thr3ads.net - Fontconfig - [Fontconfig] confused about semantics of $HOME/.fonts.cache-2 [Jan 2006]

If this information is useful, please help other people find it:
Share via:

Matthias Clasen

2006-Jan-25 09:56 UTC

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

Can somebody update me about the intended semantics of
$HOME/.fonts.cache-2. Why do we need it, when there are caches for
all font directories in /var/cache/fontconfig and in $HOME/.fonts ?

I am also puzzled by this behaviour:

$ rm $HOME/.fonts.cache-2
$ fc-list :

[long font listing]

$ ls -l .fonts.cache-2
-rw------- 1 mclasen mclasen 141434 Jan 25 12:57 .fonts.cache-2

$ fc-cat .fonts.cache-2 | wc -l
144

$ touch /usr/share/fonts/bitstream-vera
$ fc-list :

[ long font listing]

$ ls -l .fonts.cache-2
-rw------- 1 mclasen mclasen 32768 Jan 25 12:58 .fonts.cache-2

$ fc-cat .fonts.cache-2 | wc -l
0


What is going on here ?


Matthias

Frederic Crozat

2006-Jan-25 10:01 UTC

head link

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

Le mercredi 25 janvier 2006 ? 12:59 -0500, Matthias Clasen a ?crit
:> Can somebody update me about the intended semantics of
> $HOME/.fonts.cache-2. Why do we need it, when there are caches for
> all font directories in /var/cache/fontconfig and in $HOME/.fonts ?
~/.fonts.cache-2 is supposed to exist only if some informations are
missing in /var/cache/fontconfig (ie dirty cache) or if some fonts are
in ~/.fonts (cache should not be in ~/.fonts).

> I am also puzzled by this behaviour:
> 
> $ rm $HOME/.fonts.cache-2
> $ fc-list :
> 
> [long font listing]
> 
> $ ls -l .fonts.cache-2
> -rw------- 1 mclasen mclasen 141434 Jan 25 12:57 .fonts.cache-2
> 
> $ fc-cat .fonts.cache-2 | wc -l
> 144
> 
> $ touch /usr/share/fonts/bitstream-vera
> $ fc-list :
> 
> [ long font listing]
> 
> $ ls -l .fonts.cache-2
> -rw------- 1 mclasen mclasen 32768 Jan 25 12:58 .fonts.cache-2
> 
> $ fc-cat .fonts.cache-2 | wc -l
> 0
> 
> 
> What is going on here ?
I think you just hit one of the various cache bugs we are investigating
with Patrick and Mike. Try removing all files in /var/cache/fontconfig
and re-run fc-cache as root.

-- 
Frederic Crozat <fcrozat@mandriva.com>
Mandriva

Matthias Clasen

2006-Jan-25 10:08 UTC

head link

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

On Wed, 2006-01-25 at 19:04 +0100, Frederic Crozat
wrote:> Le mercredi 25 janvier 2006 ? 12:59 -0500, Matthias Clasen a ?crit :
> > Can somebody update me about the intended semantics of
> > $HOME/.fonts.cache-2. Why do we need it, when there are caches for
> > all font directories in /var/cache/fontconfig and in $HOME/.fonts ?
> 
> ~/.fonts.cache-2 is supposed to exist only if some informations are
> missing in /var/cache/fontconfig (ie dirty cache) or if some fonts are
> in ~/.fonts (cache should not be in ~/.fonts).
Hmm, I have fonts in ~/.fonts, but I get both ~/.fonts.cache-2 and 
~/.fonts/fonts.cache-2 created.

Also, whats the rationale for "should not be in ~/.fonts") ?

