FS#14846 - [go-openoffice] 3.1.0rc5-1 fails to run

Attached to Project: Arch Linux
Opened by Curtis (foxcm2000) - Wednesday, 27 May 2009, 01:02 GMT
Last edited by Andreas Radke (AndyRTR) - Wednesday, 10 June 2009, 21:31 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Thomas Bächler (brain0)
Andreas Radke (AndyRTR)
Architecture All
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
The go-openoffice packages fail to run using the 3.1.0rc5-1 release. The splash screen shows up for a split-second, then immediately disappears. Earlier releases worked, but after manually downgrading I have found that all releases I have from the 3.1 series now fail to load with the same bug.

Additional info:
Running x86_64 Arch using the testing repository kernel 2.6.29.4-1, glibc 2.10.1-2, xorg-server 1.6.1-1, gcc 4.4, under KDE 4.2.3.

There is zero debug information printed when I try to run executables from the console, so I cannot give any more information.

Steps to reproduce:

Install go-openoffice 3.1.0rc5-1
Attempt to run any program in the office suite.
This task depends upon

Closed by  Andreas Radke (AndyRTR)
Wednesday, 10 June 2009, 21:31 GMT
Reason for closing:  Works for me
Comment by Curtis (foxcm2000) - Thursday, 28 May 2009, 00:43 GMT
I just put the mainline openoffice 3.1 packages on my system and.... exact same bug, so it is not specific to go-openoffice.
Comment by Curtis (foxcm2000) - Thursday, 28 May 2009, 12:32 GMT
Some more followup: The latest builds take directly from openoffice.org have the same bug. Also, I tried building openoffice 3.1 from source, but both Sun's JDK and openjdk6 experienced bus-errors during the build process.
Comment by Greg (dolby) - Thursday, 28 May 2009, 12:46 GMT
Search, and make if there isnt one already available, a report upstream then.
Comment by Andreas Radke (AndyRTR) - Thursday, 28 May 2009, 20:47 GMT
please backup your ~/.openofficeX folder and rename it to get a fresh one. please start soffice from command line. Both, vanilla and go-oo are known to work.
Comment by Curtis (foxcm2000) - Friday, 29 May 2009, 00:23 GMT
The first thing I did was to remove all the various versions of the .openoffice folders, but it didn't help (sorry I forgot to mention that earlier).
One thing I can tell you is that openoffice is running long enough to re-generate those folders because they are recreated after each attempt at running the program.

I've been running everything from the command line, but I've never seen anything print out to the terminal... are there any debug switches I could throw?

A few other points:
I've checked my environment, and the OOO_FORCE_DESKTOP=gnome variable is present, and unsetting it didn't help.
I'm running the testing/core/extra/community repositories on a 64 bit machine, and openoffice 3.1 was working OK up until about 2 weeks ago. I'm running a couple of custom-AUR installs but I really hope they couldn't cause this issue, but here's my output of pacman -Qm just in case:
bin32-wine 1.1.21-1
google-earth 5.0.11337.1968-1
mplayer-minimal-svn 29318-1
scribus-svn 13580-1

If I run the naked soffice.bin binary without any of the wrapper shells... I get a bus error! I would expect it to fail since it probably doesn't know where all of its libraries are without the wrapper scripts, but a bus error is more serious.

OK, here's where I fly off into conjecture land, but I think something on my install is not happy when it comes to either the recent glibc 2.10 or gcc 4.4 updates since these are the types of upgrades that can cause ABI issues. For example, I don't know which packages are being compiled with gcc 4.4 vs. 4.3 (I'm guessing testing is starting to see 4.4 packages while the other repos are still on 4.3, but that's a guess). I tried doing some manual downgrades of glibc back to the 2.9 series, but it didn't help.

This bus error actually jives with bugs I've seen in a few select programs as well. First, the nvidia graphics installers (the one's from nvidia's site, not the pacman repos) have worked fine for me for a very long time on multiple distros including Arch. However, last week I did a driver upgrade and the installer program failed with a bus error, and this occurred on multiple versions of the drivers, including ones I had previously successfully installed on this machine using Arch. Additionally, when I was trying to compile openoffice from source, it went along fine for 30 minutes until the java binary failed... with a bus error.

I'm confused. I hope I haven't managed to fubar my install in any deep way. I will say that the vast majority of programs work just fine and the system is very stable, so I don't think this is some random hardware error (I've had those and they do different bad things to this system).

So I'm out of options, I hope that something in here is a big red flag you guys have seen before and it makes sense.
Comment by Curtis (foxcm2000) - Friday, 29 May 2009, 00:38 GMT
And one more thing that probably has nothing to do with this bug, but it annoys the hell out of me and might be a symptom of deeper issues:

When reinstalling gcc just to see what happens:
pacman -Sy gcc
/tmp/alpm_8aNKso/.INSTALL: 2: Syntax error: "(" unexpected
error: scriptlet failed to execute correctly

What causes these scriplet failures? Would this break software installations so it looks like something is installed but it's actually broken?
Comment by Curtis (foxcm2000) - Friday, 29 May 2009, 03:03 GMT
One last thing, I ran soffice.bin through gdb, and while I don't know if this is the cause, here's the output:

(gdb) run ./soffice.bin
Starting program: /opt/openoffice/program/soffice.bin ./soffice.bin
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
warning: Lowest section in /usr/lib/libicudata.so.40 is .hash at 0000000000000120
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[New Thread 0x7f3d05453760 (LWP 8378)]
[New Thread 0x7f3cfbb26910 (LWP 8381)]
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
---Type <return> to continue, or q <return> to quit---
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)

Program received signal SIGBUS, Bus error.
[Switching to Thread 0x7f3d05453760 (LWP 8378)]
0x00007f3cfa0f6e73 in ?? () from /opt/openoffice/basis-link/ure-link/lib/libgcc3_uno.so
(gdb) quit
The program is running. Exit anyway? (y or n) y
Comment by Andreas Radke (AndyRTR) - Friday, 29 May 2009, 03:39 GMT
please file a new bug for the scriptlet issue.

can you run strace soffice? so far you are the only one with this problem. so this doesn't seem a generic problem and more likely a problem with your system.
Comment by Curtis (foxcm2000) - Friday, 29 May 2009, 12:47 GMT
I have the full strace attached, the last few interesting lines are:
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_SHARED, 13, 0) = 0x7fe9e7493000
mmap(NULL, 4096, PROT_READ|PROT_EXEC, MAP_SHARED, 13, 0) = 0x7fe9e7492000
--- SIGBUS (Bus error) @ 0 (0) ---
+++ killed by SIGBUS +++
Comment by Curtis (foxcm2000) - Friday, 29 May 2009, 14:36 GMT
OK I know what the source of the bug is! I downgraded my kernel from the 2.6.29.4-1 package to the 2.6.29.3-1 package, and now openoffice runs just fine again. I'm not exactly sure what the bug is, but look at the diff between the two kernel versions.
Downgrading the kernel also fixed some bugs I had been seeing in wine using the steam client as well, so I think something is not happy with either the patches to 2.6.29-4, or maybe how it got built.

I have a feeling this may be a hardware-specific bug which is why I'm seeing it and other people are not, so here's more information on my hardware:
lspci:
00:00.0 Host bridge: Intel Corporation 82X38/X48 Express DRAM Controller (rev 01)
00:01.0 PCI bridge: Intel Corporation 82X38/X48 Express Host-Primary PCI Express Bridge (rev 01)
00:1a.0 USB Controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #4 (rev 02)
00:1a.1 USB Controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #5 (rev 02)
00:1a.2 USB Controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #6 (rev 02)
00:1a.7 USB Controller: Intel Corporation 82801I (ICH9 Family) USB2 EHCI Controller #2 (rev 02)
00:1b.0 Audio device: Intel Corporation 82801I (ICH9 Family) HD Audio Controller (rev 02)
00:1c.0 PCI bridge: Intel Corporation 82801I (ICH9 Family) PCI Express Port 1 (rev 02)
00:1c.2 PCI bridge: Intel Corporation 82801I (ICH9 Family) PCI Express Port 3 (rev 02)
00:1d.0 USB Controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #1 (rev 02)
00:1d.1 USB Controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #2 (rev 02)
00:1d.2 USB Controller: Intel Corporation 82801I (ICH9 Family) USB UHCI Controller #3 (rev 02)
00:1d.7 USB Controller: Intel Corporation 82801I (ICH9 Family) USB2 EHCI Controller #1 (rev 02)
00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev 92)
00:1f.0 ISA bridge: Intel Corporation 82801IR (ICH9R) LPC Interface Controller (rev 02)
00:1f.2 SATA controller: Intel Corporation 82801IR/IO/IH (ICH9R/DO/DH) 6 port SATA AHCI Controller (rev 02)
00:1f.3 SMBus: Intel Corporation 82801I (ICH9 Family) SMBus Controller (rev 02)
01:00.0 VGA compatible controller: nVidia Corporation GeForce 8800 GT (rev a2)
02:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8056 PCI-E Gigabit Ethernet Controller (rev 12)

Here's output of lsmod:
Module Size Used by
nvidia 10247240 38
w83627ehf 24256 0
hwmon_vid 3664 1 w83627ehf
coretemp 6864 0
hid_logitech 10800 0
ff_memless 5848 1 hid_logitech
usbhid 40640 1 hid_logitech
hid 49568 2 hid_logitech,usbhid
snd_seq_oss 33760 0
snd_seq_midi_event 7248 1 snd_seq_oss
snd_seq 57024 4 snd_seq_oss,snd_seq_midi_event
snd_seq_device 7268 2 snd_seq_oss,snd_seq
snd_pcm_oss 42656 0
snd_mixer_oss 16784 1 snd_pcm_oss
pcspkr 2800 0
sg 30432 0
snd_hda_codec_analog 79472 1
serio_raw 6132 0
iTCO_wdt 12912 0
iTCO_vendor_support 3540 1 iTCO_wdt
i2c_i801 10416 0
i2c_core 24000 2 nvidia,i2c_i801
snd_hda_intel 29256 1
snd_hda_codec 68112 2 snd_hda_codec_analog,snd_hda_intel
snd_hwdep 8632 1 snd_hda_codec
snd_pcm 80920 3 snd_pcm_oss,snd_hda_intel,snd_hda_codec
snd_timer 22816 2 snd_seq,snd_pcm
snd 65928 13 snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_hda_codec_analog,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_timer
soundcore 7680 1 snd
snd_page_alloc 9760 2 snd_hda_intel,snd_pcm
uhci_hcd 25424 0
sky2 53428 0
ehci_hcd 39100 0
usbcore 160400 4 usbhid,uhci_hcd,ehci_hcd
x38_edac 5120 0
edac_core 44624 2 x38_edac
thermal 18928 0
processor 44296 1 thermal
fan 5240 0
evdev 11744 8
button 7344 0
battery 12504 0
ac 4984 0
rtc_cmos 12152 0
rtc_core 20100 1 rtc_cmos
rtc_lib 2864 1 rtc_core
ext4 233096 4
mbcache 8148 1 ext4
jbd2 62912 1 ext4
crc16 1968 1 ext4
sr_mod 16804 0
cdrom 37832 1 sr_mod
sd_mod 27528 7
ahci 36128 5
libata 175468 1 ahci
scsi_mod 117784 4 sg,sr_mod,sd_mod,libata

Other information:
Core 2 e8400 clocked at 3.6Ghz (stable)
8GB of DDR2 memory
X38 motherboard
ICH9 storage controller hub
Nvidia 8800GT graphics, using the 185.18.10 beta driver installed manually
EXT4 filesystem on all partitions using noatime but otherwise default mount options

Comment by Andreas Radke (AndyRTR) - Friday, 29 May 2009, 15:08 GMT
assigned also to the kernel packagers. any idea? a guess: it's the closed binary blob once again ?
Comment by Curtis (foxcm2000) - Friday, 29 May 2009, 15:44 GMT
I'm not sure the binary driver is to blame becuase I was getting bus-errors when running the nvidia installer program itself. In other words, the binary blob wasn't even loaded into memory but the bug was still manifesting. Whenever I do get debug information, it appears the error is right after a malloc call, so maybe that would be a place to investigate.
   . (0 KiB)
Comment by Uli (Army) - Monday, 01 June 2009, 20:49 GMT
All I get is

$ soffice
/usr/lib/go-openoffice-3.1.0.6/program/soffice.bin: error while loading shared libraries: libicuuc.so.40: cannot open shared object file: No such file or directory

$ locate libicuuc.so
/opt/google-earth/libicuuc.so.38
/usr/lib/libicuuc.so
/usr/lib/libicuuc.so.42
/usr/lib/libicuuc.so.42.0

Not sure, if the google-earth stuff might cause the problem, but there definitely is no libicuuc.so.40!!! Simply recompile go-openoffice for testing?
Comment by Florent (Florut) - Tuesday, 02 June 2009, 16:40 GMT
This is not an OpenOffice-related problem : it's icu problem !

Just downgrate icu (remove testing repositories + pacman -S icu) and everything works just fine !
Comment by Uli (Army) - Thursday, 04 June 2009, 08:02 GMT
No it's not! I did a recompile of go-openoffice and now it runs just fine. So AndyRTR, could you please do a recompile for testing?
Comment by bloodniece (bloodniece) - Thursday, 04 June 2009, 15:39 GMT
I tried removing testing from my repos, downgraded icu, now soffice complains about jre missing.
javaldx: Could not find a Java Runtime Environment!

I have jre installed. I tried adding the path to the jre in my bash_profile to no avail.
PATH=$PATH:/opt/java/jre/bin




Comment by Andreas Radke (AndyRTR) - Tuesday, 09 June 2009, 16:20 GMT
all this is not related to the initial soffice strace issue.

chuck, can u test with a gtk based desktop? i guess it's related to kde and it's gtk design wrappers. already tried the kde wiki section hints on our OOo page?
Comment by Curtis (foxcm2000) - Wednesday, 10 June 2009, 03:38 GMT
Andreas: My bug involved a pretty low-level error with the 2.6.29.4-1 kernel package causing bus errors when I ran certain binaries, open office being among them. I fixed the problem first by manually downgrading to 2.6.29.3-1, and I just did a custom-compiled kernel from the Zen git branch today of 2.6.30-rc8 that also is working fine (been a WHILE since I rolled my old kernel, Arch made it relatively easy!)
Just so you know, I did try out openoffice on XFCE with the problematic kernel, and it still bus erred like it had before. That bug appears to be below the level of desktop environments.

As for Florent's bug, that appears to be a separate issue, that I can personally confirm. I have the development version of OO on my laptop and the recent changes to the ICU library broke it. I tried to put in symlinks to the renamed library files, but then OO failed with unknown symbols. Either re-compile the OO packages against the newest libicu (assuming they haven't broken the API) or do what Florent did and downgrade to have everything link up again.
Comment by Andreas Radke (AndyRTR) - Wednesday, 10 June 2009, 21:30 GMT
all OOo packages have been recompiled for icu. if something is still broken please report new issues. closing this one not beeing a OOo bug.

Loading...