Far program (and NC): make them to open sub-window to change current path position (for example, "to go" to another hard drive) [****]; > > There are no "hard drives" on Linux. What about /dev/hda, /dev/hdb. If they are in the power-set then they are on Linux. But as for me I would prefer to call them "A1", "B1" and so on :-). > There are places you go to often, > and that's what hotlist is for (ctrl+\). I added hotkeys to this > ... Very good! > > > 2) Make Ctrl-O to hide panel, but to save current state of the command line [**]; > > More knowledgeable people will likely comment on this, but I think > this requires embedding the shell into mc which is planned for 4.7 > (according to TODO). I very much look forward to this, too. According to my mind MC should not be the shell (rather to be a shell of the Shell). > > > 3) Make comands history to be saved regardless of running or exiting of MC [**]; > > It is saved, at least it works for me. See email: (From: "Jindrich Makovicka" To: "bulia byak" Cc: "Siver Andrey" ; Sent: Saturday, August 28, 2004 7:45 PM Subject: Re: Question on developing) > > 4) Make Alt+ to spring on files which begin with (as in Far and NC) [*****]. > > This is done already, except that you have to press alt+s and then Great! This feature I'm looking for 2-3 years. I have more questions. One of them: What about the message "Commands interpreter is occupied by another command" when you try to run program (russian translaters found very funny translation of this message). This message occurs very often. Standart solution Ctlr-O Ctrl-O does not always work. Another one. Why does MC forget about sizes of the directories (after the command "show directory size")? And last one to this email. Why does neither Ctrl-Ins (Shift-Ins) nor Ctrl-C(Ctrl-V) work properly in the editor? Regards, Andrey From leonard at den.ottolander.nl Mon Aug 30 09:46:51 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Mon, 30 Aug 2004 11:46:51 +0200 Subject: Committing trivial changes In-Reply-To: <4131F2B9.20608@gmx.de> References: <4131F2B9.20608@gmx.de> Message-ID: <1093859210.4786.42.camel@athlon.localdomain> Hi Roland, On Sun, 2004-08-29 at 17:14, Roland Illig wrote: > I'd like to commit some changes that convert many declarations of "char > *" into "const char *". > Lately, Pavel banned me from doing these changes because mc-4.6.1 will > be released soon, but I would nevertheless like to commit these, as they > improve the code quality and also help me find memory leaks. Although I do agree that we might want to focus on the TODO for 4.6.1 and wait with work on new features I also agree that these proposed changes are so basic and unintrusive that I can't see much harm in implementing them before 4.6.1. Make sure the headers and code match though, and don't touch functions that should actually modify the string ;) . Leonard. -- mount -t life -o ro /dev/dna /genetic/research From andrew at email.zp.ua Mon Aug 30 10:02:48 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Mon, 30 Aug 2004 13:02:48 +0300 (EEST) Subject: Committing trivial changes In-Reply-To: <4131F2B9.20608@gmx.de> Message-ID: <200408301002.i7UA2mCV062196@email.zp.ua> Hi, Roland, your trivial changes break compilation with --enable-charset and produce a lot of warnings about slang functions. Patch attached. -- Regards, Andrew V. Samoilov. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: charsets.patch URL: From roland.illig at gmx.de Mon Aug 30 10:04:42 2004 From: roland.illig at gmx.de (Roland Illig) Date: Mon, 30 Aug 2004 12:04:42 +0200 Subject: Committing trivial changes In-Reply-To: <1093859210.4786.42.camel@athlon.localdomain> References: <4131F2B9.20608@gmx.de> <1093859210.4786.42.camel@athlon.localdomain> Message-ID: <4132FBBA.60002@gmx.de> Leonard den Ottolander wrote: > [...] Make sure the headers and code match > though, and don't touch functions that should actually modify the string > ;) . Well said, during my patch work I really noticed some arguments that I had declared "const" first but which were really freed. Currently the result looks quite good. I've gone through many files in edit and src that don't require functional changes. I will continue when I'm back from holidays. Roland From roland.illig at gmx.de Mon Aug 