1. 31 Aug, 2013 6 commits
  2. 01 Apr, 2012 8 commits
  3. 10 Dec, 2011 6 commits
  4. 04 Oct, 2011 5 commits
  5. 04 Aug, 2011 6 commits
    • Markus Steinborn's avatar
      NEWS file updated · ddf1fa9e
      Markus Steinborn authored
    • Markus Steinborn's avatar
      Updated changelog · 819601e0
      Markus Steinborn authored
    • Markus Steinborn's avatar
      Fix usage of sec_sscanf · a5405341
      Markus Steinborn authored
      This should fix http://bugs.debian.org/627471
      Bernhard R. Link finally found the bug (excellent job!). He
      describes the bug as follows:
      ps.c is using some sec_sscanf (from secscanf.c) instead of
      regular sscanf or instead of doing some proper parsing.
      As sec_sscanf differs from regular sscanf about it variadic
      arguments gcc cannot test if the arguments given match the
      format string, especially it is lost about sec_sscanf
      wanting a 'char *' and a 'size_t' for ever '%s' or '%256s'
      it gets. Thus when ps.c does
      sec_sscanf(line+lenght("%%BoundingBox:), "%256s", text);
      the size of text field is not given, so some random value
      is returned by the 'va_arg(ap, size_t)' in secscanf.c
      If that random value is smaller than the length of "(atend)"
      then this will be copied incompletely and thus
      not be recognized.
    • Bernhard Link's avatar
      do not define errno, as it might be a macro already. · 0ecbe97f
      Bernhard Link authored
      Errno will usually be a macro nowadays (to make use of thread local
      storage), so defining it as
      extern int errno;
      is made by the preprocessor to:
      extern int (*__errno_location ());
      which luckily currently still works, but shows that is really not
      what should be done, so remove the definitions. (Which also silences
      a compiler warning).
    • Bernhard Link's avatar
      make numpages and nummedia signed to avoid warnings · 39b76a7b
      Bernhard Link authored
      As numpages and nummedia are usually compared to int
      and assigned to them, the bigger unsigned values cannot
      be used anway, so make them int to avoid all kind of warnings.
      Also remove some casts no longer necessary due to that.
    • Bernhard Link's avatar
      get rid of core K&R, use gnulib's _GL_UNUSED and avoid some signed/unsigned comparisons · 0b7b4993
      Bernhard Link authored
      This significantly reduces the number of warnings if requesting a
      slightly bigger set of warnings.
  6. 01 Aug, 2011 5 commits
    • Bernhard Link's avatar
    • Bernhard Link's avatar
      don't install gv_spartan.dat and gv_widgetless.dat with fallback styles · 6dcafe6c
      Bernhard Link authored
      * Markus Steinborn <gnugv_maintainer@yahoo.de> [110506 13:28]:
      > Hello Bernhard,
      > I've applied the other three patches.
      > Bernhard R. Link schrieb:
      >> With fallback styles enabled (currently default), gv_spartan.dat and
      >> gv_widgetless.dat are always replaced with compiled in versions,
      >> so do not install those files to not confuse users with installing
      >> files that have no effect if changed.
      >> ---
      > On my system, GNU gv 3.7.2 installs the following files within
      > /usr/share/gv/
      > /usr/share/gv/GV
      that is stored in appdefaultsdir (which by default might be the same,
      but can thus be installed in a way that it is used.
      > /usr/share/gv/gv_class.ad
      > /usr/share/gv/gv_copyright.dat
      > /usr/share/gv/gv_spartan.dat
      > /usr/share/gv/gv_system.ad
      > /usr/share/gv/gv_user.ad
      > /usr/share/gv/gv_user_res.dat
      > /usr/share/gv/gv_widgetless.dat
      > Except "/usr/share/gv/gv_system.ad" none is read iff fallback styles are
      > enabled. I do not see any reason to handle gv_spartan.dat and
      > gv_widgetless.dat different from the raiming gv_*-files (except
      > gv_system.ad).
      I did not realize that. I only saw that those two are used with fallback
      styles and not without. What about the following then?
      >From 6540ff7072006cc84768704c1e0ceef80544be13 Mon Sep 17 00:00:00 2001
      From: Bernhard R. Link <brlink@debian.org>
      Date: Thu, 28 Apr 2011 16:24:16 +0200
      Subject: [PATCH] don't install .ad files not used
      Only install files gv actually reads. Having other files around
      is confusing as changing them has no effect.
      With fallback styles enabled (currently default), gv_spartan.dat and
      gv_widgetless.dat are always replaced with compiled in versions,
      so only install them if fallback styles are disabled.
    • Bernhard Link's avatar
      do not replace ~ a shell would not · 6860bdce
      Bernhard Link authored
      file_translateTildeInPath currently translates the first ~ it finds.
      This is different from shell which only translates it if it is the
      first character of a word and not if other letters follow.
      As file_translateTildeInPath is always called with a single filename,
      only replace it if this string is "~" or starts with "~/".
      Additionally add an size argument to the function, so that no string
      buffers can overrun and do not access a NULL pointer if HOME is not set.
    • Bernhard Link's avatar
      ignore resource and style files without absolute paths · 4eeca7eb
      Bernhard Link authored
      The current search path is current directory, home directory,
      and then GV_LIBDIR (but no /) prepended.
      As GV_LIBDIR without the slash cannot have worked for anyone
      yet and looking into the current directory is quite some security
      risk (in case anyone actually used that option), reject filenames
      not being absolute.
    • Markus Steinborn's avatar
      Update gnulib, import signal · 0a1be09f
      Markus Steinborn authored
  7. 04 May, 2011 4 commits