Bug 176680

Created attachment 122630 [details]
screenshot of gnucash with unreadable font

Is this a rawhide or test1 install?

It's a test1 base, but as I used yum to install gnucash, and test1's yum
repositories default to rawhide, I suppose it's a hybrid. 

See: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=173648#c3

Does that procedure help?

In response to comment #4, I tried upgrading libfontenc. I get the same result.
(Did I understand your suggestion correctly?)

It installed upgrades to the 32 and 64 bit versions of this package, as well as
libfontenc-devel. It also installed xorg-x11-filesystem. 

Just tried it on my FC5 test1 32-bit system on my laptop, and I get the same
result (I tried it with gnucash-1.8.11-4. I've had problems trying to upgrade it
to gnucash 1.8.12, but I don't that upgrade issue is relevant to this bug).

No, I mean the part about regenerating fonts.dir/fonts.scale.

chkfontpath --list | grep /usr | awk -F ':' '{ print $2 }'| while read foo ; do
rm -f $foo/fonts.{dir,scale} ; done
service xfs restart

(Note, at the end of this you will need to restart the X server as well.)

It works on the FedoraCore 64 bit system with gnucash-1.8.12. It does not work
on the FedoraCore 32 bit system with gnucash-1.8.11. In fact, it borks the GUI;
when I restart, I get a

"Fatal server error
Could not open default font 'fixed'" 

error. I get the sequence where I'm prompted to reconfigure X, but never get to
the config tool.

Am I missing something, or should I open a separate bug report? Xorg log file
available on request.

Hm, I wonder if something didn't get remade right.

Is there a 'fixed' in /usr/share/X11/fonts/misc/fonts.alias, and is that
directory in chkfontpath --list?

Yes and no. Specifically, in fonts.alias, there is

fixed        -misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-iso8859-1

(most other fonts refer to -misc-fixed as well, e.g.
7x13euro     -misc-fixed-medium-r-normal--13-120-75-75-c-70-iso8859-15)

But from chkfontpath --list:

1: /usr/share/X11/fonts/misc:unscaled
2: /usr/share/X11/fonts/75dpi:unscaled
3: /usr/share/X11/fonts/100dpi:unscaled
4: /usr/share/X11/fonts/Type1
5: /usr/share/X11/fonts/TTF
6: /usr/share/fonts/default/Type1
7:
8: /usr/share/fonts/japanese/misc:unscaled
9: /usr/share/fonts/japanese/misc
10: /usr/share/fonts/japanese/TrueType

In this respect, both 32 bit and 64 bit FC5 test1 systems are identical.

However, there's one more interesting bit from my 32-bit /var/log/messages,
before the XKeepsCrashing script.


On boot, I get this sequence several times:

Jan  7 09:28:05 FedoraCore5 kernel: agpgart: Found an AGP 2.0 compliant device
at 0000:00:00.0.
Jan  7 09:28:05 FedoraCore5 kernel: agpgart: Putting AGP V2 device at
0000:00:00.0 into 4x mode
Jan  7 09:28:05 FedoraCore5 kernel: agpgart: Putting AGP V2 device at
0000:01:05.0 into 4x mode
Jan  7 09:28:11 FedoraCore5 gdm[2600]: gdm_slave_xioerror_handler: Fatal X error
- Restarting :0


followed by the XKeepsCrashing script message and 

Jan  7 09:28:41 FedoraCore5 gpm[2139]: Request on 6 (console 7)
Jan  7 09:28:41 FedoraCore5 gpm[2139]: *** info [client.c(284)]:
Jan  7 09:28:41 FedoraCore5 gpm[2139]: Closing
Jan  7 09:28:41 FedoraCore5 gdm[2546]: Failed to start X server several times in
a short time period; disabling display :0


When I try gdm from runlevel 3 these show up in /var/log/messages:

Jan  7 09:34:21 FedoraCore5 gconfd (root-3509): starting (version 2.12.1), pid
3509 user 'root'
Jan  7 09:34:21 FedoraCore5 gconfd (root-3509): Resolved address
"xml:readonly:/etc/gconf/gconf.xml.mandatory" to a read-only configuration
source at position 0
Jan  7 09:34:21 FedoraCore5 gconfd (root-3509): Resolved address
"xml:readwrite:/root/.gconf" to a writable configuration source at position 1
Jan  7 09:34:21 FedoraCore5 gconfd (root-3509): Resolved address
"xml:readonly:/etc/gconf/gconf.xml.defaults" to a read-only configuration source
at position 2
Jan  7 09:34:23 FedoraCore5 gpm[2139]: *** info [client.c(137)]:
Jan  7 09:34:23 FedoraCore5 gpm[2139]: Connecting at fd 6
Jan  7 09:34:25 FedoraCore5 gpm[2139]: *** info [client.c(275)]:
Jan  7 09:34:25 FedoraCore5 gpm[2139]: Request on 6 (console 7)
Jan  7 09:34:25 FedoraCore5 gpm[2139]: *** info [client.c(284)]:
Jan  7 09:34:25 FedoraCore5 gpm[2139]: Closing
Jan  7 09:34:25 FedoraCore5 kernel: Slab corruption: (Tainted: P     ) start=e69
f9000, len=4096
Jan  7 09:34:25 FedoraCore5 kernel:  [<c0148a4c>] check_poison_obj+0xb8/0x17b
Jan  7 09:34:25 FedoraCore5 kernel:  [<c014a2c0>] cache_alloc_debugcheck_after+0
x66/0x11b
Jan  7 09:34:25 FedoraCore5 kernel:  [<c01bd70d>] selinux_file_permission+0xe0/0
x150
Jan  7 09:34:25 FedoraCore5 kernel:  [<c017db4a>] seq_read+0x0/0x2b9
Jan  7 09:34:25 FedoraCore5 kernel:  [<c014a69b>] kmem_cache_alloc+0x52/0x86
Jan  7 09:34:25 FedoraCore5 kernel:  [<c017dd07>] seq_read+0x1bd/0x2b9
Jan  7 09:34:25 FedoraCore5 kernel:  [<c017dd07>] seq_read+0x1bd/0x2b9
Jan  7 09:34:25 FedoraCore5 kernel:  [<c017db4a>] seq_read+0x0/0x2b9
Jan  7 09:34:25 FedoraCore5 kernel:  [<c015f58e>] vfs_read+0xa0/0x158
Jan  7 09:34:25 FedoraCore5 kernel:  [<c015f8f7>] sys_read+0x41/0x6a
Jan  7 09:34:25 FedoraCore5 kernel:  [<c0102ec1>] syscall_call+0x7/0xb
Jan  7 09:34:25 FedoraCore5 kernel: 0c0: f3 3d ff ff 00 00 00 00 6b 6b 6b 6b 6b
6b 6b 6b

Actually, the FC5 test1 64 bit GUI borks too... I was testing gnucash (from
another room) via SSH -X, so the results are consistent.

I missed a step - it all works fine - after - updating the applicable xorg-x11
packages (from FC5 test1 to today's rawhide)

OK, closing then. This should mean anyone installing or upgrading to final FC5
shouldn't see this.