30 10:38:13 2004 From: roland.illig at gmx.de (Roland Illig) Date: Mon, 30 Aug 2004 12:38:13 +0200 Subject: Committing trivial changes In-Reply-To: <200408301002.i7UA2mCV062196@email.zp.ua> References: <200408301002.i7UA2mCV062196@email.zp.ua> Message-ID: <41330395.6010706@gmx.de> Andrew V. Samoilov wrote: > Hi, Roland, > > your trivial changes break compilation with --enable-charset and produce a lot of warnings about slang functions. > > Patch attached. Patch applied. Thanks for reporting. I had forgotten to --enable-charset. I hope I didn't forget too much else. Roland From andrew at email.zp.ua Mon Aug 30 13:09:00 2004 From: andrew at email.zp.ua (Andrew V. Samoilov) Date: Mon, 30 Aug 2004 16:09:00 +0300 (EEST) Subject: Bug#248646: mc: ftpfs can't show file size larger than 2 Gigabytes In-Reply-To: <200407211842.i6LIgR2c099079@email.zp.ua> Message-ID: <200408301309.i7UD90LU067575@email.zp.ua> Hi, I hope this patch incarnation is ready for production. -- Regards, Andrew V. Samoilov. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: vfs.atoll.patch URL: From pavelsh at mail.ru Mon Aug 30 19:06:09 2004 From: pavelsh at mail.ru (pavelsh) Date: Tue, 31 Aug 2004 01:06:09 +0600 Subject: Bug#248646: mc: ftpfs can't show file size larger than 2 Gigabytes In-Reply-To: <200408301309.i7UD90LU067575@email.zp.ua> References: <200407211842.i6LIgR2c099079@email.zp.ua> <200408301309.i7UD90LU067575@email.zp.ua> Message-ID: <1964742099.20040831010609@mail.ru> Hello Andrew, Monday, August 30, 2004, 7:09:00 PM, you wrote: AVS> Hi, AVS> I hope this patch incarnation is ready for production. Committed. Thx. -- Best regards, pavelsh mailto:pavelsh at mail.ru From pavelsh at mail.ru Mon Aug 30 20:50:57 2004 From: pavelsh at mail.ru (pavelsh) Date: Tue, 31 Aug 2004 02:50:57 +0600 Subject: Current code snapshot Message-ID: <1434890708.20040831025057@mail.ru> Hello, Please, check new place for code snapshots. http://pavelsh.pp.ru/mc/ P.S.: Current code snapshot without changes in slang from Roland. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Mon Aug 30 21:22:26 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Mon, 30 Aug 2004 23:22:26 +0200 Subject: Current code snapshot In-Reply-To: <1434890708.20040831025057@mail.ru> References: <1434890708.20040831025057@mail.ru> Message-ID: <1093900945.4786.48.camel@athlon.localdomain> Hi Pavel, Roland, On Mon, 2004-08-30 at 22:50, pavelsh wrote: > P.S.: Current code snapshot without changes in slang from Roland. Which changes are those exactly? Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Mon Aug 30 21:32:51 2004 From: pavelsh at mail.ru (pavelsh) Date: Tue, 31 Aug 2004 03:32:51 +0600 Subject: Current code snapshot In-Reply-To: <1093900945.4786.48.camel@athlon.localdomain> References: <1434890708.20040831025057@mail.ru> <1093900945.4786.48.camel@athlon.localdomain> Message-ID: <669048390.20040831033251@mail.ru> Hello Leonard, Tuesday, August 31, 2004, 3:22:26 AM, you wrote: LdO> Hi Pavel, Roland, LdO> On Mon, 2004-08-30 at 22:50, pavelsh wrote: >> P.S.: Current code snapshot without changes in slang from Roland. LdO> Which changes are those exactly? 2004-08-29 Roland Illig * slang.h: Added const qualifier to some of the SLang functions. -- Best regards, pavelsh mailto:pavelsh at mail.ru From proski at gnu.org Mon Aug 30 21:28:54 2004 From: proski at gnu.org (Pavel Roskin) Date: Mon, 30 Aug 2004 17:28:54 -0400 (EDT) Subject: Current code snapshot In-Reply-To: <1434890708.20040831025057@mail.ru> References: <1434890708.20040831025057@mail.ru> Message-ID: On Tue, 31 Aug 2004, pavelsh wrote: > Hello, > > Please, check new place for code snapshots. > > http://pavelsh.pp.ru/mc/ > > P.S.: Current code snapshot without changes in slang from Roland. Sorry that I stopped making snapshots for some time. I've just made a snapshot from CVS as is, without excluding any changes and put it to the usual place: http://www.ibiblio.org/pub/Linux/utils/file/managers/mc/snapshots/ Unfortunately, there are massive warnings even in the default build on Fedora Core 2 (I didn't even try mctest). The worst thing, the warnings are spread throughout the code. The source was warning free just recently. I'm afraid I entrusted the write access to CVS to wrong people. Please fix the warnings within