Description of problem: With perl-Tk installed and assuming that texdoctk started (see below) various documents cannot be found. It turns out that texdoctk is using /usr/share/texmf/texdoctk/texdoctk.dat (and optionally also $TEXMFLOCAL/texdoctk/texdoctk-local.dat and $TEXMFHOME/texdoctk/texdoctk-pers.dat) but supplied texdoctk.dat is obsolete. It includes entries for files like fontinstallationguide.pdf, dtxtut.pdf, ... not to be found anywhere on the disk. OTOH files which are present, say intro98.pdf from fontinst documentation, and not listed. All of that assuming that texdoctk can be even started. Bug 210718 is still there and it seems that it is getting worse over time. I can still start that on a local display of my system, although even that is not assured anymore, but touching "Search" or "Settings" looks like a guaranteed recipe for an instant segfault. It also looks like that current texdoctk ignores private configuration in an existing ~/.texdocrc; but it is hard to be sure what is really happening due to problems mentioned in the previous paragraph. A quick automated check seems to indicate that the following referenced in texdoctk.dat files are missing. Results may be not entirely correct or complete but mostly should be ok. Some files may show up after an installation of additional packages. Version-Release number of selected component (if applicable): texlive-texmf-2007-2.fc9 How reproducible: always
Created attachment 282121 [details] a list of files not found by a simple check
$ texdoctk NOTE: Button colours not set in /usr/share/texmf/texdoctk/texdocrc.defaults; using defaults. Segmentation fault ... seems that it really needs fixing ...
Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 46912496263824 (LWP 24522)] 0x0000003285e16c7a in FcCompare (pat=0xcd8050, fnt=0x2aaaafe700c0, value=0xd06c40, result=0x7fffada3baf8) at fcmatch.c:392 392 while (i1 < pat->num && i2 < fnt->num) (gdb) bt #0 0x0000003285e16c7a in FcCompare (pat=0xcd8050, fnt=0x2aaaafe700c0, value=0xd06c40, result=0x7fffada3baf8) at fcmatch.c:392 #1 0x0000003285e170c4 in IA__FcFontSetSort (config=<value optimized out>, sets=<value optimized out>, nsets=1, p=0xcd8050, trim=1, csp=0x7fffada3bae0, result=0x7fffada3baf8) at fcmatch.c:708 #2 0x0000003285e175d6 in IA__FcFontSort (config=0xcd8050, p=0xcd8050, trim=1, csp=0x7fffada3bae0, result=0x7fffada3baf8) at fcmatch.c:836 #3 0x00002aaaafc14ca2 in ?? () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #4 0x00002aaaafc150f6 in TkpGetFontFromAttributes () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #5 0x00002aaaafbd8379 in Tk_AllocFontFromObj () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #6 0x00002aaaafc20230 in ?? () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #7 0x00002aaaafc206bb in Tk_InitOptions () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #8 0x00002aaaafbcf9b3 in ?? () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #9 0x00002aaaafbb0bbb in Call_Tk () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #10 0x00002aaaafbb1439 in XSTkCommand () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #11 0x00002aaaafbb156f in ?? () from /usr/lib64/perl5/vendor_perl/5.8.8/x86_64-linux-thread-multi/auto/Tk/Tk.so #12 0x00000039cae90866 in Perl_pp_entersub () from /usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi/CORE/libperl.so #13 0x00000039cae8a10e in Perl_runops_standard () from /usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi/CORE/libperl.so #14 0x00000039cae37ddc in perl_run () from /usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi/CORE/libperl.so #15 0x00000000004017cc in main () hmmm...
> hmmm... You are hitting precisely bug 210718 (reported over a year ago). Quite a bit of information there. I would love to know what comments in upstream NOTOURBUG claiming https://bugs.freedesktop.org/show_bug.cgi?id=8665 https://bugs.freedesktop.org/show_bug.cgi?id=8791 are trying to say but I failed to get anything specific/informative. BTW - I strongly suspect that bug 210718 is also an underlying cause of bug 375131 which kills xfs if you will look at it wrong way. Fun, eh?
It looks like that an update to perl-Tk-804.028-1 (see bug 210718) brings a long awaited fix to issues noted in comment 2 and comment 3.
Yes, texdoctk now works as expected, closing RAWHIDE.