Following an upgrade from F23 to F24 yesterday, an attempt to start Open Office gives the following in /var/log/messages (worked fine under F23):
Nov 28 12:22:13 mustang kwin_x11: QXcbWindow: Unhandled client message: "_NET_CURRENT_DESKTOP" Nov 28 12:22:24 mustang kwin_x11: QXcbConnection: XCB error: 3 (BadWindow), sequence: 7727, resource id: 33558153, major code: 18 (ChangeProperty), minor code: 0 Nov 28 12:22:26 mustang plasmashell: QXcbConnection: XCB error: 2 (BadValue), sequence: 5590, resource id: 2097157, major code: 141 (Unknown), minor code: 3 Nov 28 12:22:26 mustang plasmashell: QXcbConnection: XCB error: 2 (BadValue), sequence: 5645, resource id: 94371842, major code: 141 (Unknown), minor code: 3
On 11/28/16 09:55, Stephen Davies wrote:
Following an upgrade from F23 to F24 yesterday, an attempt to start Open Office gives the following in /var/log/messages (worked fine under F23):
Nov 28 12:22:13 mustang kwin_x11: QXcbWindow: Unhandled client message: "_NET_CURRENT_DESKTOP" Nov 28 12:22:24 mustang kwin_x11: QXcbConnection: XCB error: 3 (BadWindow), sequence: 7727, resource id: 33558153, major code: 18 (ChangeProperty), minor code: 0 Nov 28 12:22:26 mustang plasmashell: QXcbConnection: XCB error: 2 (BadValue), sequence: 5590, resource id: 2097157, major code: 141 (Unknown), minor code: 3 Nov 28 12:22:26 mustang plasmashell: QXcbConnection: XCB error: 2 (BadValue), sequence: 5645, resource id: 94371842, major code: 141 (Unknown), minor code:
Works fine here. Just started it from the menu, picking "Spreadsheet".
Have you tried starting from a konsole session to see if you get more info?
On 28/11/16 14:26, Ed Greshko wrote:
On 11/28/16 09:55, Stephen Davies wrote:
Following an upgrade from F23 to F24 yesterday, an attempt to start Open Office gives the following in /var/log/messages (worked fine under F23):
Nov 28 12:22:13 mustang kwin_x11: QXcbWindow: Unhandled client message: "_NET_CURRENT_DESKTOP" Nov 28 12:22:24 mustang kwin_x11: QXcbConnection: XCB error: 3 (BadWindow), sequence: 7727, resource id: 33558153, major code: 18 (ChangeProperty), minor code: 0 Nov 28 12:22:26 mustang plasmashell: QXcbConnection: XCB error: 2 (BadValue), sequence: 5590, resource id: 2097157, major code: 141 (Unknown), minor code: 3 Nov 28 12:22:26 mustang plasmashell: QXcbConnection: XCB error: 2 (BadValue), sequence: 5645, resource id: 94371842, major code: 141 (Unknown), minor code:
Works fine here. Just started it from the menu, picking "Spreadsheet".
Have you tried starting from a konsole session to see if you get more info?
If I run /usr/bin/soffice as root, it works and the Office menu appears. If I do the same as myself (sh -x /usr/bin/soffice), I get:
+ SAL_ENABLE_FILE_LOCKING=1 + export SAL_ENABLE_FILE_LOCKING ++ pwd + sd_cwd=/home/scldad + sd_res=/usr/bin/soffice + '[' -h /usr/bin/soffice ']' ++ dirname /usr/bin/soffice + cd /usr/bin ++ basename /usr/bin/soffice + sd_basename=soffice ++ ls -l soffice ++ sed 's/.*soffice -> //g' + sd_res=/opt/openoffice4/program/soffice + '[' -h /opt/openoffice4/program/soffice ']' ++ dirname /opt/openoffice4/program/soffice + cd /opt/openoffice4/program ++ pwd + sd_prog=/opt/openoffice4/program + cd /home/scldad ++ basename /usr/bin/soffice + sd_binary=soffice.bin + sd_pagein_args=@pagein-common + /opt/openoffice4/program/pagein -L/opt/openoffice4/program @pagein-common + '[' -x /opt/openoffice4/program/javaldx ']' + case "`uname -s`" in ++ uname -s ++ /opt/openoffice4/program/javaldx -env:INIFILENAME=vnd.sun.star.pathname:/opt/openoffice4/program/redirectrc + my_path=/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64 + '[' -n /usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64 ']' + LD_LIBRARY_PATH=/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64:/opt/LCS/bv103/doci:/opt/LCS/bv103/dm/lib + export LD_LIBRARY_PATH + unset XENVIRONMENT + '[' -f /etc/adabasrc ']' + trap 'kill -9 $!' TERM + wait 11722 + /opt/openoffice4/program/soffice.bin + sd_ret=0 + '[' 0 -eq 79 -o 0 -eq 81 ']' + exit 0
Running soffice without sh -x gives absolutely nothing.