one week or I'll revoke write permissions for those who introduced them. ftpfs.c: In function `ftpfs_netrc_next': ftpfs.c:1701: warning: assignment discards qualifiers from pointer target type sldisply.c: In function `SLtt_set_color': sldisply.c:1206: warning: passing arg 1 of `make_color_fgbg' discards qualifiers from pointer target type sldisply.c:1206: warning: passing arg 2 of `make_color_fgbg' discards qualifiers from pointer target type sldisply.c: In function `SLtt_tgetstr': sldisply.c:2035: warning: passing arg 2 of `_SLtt_tigetstr' discards qualifiers from pointer target type sldisply.c: In function `SLtt_tgetnum': sldisply.c:2055: warning: passing arg 2 of `_SLtt_tigetnum' discards qualifiers from pointer target type sldisply.c: In function `SLtt_tgetflag': sldisply.c:2066: warning: passing arg 2 of `_SLtt_tigetflag' discards qualifiers from pointer target type slerr.c: In function `SLang_exit_error': slerr.c:149: warning: passing arg 1 of pointer to function discards qualifiers from pointer target type slsmg.c: In function `SLsmg_write_nchars': slsmg.c:323: warning: assignment discards qualifiers from pointer target type slsmg.c: In function `SLsmg_printf': slsmg.c:1200: warning: assignment discards qualifiers from pointer target type sltermin.c: In function `SLtt_tigetstr': sltermin.c:1154: warning: passing arg 2 of `_SLtt_tigetstr' discards qualifiers from pointer target type editcmd.c: In function `menu_save_mode_cmd': editcmd.c:391: warning: initialization from incompatible pointer type editwidget.c: In function `edit_adjust_size': editwidget.c:130: warning: passing arg 2 of `find_widget_type' from incompatible pointer type editwidget.c: In function `edit_dialog_callback': editwidget.c:154: warning: passing arg 2 of `find_widget_type' from incompatible pointer type help.c: In function `help_handle_key': help.c:612: warning: passing arg 3 of `check_movement_keys' discards qualifiers from pointer target type help.c: In function `interactive_display': help.c:778: warning: passing arg 2 of `search_string' discards qualifiers from pointer target type menu.c: In function `destroy_menu': menu.c:529: warning: passing arg 1 of `g_free' discards qualifiers from pointer target type menu.c:530: warning: passing arg 1 of `g_free' discards qualifiers from pointer target type popt.c: In function `invokeCallbacks': popt.c:55: warning: passing arg 5 of pointer to function discards qualifiers from pointer target type popt.c: In function `findOption': popt.c:254: warning: assignment discards qualifiers from pointer target type popt.c:275: warning: assignment discards qualifiers from pointer target type profile.c: In function `GetSetProfileChar': profile.c:319: warning: return discards qualifiers from pointer target type view.c: In function `view_adjust_size': view.c:2607: warning: passing arg 2 of `find_widget_type' from incompatible pointer type widget.c: In function `find_buttonbar': widget.c:2322: warning: passing arg 2 of `find_widget_type' from incompatible pointer type wtools.c: In function `fg_input_dialog_help': wtools.c:503: warning: passing arg 1 of `g_free' discards qualifiers from pointer target type -- Regards, Pavel Roskin From pavelsh at mail.ru Mon Aug 30 22:09:03 2004 From: pavelsh at mail.ru (pavelsh) Date: Tue, 31 Aug 2004 04:09:03 +0600 Subject: Current code snapshot In-Reply-To: References: <1434890708.20040831025057@mail.ru> Message-ID: <1301787532.20040831040903@mail.ru> Hello Pavel, Tuesday, August 31, 2004, 3:28:54 AM, you wrote: PR> On Tue, 31 Aug 2004, pavelsh wrote: >> Hello, >> >> Please, check new place for code snapshots. >> >> http://pavelsh.pp.ru/mc/ >> >> P.S.: Current code snapshot without changes in slang from Roland. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Pavel, look at Problem is conflict among system slang.h and mc/slang/*.c sources. These errors is not present in my version of snapshot. Pavel, please, give me your scripts for making snapshots. PR> Sorry that I stopped making snapshots for some time. I've just made a PR> snapshot from CVS as is, without excluding any changes and put it to the PR> usual place: PR> http://www.ibiblio.org/pub/Linux/utils/file/managers/mc/snapshots/ Your version of snapshots has a problems also. :) .......................................................................................................... done. rm -f it.gmo && /usr/local/bin/msgfmt -c --statistics -o it.gmo it.po 942 translated messages, 10 fuzzy translations, 7 untranslated messages. /usr/local/bin/msgmerge --update ja.po mc.pot .......................................................................................................... done. rm -f ja.gmo && /usr/local/bin/msgfmt -c --statistics -o ja.gmo ja.po 906 translated messages, 29 fuzzy translations, 24 untranslated messages. /usr/local/bin/msgmerge --update ko.po mc.pot .......................................................................................................... done. rm -f ko.gmo && /usr/local/bin/msgfmt -c --statistics -o ko.gmo ko.po 929 translated messages, 10 fuzzy translations, 20 untranslated messages. /usr/local/bin/msgmerge --update lt.po mc.pot /usr/local/bin/msgmerge: error while opening "lt.po" for reading: No such file or directory *** Error code 1 Stop in /home/pavelsh/test/mc-4.6.0/po. *** Error code 1 Stop in /home/pavelsh/test/mc-4.6.0. *** Error code 1 Stop in /home/pavelsh/test/mc-4.6.0. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Mon Aug 30 22:16:20 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 31 Aug 2004 00:16:20 +0200 Subject: Current code snapshot In-Reply-To: References: <1434890708.20040831025057@mail.ru> Message-ID: <1093904180.4786.100.camel@athlon.localdomain> Hello Pavel, On Mon, 2004-08-30 at 23:28, Pavel Roskin wrote: > Unfortunately, there are massive warnings even in the default build on > Fedora Core 2 (I didn't even try mctest). The worst thing, the warnings > are spread throughout the code. The source was warning free just > recently. I'm afraid I entrusted the write access to CVS to wrong people. That probably is singular. You might want to contact the culprit directly as per his private email address. I don't like finger pointing, but to avoid confusion and false accusations to people not responsible, these changes have been introduced by Roland. He should be much more careful, and probably not commit changes just before going on holiday (and without testing them properly). Pavel, now that you are around, would you mind giving some answers to the questions raised in the "Maintenance question" thread that you yourself started a couple of weeks ago? TIA. Leonard. -- mount -t life -o ro /dev/dna /genetic/research From dmi_a at qnx.org.ru Mon Aug 30 22:43:18 2004 From: dmi_a at qnx.org.ru (Dmitry Alexeyev) Date: Tue, 31 Aug 2004 02:43:18 +0400 Subject: autogen.sh problem Message-ID: <200408310243.18186.dmi_a@qnx.org.ru> Hi, Just checked out mc from CVS and tried to run autogen.sh It fails giving this message: configure: creating ./config.status config.status: creating Makefile config.status: creating mc.spec config.status: creating doc/Makefile config.status: creating vfs/Makefile config.status: creating vfs/extfs/Makefile config.status: error: cannot find input file: vfs/extfs/Makefile.in ^^^ WBR Dmitry From proski at gnu.org Tue Aug 31 03:01:50 2004 From: proski at gnu.org (Pavel Roskin) Date: Mon, 30 Aug 2004 23:01:50 -0400 (EDT) Subject: Current code snapshot In-Reply-To: <1301787532.20040831040903@mail.ru> References: <1434890708.20040831025057@mail.ru> <1301787532.20040831040903@mail.ru> Message-ID: On Tue, 31 Aug 2004, pavelsh wrote: >>> P.S.: Current code snapshot without changes in slang from Roland. > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > Pavel, look at > Problem is conflict among system slang.h and mc/slang/*.c > sources. These errors is not present in my version of snapshot. I believe the snapshot by definition is what is in CVS and a certain point of time without changes (other than version number). > Pavel, please, give me your scripts for making snapshots. I don't keep any private scripts. I'm using maint/mcsnap from CVS. I recommend that all scripts for maintainers are kept in CVS under maint. > PR> Sorry that I stopped making snapshots for some time. I've just made a > PR> snapshot from CVS as is, without excluding any changes and put it to the > PR> usual place: > PR> http://www.ibiblio.org/pub/Linux/utils/file/managers/mc/snapshots/ > > Your version of snapshots has a problems also. :) I don't see anything