Matthias

Behdad Esfahbod

2006-Jan-25 10:30 UTC

head link

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

I can reproduce this consistently:

[behdad@home ~]$ rm .fonts.cache-2* -f
[behdad@home ~]$ \time fc-list | wc -l
2.49user 0.09system 0:02.59elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+29969minor)pagefaults 0swaps
198
[behdad@home ~]$ ll .fonts.cache-2*
-rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
[behdad@home ~]$ sudo touch /usr/share/fonts/bitstream-vera
[behdad@home ~]$ fc-list
Segmentation fault
[behdad@home ~]$ ll .fonts.cache-2*
-rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
-rw-------  1 behdad behdad      6 Jan 25 13:27 .fonts.cache-2.LCK
-rw-------  1 behdad behdad  14522 Jan 25 13:27 .fonts.cache-2.NEW
[behdad@home ~]$ \time fc-list | wc -l
0.03user 0.01system 0:00.05elapsed 84%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+806minor)pagefaults 0swaps
198
[behdad@home ~]$ ll .fonts.cache-2*
-rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
-rw-------  1 behdad behdad      6 Jan 25 13:27 .fonts.cache-2.LCK
-rw-------  1 behdad behdad  14522 Jan 25 13:27 .fonts.cache-2.NEW



And the backtrace:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1209092416 (LWP 28257)]
0x004d4d42 in FcGlobalCacheSave (cache=0x9b4cad0,
    cache_file=0x9b440e8 "/home/behdad/.fonts.cache-2",
config=0x9b44008)
    at fccache.c:152
152         if (write (fd, chars, strlen(chars)+1) != strlen(chars)+1)
(gdb) bt
#0  0x004d4d42 in FcGlobalCacheSave (cache=0x9b4cad0,
    cache_file=0x9b440e8 "/home/behdad/.fonts.cache-2",
config=0x9b44008)
    at fccache.c:152
#1  0x004d8910 in FcConfigBuildFonts (config=0x9b44008) at fccfg.c:328
#2  0x004e19cd in FcInitLoadConfigAndFonts () at fcinit.c:85
#3  0x004e1c66 in FcInit () at fcinit.c:103
#4  0x08048b16 in main (argc=1, argv=0xbfd03b84) at fc-list.c:119
(gdb)


(gdb) p cache->dirs[0]
$4 = {next = 0x9b46538, name = 0x9b5ba88
"/usr/share/fonts/bitstream-vera",
  metadata = {magic = -66913278, count = 10361, bank = 778904322,
    pattern_count = 10, patternelt_count = 164, valuelist_count = 164,
    str_count = 238, langset_count = 10, charset_count = 10,
    charset_numbers_count = 90, charset_leaf_count = 90,
    charset_leaf_idx_count = 10}, offset = 0, ent = 0x9b54998}


I cannot do p chars, because of inlining I guess.


--behdad
http://behdad.org/

"Commandment Three says Do Not Kill, Amendment Two says Blood Will
Spill"
	-- Dan Bern, "New American Language"

Mike FABIAN

2006-Jan-26 03:59 UTC

head link

[Fontconfig] Re: confused about semantics of $HOME/.fonts.cache-2

Behdad Esfahbod <behdad@cs.toronto.edu> ????????:
> I can reproduce this consistently:
>
> [behdad@home ~]$ rm .fonts.cache-2* -f
> [behdad@home ~]$ \time fc-list | wc -l
> 2.49user 0.09system 0:02.59elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (0major+29969minor)pagefaults 0swaps
> 198
> [behdad@home ~]$ ll .fonts.cache-2*
> -rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
> [behdad@home ~]$ sudo touch /usr/share/fonts/bitstream-vera
> [behdad@home ~]$ fc-list
> Segmentation fault
I cannot reproduce that crash with the current CVS.

(with my patch improve-normalization-of-font-path-add-subdirs.patch,
see

http://lists.freedesktop.org/archives/fontconfig/2006-January/001939.html

but I doubt this makes a difference).
> [behdad@home ~]$ ll .fonts.cache-2*
> -rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
> -rw-------  1 behdad behdad      6 Jan 25 13:27 .fonts.cache-2.LCK
> -rw-------  1 behdad behdad  14522 Jan 25 13:27 .fonts.cache-2.NEW
> [behdad@home ~]$ \time fc-list | wc -l
> 0.03user 0.01system 0:00.05elapsed 84%CPU (0avgtext+0avgdata 0maxresident)k
> 0inputs+0outputs (0major+806minor)pagefaults 0swaps
> 198
> [behdad@home ~]$ ll .fonts.cache-2*
> -rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
> -rw-------  1 behdad behdad      6 Jan 25 13:27 .fonts.cache-2.LCK
> -rw-------  1 behdad behdad  14522 Jan 25 13:27 .fonts.cache-2.NEW
>
>
>
> And the backtrace:
>
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread -1209092416 (LWP 28257)]
> 0x004d4d42 in FcGlobalCacheSave (cache=0x9b4cad0,
>     cache_file=0x9b440e8 "/home/behdad/.fonts.cache-2",
config=0x9b44008)
>     at fccache.c:152
> 152         if (write (fd, chars, strlen(chars)+1) != strlen(chars)+1)
> (gdb) bt
> #0  0x004d4d42 in FcGlobalCacheSave (cache=0x9b4cad0,
>     cache_file=0x9b440e8 "/home/behdad/.fonts.cache-2",
config=0x9b44008)
>     at fccache.c:152
> #1  0x004d8910 in FcConfigBuildFonts (config=0x9b44008) at fccfg.c:328
> #2  0x004e19cd in FcInitLoadConfigAndFonts () at fcinit.c:85
> #3  0x004e1c66 in FcInit () at fcinit.c:103
> #4  0x08048b16 in main (argc=1, argv=0xbfd03b84) at fc-list.c:119
> (gdb)
>
>
> (gdb) p cache->dirs[0]
> $4 = {next = 0x9b46538, name = 0x9b5ba88
> "/usr/share/fonts/bitstream-vera",
>   metadata = {magic = -66913278, count = 10361, bank = 778904322,
>     pattern_count = 10, patternelt_count = 164, valuelist_count = 164,
>     str_count = 238, langset_count = 10, charset_count = 10,
>     charset_numbers_count = 90, charset_leaf_count = 90,
>     charset_leaf_idx_count = 10}, offset = 0, ent = 0x9b54998}
>
>
> I cannot do p chars, because of inlining I guess.
>
>
> --behdad
> http://behdad.org/
>
> "Commandment Three says Do Not Kill, Amendment Two says Blood Will
Spill"
> 	-- Dan Bern, "New American Language"
> _______________________________________________
> Fontconfig mailing list
> Fontconfig@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/fontconfig
>
-- 
Mike FABIAN   <mfabian@suse.de>   http://www.suse.de/~mfabian
?????????????

Behdad Esfahbod

2006-Jan-26 05:38 UTC

head link

[Fontconfig] Re: confused about semantics of $HOME/.fonts.cache-2

On Thu, 26 Jan 2006, Mike FABIAN wrote:
> Behdad Esfahbod <behdad@cs.toronto.edu>
=A4=B5=A4=F3=A4=CF=BD=F1=A4=AD=A4=DE=A4=B7=A4=BF:
>
> > I can reproduce this consistently:
> >
> > [behdad@home ~]$ rm .fonts.cache-2* -f
> > [behdad@home ~]$ \time fc-list | wc -l
> > 2.49user 0.09system 0:02.59elapsed 99%CPU (0avgtext+0avgdata
0maxresident)k
> > 0inputs+0outputs (0major+29969minor)pagefaults 0swaps
> > 198
> > [behdad@home ~]$ ll .fonts.cache-2*
> > -rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
> > [behdad@home ~]$ sudo touch /usr/share/fonts/bitstream-vera
> > [behdad@home ~]$ fc-list
> > Segmentation fault
>
> I cannot reproduce that crash with the current CVS.
Note that I have installed fc-branch in my home user and not
created global -2 caches.  And I got random apps crashing
randomly all the time after yesterday''s cvs up, so had to
uninstall it.

behdad

> (with my patch improve-normalization-of-font-path-add-subdirs.patch,
> see
>
> http://lists.freedesktop.org/archives/fontconfig/2006-January/001939.html
>
> but I doubt this makes a difference).
>
> > [behdad@home ~]$ ll .fonts.cache-2*
> > -rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
> > -rw-------  1 behdad behdad      6 Jan 25 13:27 .fonts.cache-2.LCK
> > -rw-------  1 behdad behdad  14522 Jan 25 13:27 .fonts.cache-2.NEW
> > [behdad@home ~]$ \time fc-list | wc -l
> > 0.03user 0.01system 0:00.05elapsed 84%CPU (0avgtext+0avgdata
0maxresident)k
> > 0inputs+0outputs (0major+806minor)pagefaults 0swaps
> > 198
> > [behdad@home ~]$ ll .fonts.cache-2*
> > -rw-------  1 behdad behdad 286721 Jan 25 13:27 .fonts.cache-2
> > -rw-------  1 behdad behdad      6 Jan 25 13:27 .fonts.cache-2.LCK
> > -rw-------  1 behdad behdad  14522 Jan 25 13:27 .fonts.cache-2.NEW
> >
> >
> >
> > And the backtrace:
> >
> > Program received signal SIGSEGV, Segmentation fault.
> > [Switching to Thread -1209092416 (LWP 28257)]
> > 0x004d4d42 in FcGlobalCacheSave (cache=3D0x9b4cad0,
> >     cache_file=3D0x9b440e8 "/home/behdad/.fonts.cache-2",
config=3D0x9b44008)
> >     at fccache.c:152
> > 152         if (write (fd, chars, strlen(chars)+1) !=3D
strlen(chars)+1)
> > (gdb) bt
> > #0  0x004d4d42 in FcGlobalCacheSave (cache=3D0x9b4cad0,
> >     cache_file=3D0x9b440e8 "/home/behdad/.fonts.cache-2",
config=3D0x9b44008)
> >     at fccache.c:152
> > #1  0x004d8910 in FcConfigBuildFonts (config=3D0x9b44008) at
fccfg.c:328
> > #2  0x004e19cd in FcInitLoadConfigAndFonts () at fcinit.c:85
> > #3  0x004e1c66 in FcInit () at fcinit.c:103
> > #4  0x08048b16 in main (argc=3D1, argv=3D0xbfd03b84) at fc-list.c:119
> > (gdb)
> >
> >
> > (gdb) p cache->dirs[0]
> > $4 =3D {next =3D 0x9b46538, name =3D 0x9b5ba88
> > "/usr/share/fonts/bitstream-vera",
> >   metadata =3D {magic =3D -66913278, count =3D 10361, bank =3D
778904322,
> >     pattern_count =3D 10, patternelt_count =3D 164, valuelist_count
=3D 164,
> >     str_count =3D 238, langset_count =3D 10, charset_count =3D 10,
> >     charset_numbers_count =3D 90, charset_leaf_count =3D 90,
> >     charset_leaf_idx_count =3D 10}, offset =3D 0, ent =3D 0x9b54998}
> >
> >
> > I cannot do p chars, because of inlining I guess.
> >
> >
> > --behdad
> > http://behdad.org/
> >
> > "Commandment Three says Do Not Kill, Amendment Two says Blood
Will Spill"
> > =09-- Dan Bern, "New American Language"
> > _______________________________________________
> > Fontconfig mailing list
> > Fontconfig@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/fontconfig
> >
>
>
--behdad
http://behdad.org/

"Commandment Three says Do Not Kill, Amendment Two says Blood Will
Spill"
=09-- Dan Bern, "New American Language"

Patrick Lam

2006-Jan-26 16:35 UTC

head link

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

Matthias Clasen wrote:> Can somebody update me about the intended semantics of
> $HOME/.fonts.cache-2. Why do we need it, when there are caches for
> all font directories in /var/cache/fontconfig and in $HOME/.fonts ?
Just to confirm Frederic''s answer, the fonts.cache-2 files in
/var/cache/fontconfig are directory caches, while ~/.fonts.cache-2 is a
global cache.  The directory caches are generated by fc-cache (so
regular users can''t generate them), whereas the global cache is
generated on a per-user basis to handle any missing cache information
that''s no in the directory caches.
> $ fc-cat .fonts.cache-2 | wc -l
> 0
Probably a bug that I''ve squashed.

pat

Fontconfig - Jan 2006 - confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] Re: confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] Re: confused about semantics of $HOME/.fonts.cache-2

[Fontconfig] confused about semantics of $HOME/.fonts.cache-2