On 11/28/16 12:22, Stephen Davies wrote:
If I run /usr/bin/soffice as root, it works and the Office menu appears. If I do the same as myself (sh -x /usr/bin/soffice), I get:
- SAL_ENABLE_FILE_LOCKING=1
- export SAL_ENABLE_FILE_LOCKING
++ pwd
- sd_cwd=/home/scldad
- sd_res=/usr/bin/soffice
- '[' -h /usr/bin/soffice ']'
++ dirname /usr/bin/soffice
- cd /usr/bin
++ basename /usr/bin/soffice
- sd_basename=soffice
++ ls -l soffice ++ sed 's/.*soffice -> //g'
- sd_res=/opt/openoffice4/program/soffice
- '[' -h /opt/openoffice4/program/soffice ']'
++ dirname /opt/openoffice4/program/soffice
- cd /opt/openoffice4/program
++ pwd
- sd_prog=/opt/openoffice4/program
- cd /home/scldad
++ basename /usr/bin/soffice
- sd_binary=soffice.bin
- sd_pagein_args=@pagein-common
- /opt/openoffice4/program/pagein -L/opt/openoffice4/program @pagein-common
- '[' -x /opt/openoffice4/program/javaldx ']'
- case "`uname -s`" in
++ uname -s ++ /opt/openoffice4/program/javaldx -env:INIFILENAME=vnd.sun.star.pathname:/opt/openoffice4/program/redirectrc
my_path=/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64
- '[' -n
/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64 ']'
LD_LIBRARY_PATH=/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64:/opt/LCS/bv103/doci:/opt/LCS/bv103/dm/lib
- export LD_LIBRARY_PATH
- unset XENVIRONMENT
- '[' -f /etc/adabasrc ']'
- trap 'kill -9 $!' TERM
- wait 11722
- /opt/openoffice4/program/soffice.bin
- sd_ret=0
- '[' 0 -eq 79 -o 0 -eq 81 ']'
- exit 0
Running soffice without sh -x gives absolutely nothing.
Well, I would check the environment compared to root's. I also find the reference to LD_LIBRARY_PATH suspect.
Here is what I get.
[egreshko@acer ~]$ sh -x /usr/bin/soffice + LO_SAVE_LC_ALL= + LC_ALL=C + export LC_ALL + SAL_ENABLE_FILE_LOCKING=1 + export SAL_ENABLE_FILE_LOCKING ++ pwd + sd_cwd=/home/egreshko + sd_res=/usr/bin/soffice + '[' -h /usr/bin/soffice ']' ++ dirname /usr/bin/soffice + cd /usr/bin ++ basename /usr/bin/soffice + sd_basename=soffice ++ ls -l soffice ++ sed 's/.*soffice -> //g' + sd_res=/usr/lib64/libreoffice/program/soffice + '[' -h /usr/lib64/libreoffice/program/soffice ']' ++ dirname /usr/lib64/libreoffice/program/soffice + cd /usr/lib64/libreoffice/program ++ pwd + sd_prog=/usr/lib64/libreoffice/program + cd /home/egreshko + '[' -e /usr/lib64/libreoffice/program/ooenv ']' + GDBTRACECHECK= + STRACECHECK= + VALGRINDCHECK= + RRCHECK= + checks= + EXTRAOPT= + test -n '' + test -n '' + echo '' + grep -q cc + case "`uname -s`" in ++ uname -s + LC_ALL= + '[' -n '' ']' + '[' -n '' -a -z '' ']' + exec /usr/lib64/libreoffice/program/oosplash
On 28/11/16 15:32, Ed Greshko wrote:
On 11/28/16 12:22, Stephen Davies wrote:
If I run /usr/bin/soffice as root, it works and the Office menu appears. If I do the same as myself (sh -x /usr/bin/soffice), I get:
- SAL_ENABLE_FILE_LOCKING=1
- export SAL_ENABLE_FILE_LOCKING
++ pwd
- sd_cwd=/home/scldad
- sd_res=/usr/bin/soffice
- '[' -h /usr/bin/soffice ']'
++ dirname /usr/bin/soffice
- cd /usr/bin
++ basename /usr/bin/soffice
- sd_basename=soffice
++ ls -l soffice ++ sed 's/.*soffice -> //g'
- sd_res=/opt/openoffice4/program/soffice
- '[' -h /opt/openoffice4/program/soffice ']'
++ dirname /opt/openoffice4/program/soffice
- cd /opt/openoffice4/program
++ pwd
- sd_prog=/opt/openoffice4/program
- cd /home/scldad
++ basename /usr/bin/soffice
- sd_binary=soffice.bin
- sd_pagein_args=@pagein-common
- /opt/openoffice4/program/pagein -L/opt/openoffice4/program @pagein-common
- '[' -x /opt/openoffice4/program/javaldx ']'
- case "`uname -s`" in
++ uname -s ++ /opt/openoffice4/program/javaldx -env:INIFILENAME=vnd.sun.star.pathname:/opt/openoffice4/program/redirectrc
my_path=/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64
- '[' -n
/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64 ']'
LD_LIBRARY_PATH=/usr/java/jdk1.7.0_71/jre/lib/amd64/client:/usr/java/jdk1.7.0_71/jre/lib/amd64/native_threads:/usr/java/jdk1.7.0_71/jre/lib/amd64:/opt/LCS/bv103/doci:/opt/LCS/bv103/dm/lib
- export LD_LIBRARY_PATH
- unset XENVIRONMENT
- '[' -f /etc/adabasrc ']'
- trap 'kill -9 $!' TERM
- wait 11722
- /opt/openoffice4/program/soffice.bin
- sd_ret=0
- '[' 0 -eq 79 -o 0 -eq 81 ']'
- exit 0
Running soffice without sh -x gives absolutely nothing.
Well, I would check the environment compared to root's. I also find the reference to LD_LIBRARY_PATH suspect.
Here is what I get.
[egreshko@acer ~]$ sh -x /usr/bin/soffice
- LO_SAVE_LC_ALL=
- LC_ALL=C
- export LC_ALL
- SAL_ENABLE_FILE_LOCKING=1
- export SAL_ENABLE_FILE_LOCKING
++ pwd
- sd_cwd=/home/egreshko
- sd_res=/usr/bin/soffice
- '[' -h /usr/bin/soffice ']'
++ dirname /usr/bin/soffice
- cd /usr/bin
++ basename /usr/bin/soffice
- sd_basename=soffice
++ ls -l soffice ++ sed 's/.*soffice -> //g'
- sd_res=/usr/lib64/libreoffice/program/soffice
- '[' -h /usr/lib64/libreoffice/program/soffice ']'
++ dirname /usr/lib64/libreoffice/program/soffice
- cd /usr/lib64/libreoffice/program
++ pwd
- sd_prog=/usr/lib64/libreoffice/program
- cd /home/egreshko
- '[' -e /usr/lib64/libreoffice/program/ooenv ']'
- GDBTRACECHECK=
- STRACECHECK=
- VALGRINDCHECK=
- RRCHECK=
- checks=
- EXTRAOPT=
- test -n ''
- test -n ''
- echo ''
- grep -q cc
- case "`uname -s`" in
++ uname -s
- LC_ALL=
- '[' -n '' ']'
- '[' -n '' -a -z '' ']'
- exec /usr/lib64/libreoffice/program/oosplash
Looks as if you are running libreoffice rather than openoffice (from Apache).
I also suspected the LD_LIBRARY_PATH but the root version is the same.
On 11/28/16 13:10, Stephen Davies wrote:
Looks as if you are running libreoffice rather than openoffice (from Apache).
I also suspected the LD_LIBRARY_PATH but the root version is the same.
Oh, you are right. I am running libreoffice 5.2 as supplied by Fedora Project. I missed that you're running something else.
Maybe Apache has a support mailing list? It would seem best to ask them.
On 11/28/16 13:10, Stephen Davies wrote:
Looks as if you are running libreoffice rather than openoffice (from Apache).
I also suspected the LD_LIBRARY_PATH but the root version is the same.
Oh, FWIW, I have a VM. So, I erases libreoffice from my system and installed Apache_OpenOffice_4.1.3.
I can't start it in either my user or root.
[egreshko@meimei ~]$ cat err + SAL_ENABLE_FILE_LOCKING=1 + export SAL_ENABLE_FILE_LOCKING ++ pwd + sd_cwd=/home/egreshko + sd_res=/usr/bin/soffice + '[' -h /usr/bin/soffice ']' ++ dirname /usr/bin/soffice + cd /usr/bin ++ basename /usr/bin/soffice + sd_basename=soffice ++ sed 's/.*soffice -> //g' ++ ls -l soffice + sd_res=/opt/openoffice4/program/soffice + '[' -h /opt/openoffice4/program/soffice ']' ++ dirname /opt/openoffice4/program/soffice + cd /opt/openoffice4/program ++ pwd + sd_prog=/opt/openoffice4/program + cd /home/egreshko ++ basename /usr/bin/soffice + sd_binary=soffice.bin + sd_pagein_args=@pagein-common + /opt/openoffice4/program/pagein -L/opt/openoffice4/program @pagein-common + '[' -x /opt/openoffice4/program/javaldx ']' + case "`uname -s`" in ++ uname -s ++ /opt/openoffice4/program/javaldx -env:INIFILENAME=vnd.sun.star.pathname:/opt/openoffice4/program/redirectrc + my_path=/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64/client:/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64/native_threads:/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64 + '[' -n /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64/client:/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64/native_threads:/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64 ']' + LD_LIBRARY_PATH=/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64/client:/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64/native_threads:/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.111-3.b16.fc25.x86_64/jre/lib/amd64 + export LD_LIBRARY_PATH + unset XENVIRONMENT + SAL_NO_XINITTHREADS=true + export SAL_NO_XINITTHREADS + '[' -f /etc/adabasrc ']' + trap 'kill -9 $!' TERM + wait 2426 + /opt/openoffice4/program/soffice.bin Application Error
Fatal exception: Signal 6 /usr/bin/soffice: line 121: 2426 Aborted (core dumped) "$sd_prog/$sd_binary" "$@" + sd_ret=134 + '[' 134 -eq 79 -o 134 -eq 81 ']' + exit 134
On 11/28/16 14:42, Ed Greshko wrote:
Ooops... Forget this last try.... I was on an F25 VM.
Going to do same on F24.
Wait a bit....
Well the core dump was the same in F24 as F25.
I have found this.... https://community.spiceworks.com/topic/1808558-openoffice-4-1-2-4-1-3-crashe...
But can't connect to the bugzilla system to see what https://bz.apache.org/ooo/show_bug.cgi?id=124948 has to say.
This is probably out of scope for Fedora. So, not sure if I'll try and track it down.
On 28 November 2016 at 07:26, Ed Greshko ed.greshko@greshko.com wrote:
On 11/28/16 14:42, Ed Greshko wrote:
Ooops... Forget this last try.... I was on an F25 VM.
Going to do same on F24.
Wait a bit....
Well the core dump was the same in F24 as F25.
I have found this.... https://community.spiceworks.com/topic/1808558-openoffice-4-1-2-4-1-3-crashe...
But can't connect to the bugzilla system to see what https://bz.apache.org/ooo/show_bug.cgi?id=124948 has to say.
This is probably out of scope for Fedora. So, not sure if I'll try and track it down.
You're right it is out of scope as Apache OpenOffice has never been in Fedora, and given the state of the project I doubt it ever will be.
OP for your own sanity and security stop using AOO. The recent 4.1.3 release fixed a single arbitrary code execution bug that they were notified of in October last year (shortly after the 4.1.2 bugfix release in August 2015) and it's taken them over a year to get a release out for a arbitrary code execution security issue (note this particular issue was found and fixed in LibreOffice over 2 years ago).
The project is only hanging on through sheer stubbornness of half a doze individuals and development is at a standstill, with the few people active on the development mailing lists frequently failing to even be able to build the source code.
There really is no point or use in using Apache OpenOffice at this point (and hasn't been for a year now).
Do you have the same problems opening LO in your system?
On 11/28/16 19:11, James Hogarth wrote:
You're right it is out of scope as Apache OpenOffice has never been in Fedora, and given the state of the project I doubt it ever will be.
Yes, but I've never been too much of a fan of telling people what they should and shouldn't run. And I like a challenge from time to time when the list isn't that busy. :-)
FWIW, I was able to access the Apache BZ and fixed the crash I was having by installing gdk-pixbuf2-xlib. And OpenOffice works just fine.
So, the OP has to figure out what in his specific user environment is causing the issue. Many times these things are fixed by moving the ~/.openoffice directory out of the way and starting over. :-)
On 28 Nov 2016 12:01 pm, "Ed Greshko" ed.greshko@greshko.com wrote:
On 11/28/16 19:11, James Hogarth wrote:
You're right it is out of scope as Apache OpenOffice has never been in Fedora, and given the state of the project I doubt it ever will be.
Yes, but I've never been too much of a fan of telling people what they
should and
shouldn't run. And I like a challenge from time to time when the list
isn't that busy. :-)
FWIW, I was able to access the Apache BZ and fixed the crash I was having
by installing
gdk-pixbuf2-xlib. And OpenOffice works just fine.
"Just fine" ... If you ignore all the AOO problems indeed...
An important thing to be aware of is that they build the Linux binaries on CentOS 5 which will have a range of poor behavioural issues due to being built against ancient glibc and gcc versions etc.
Generally I agree on the whole "don't tell someone what they should or shouldn't run" mantra but specifically in the AOO situation given the very poor attitude to security issues, almost complete standstill of development, that it is a subset of LO (all AOO commits, as few and far between add they are, get evaluated for inclusion in LO) etc etc I deem it required to ensure people are properly informed about its near-dead status and the potential for security issues (and general exaggerated bugginess due to lack of activity) ...
If he continues to want to use a project that is in such a dire state then so be it, but LO is the supported fork of the old Oo.org code for a reason.