funny in that. I trusted you to make changes, but it looks like we are further from the release than we used to be. I'll be forced to remove write permission from everybody who breaks the project and finds it funny. Write access won't be restored until the regressions are fixed. > .......................................................................................................... done. > rm -f it.gmo && /usr/local/bin/msgfmt -c --statistics -o it.gmo it.po > 942 translated messages, 10 fuzzy translations, 7 untranslated messages. > /usr/local/bin/msgmerge --update ja.po mc.pot > .......................................................................................................... done. > rm -f ja.gmo && /usr/local/bin/msgfmt -c --statistics -o ja.gmo ja.po > 906 translated messages, 29 fuzzy translations, 24 untranslated messages. > /usr/local/bin/msgmerge --update ko.po mc.pot > .......................................................................................................... done. > rm -f ko.gmo && /usr/local/bin/msgfmt -c --statistics -o ko.gmo ko.po > 929 translated messages, 10 fuzzy translations, 20 untranslated messages. > /usr/local/bin/msgmerge --update lt.po mc.pot > /usr/local/bin/msgmerge: error while opening "lt.po" for reading: No such file or directory You didn't provide any details how to reproduce the problem. mcsnap would not upload the snapshot without successful "make distcheck". This measure is intentional to prevent upload of snapshots that don't compile. -- Regards, Pavel Roskin From pavelsh at mail.ru Tue Aug 31 09:11:40 2004 From: pavelsh at mail.ru (pavelsh) Date: Tue, 31 Aug 2004 15:11:40 +0600 Subject: Current code snapshot In-Reply-To: References: <1434890708.20040831025057@mail.ru> Message-ID: <55345565.20040831151140@mail.ru> Hello Pavel, Tuesday, August 31, 2004, 3:28:54 AM, you wrote: PR> ftpfs.c: In function `ftpfs_netrc_next': PR> ftpfs.c:1701: warning: assignment discards qualifiers from pointer target type PR> sldisply.c: In function `SLtt_set_color': PR> sldisply.c:1206: warning: passing arg 1 of `make_color_fgbg' Fixed. -- Best regards, pavelsh mailto:pavelsh at mail.ru From leonard at den.ottolander.nl Tue Aug 31 09:28:29 2004 From: leonard at den.ottolander.nl (Leonard den Ottolander) Date: Tue, 31 Aug 2004 11:28:29 +0200 Subject: Current code snapshot In-Reply-To: References: <1434890708.20040831025057@mail.ru> <1301787532.20040831040903@mail.ru> Message-ID: <1093944509.4786.10.camel@athlon.localdomain> Hello Pavel, On Tue, 2004-08-31 at 05:01, Pavel Roskin wrote: > > Your version of snapshots has a problems also. :) > > I don't see anything funny in that. I trusted you to make changes, but it > looks like we are further from the release than we used to be. I'll be > forced to remove write permission from everybody who breaks the project > and finds it funny. Well, luckily neither Roland nor Pavel are guilty of both. Ease up a bit, will you? If you'd read my mail on this subject you would know these changes are not introduced by Pavel. If you had read Pavel's mail more carefully you would have understood that this might be an issue with slang.h that is not as serious as it looks at first sight (well, that is what I understand from Pavel, so I might be wrong). Also, for years Red Hat/Fedora use patches to change the build process with regard to slang. I haven't discussed this with Jakub Jelinek (and Vladimir Nadvornik for that matter) yet, but it might be that these patches (specifically the mc-4.6.0-slang.patch which you can find inside the Fedora mc src rpms) are still required to build on Fedora, even without the utf-8 patches for mc. Regards, Leonard. -- mount -t life -o ro /dev/dna /genetic/research From pavelsh at mail.ru Tue Aug 31 22:53:54 2004 From: pavelsh at mail.ru (pavelsh) Date: Wed, 1 Sep 2004 04:53:54 +0600 Subject: Check patch for a separate color for hotkeys in red dialogs Message-ID: <569391063.20040901045354@mail.ru> Hello mc-devel, Check patch. It's make a separate color for hotkeys in red dialogs. -- Best regards, pavelsh mailto:pavelsh at mail.ru -------------- next part -------------- A non-text attachment was scrubbed... Name: fix-error-dialog.patch Type: application/octet-stream Size: 2273 bytes Desc: not available URL: