[Bug 1079744] New: Java throws exception on start of oowriter
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 Bug ID: 1079744 Summary: Java throws exception on start of oowriter Classification: openSUSE Product: openSUSE Distribution Version: Leap 42.3 Hardware: x86-64 OS: openSUSE 42.3 Status: NEW Severity: Normal Priority: P5 - None Component: LibreOffice Assignee: bnc-team-screening@forge.provo.novell.com Reporter: Ulrich.Windl@rz.uni-regensburg.de QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- After installing the LO Upgrade 5.3 -> 5.4 (libreoffice-5.4.4.2-12.6.x86_64), a Java exception was thrown on the first start: Exception in thread "Thread-0" java.lang.NoClassDefFoundError: com/sun/star/loader/XImplementationLoader at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:763) at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142) at java.net.URLClassLoader.defineClass(URLClassLoader.java:467) at java.net.URLClassLoader.access$100(URLClassLoader.java:73) at java.net.URLClassLoader$1.run(URLClassLoader.java:368) at java.net.URLClassLoader$1.run(URLClassLoader.java:362) at java.security.AccessController.doPrivileged(Native Method) at java.net.URLClassLoader.findClass(URLClassLoader.java:361) at java.lang.ClassLoader.loadClass(ClassLoader.java:424) at java.lang.ClassLoader.loadClass(ClassLoader.java:357) Caused by: java.lang.ClassNotFoundException: com.sun.star.loader.XImplementationLoader at java.net.URLClassLoader.findClass(URLClassLoader.java:381) at java.lang.ClassLoader.loadClass(ClassLoader.java:424) at java.lang.ClassLoader.loadClass(ClassLoader.java:357) ... 11 more Exception in thread "Thread-0" java.lang.NoClassDefFoundError: com/sun/star/loader/XImplementationLoader at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:763) at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142) at java.net.URLClassLoader.defineClass(URLClassLoader.java:467) at java.net.URLClassLoader.access$100(URLClassLoader.java:73) at java.net.URLClassLoader$1.run(URLClassLoader.java:368) at java.net.URLClassLoader$1.run(URLClassLoader.java:362) at java.security.AccessController.doPrivileged(Native Method) at java.net.URLClassLoader.findClass(URLClassLoader.java:361) at java.lang.ClassLoader.loadClass(ClassLoader.java:424) at java.lang.ClassLoader.loadClass(ClassLoader.java:357) Caused by: java.lang.ClassNotFoundException: com.sun.star.loader.XImplementationLoader at java.net.URLClassLoader.findClass(URLClassLoader.java:381) at java.lang.ClassLoader.loadClass(ClassLoader.java:424) at java.lang.ClassLoader.loadClass(ClassLoader.java:357) ... 11 more The system has Java 1.8.0_151 and 1.7.0_161 installed, and it's not obvious from LO settings which JRE is being used. In a terminal "java -version" prints: OpenJDK Runtime Environment (IcedTea 3.6.0) (build 1.8.0_151-b12 suse-18.1-x86_64) OpenJDK 64-Bit Server VM (build 25.151-b12, mixed mode) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 Andreas Stieger <astieger@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |astieger@suse.com Assignee|bnc-team-screening@forge.pr |tchvatal@suse.com |ovo.novell.com | -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c1 Tomáš Chvátal <tchvatal@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |Ulrich.Windl@rz.uni-regensb | |urg.de Flags| |needinfo?(Ulrich.Windl@rz.u | |ni-regensburg.de) --- Comment #1 from Tomáš Chvátal <tchvatal@suse.com> --- You can try to run it in gdb and trace it to see why it crashed... without core this is not possible to determine what happened and what needs to be fixed. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c2 Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(Ulrich.Windl@rz.u | |ni-regensburg.de) | --- Comment #2 from Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> --- Created attachment 759211 --> http://bugzilla.opensuse.org/attachment.cgi?id=759211&action=edit strace of LibreOffice Writer start causing Java exception (In reply to Tomáš Chvátal from comment #1)
You can try to run it in gdb and trace it to see why it crashed... without core this is not possible to determine what happened and what needs to be fixed.
"it" being LO Writer did NOT crash. Not even sure anything crashed, but I see the exception stack trace every time I launch Writer. So I guess it's either a problem with IcedTea not being Oracle Java, or some packaging issue, or compatibility, etc. The gdb thing suggested probably won't work either, as "oowriter" is a shell script calling `/usr/lib64/libreoffice/program/soffice --writer "$@"':
file /usr/bin/oowriter /usr/bin/oowriter: POSIX shell script, ASCII text executable
/usr/lib64/libreoffice/program/soffice in turn is another shell script: usr/lib64/libreoffice/program/soffice: POSIX shell script, ASCII text executable So I hacked the latter script to inject an strace (not GDB). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c4 --- Comment #4 from Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> --- (In reply to Tomáš Chvátal from comment #3) They all look like valid JAR files:
file /usr/lib64/libreoffice/program/classes/* /usr/lib64/libreoffice/program/classes/commonwizards.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/form.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/hsqldb.jar: Zip archive data, at least v1.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/java_uno.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/js.jar: Zip archive data, at least v1.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/juh.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/jurt.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/query.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/reportbuilder.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/reportbuilderwizard.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/report.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/ridl.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/ScriptFramework.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/ScriptProviderForJava.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/ScriptProviderForJavaScript.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/sdbc_hsqldb.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/smoketest.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/table.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/unoil.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/unoloader.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/XMergeBridge.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR) /usr/lib64/libreoffice/program/classes/xmerge.jar: Zip archive data, at least v2.0 to extract Java archive data (JAR)
-- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c5 --- Comment #5 from Tomáš Chvátal <tchvatal@suse.com> --- If there is at least META-INF folder with MANIFEST.MF file it is valid jar, thats why I asked if they have reasonable size... anything bigger than couple hundred bytes... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c9 Stefan Münkner <cluster15@web.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cluster15@web.de --- Comment #9 from Stefan Münkner <cluster15@web.de> --- sorry, coming from via different route, I reopened an older bug, specifying that unoil.jar only contains MANIFEST.MF. The empty unoil.jar breaks report generation in base completely. Using the jars in /usr/lib64/libreoffice/program/classes/ from the previous version 5.3.5.2-3.4 fixes the error. Just replacing unoil.jar did not work. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c10 Wolfgang Bauer <wbauer@tmo.at> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |wbauer@tmo.at --- Comment #10 from Wolfgang Bauer <wbauer@tmo.at> --- (In reply to Stefan Münkner from comment #9)
Using the jars in /usr/lib64/libreoffice/program/classes/ from the previous version 5.3.5.2-3.4 fixes the error. Just replacing unoil.jar did not work.
Yes, ridl.jar is again incomplete as well. The situation seems to be exactly the same as in 13.2 3 years ago (bug#913042) ... And has been mentioned, the package in LibreOffice:5.4 is fine as well too, the problem again only exists in the update repo. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c11 --- Comment #11 from Tomáš Chvátal <tchvatal@suse.com> --- (In reply to Wolfgang Bauer from comment #10)
(In reply to Stefan Münkner from comment #9)
Using the jars in /usr/lib64/libreoffice/program/classes/ from the previous version 5.3.5.2-3.4 fixes the error. Just replacing unoil.jar did not work.
Yes, ridl.jar is again incomplete as well.
The situation seems to be exactly the same as in 13.2 3 years ago (bug#913042) ... And has been mentioned, the package in LibreOffice:5.4 is fine as well too, the problem again only exists in the update repo.
Weird problem is that if we do not run %check phase the unoil.jar and others are fine... :/ -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c12 --- Comment #12 from Wolfgang Bauer <wbauer@tmo.at> --- (In reply to Tomáš Chvátal from comment #11)
Weird problem is that if we do not run %check phase the unoil.jar and others are fine... :/
IIRC, that was the problem 3 years ago as well, wasn't it? This was the "fix" back then: --- libreoffice.spec (revision 1) +++ libreoffice.spec (revision 1) @@ -1013,11 +1013,9 @@ sed -i -e "s|@INSTALLDIR@|%{_libdir}/%{name}|" pyuno/source/module/uno.py scripting/source/pyprov/officehelper.py # just call make here as we added the jobs in configure +# The check phase is run here too if it is split with nocheck +# install of jars get broken sometimes +make VERBOSE=true -make build-nocheck - -%check -make unitcheck -make slowcheck %install make DESTDIR=%{buildroot} distro-pack-install Just mentioning... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c13 --- Comment #13 from Tomáš Chvátal <tchvatal@suse.com> --- (In reply to Wolfgang Bauer from comment #12)
(In reply to Tomáš Chvátal from comment #11)
Weird problem is that if we do not run %check phase the unoil.jar and others are fine... :/
IIRC, that was the problem 3 years ago as well, wasn't it? This was the "fix" back then: --- libreoffice.spec (revision 1) +++ libreoffice.spec (revision 1) @@ -1013,11 +1013,9 @@ sed -i -e "s|@INSTALLDIR@|%{_libdir}/%{name}|" pyuno/source/module/uno.py scripting/source/pyprov/officehelper.py
# just call make here as we added the jobs in configure +# The check phase is run here too if it is split with nocheck +# install of jars get broken sometimes +make VERBOSE=true -make build-nocheck - -%check -make unitcheck -make slowcheck
%install make DESTDIR=%{buildroot} distro-pack-install
Just mentioning...
Yep, but the problem now is we build the tests conditionally and only on 64b intel, that means this would not solve it this time. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c14 --- Comment #14 from Wolfgang Bauer <wbauer@tmo.at> --- (In reply to Tomáš Chvátal from comment #13)
Yep, but the problem now is we build the tests conditionally and only on 64b intel, that means this would not solve it this time.
Hm, maybe this would work? +%ifarch x86_64 +make +%else make build-nocheck +%endif -%check -# Run tests only on 64b intel as they are resource hogs -%ifarch x86_64 -make -%endif AIUI, that should still run the tests on x86_64, just in %build instead of %check... -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c15 Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|unoil.jar does not contain |unoil.jar does not contain |any content in 5.4.4.2 |any content in LibreOffice |update |5.4.4.2 update --- Comment #15 from Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> --- On comments 12, 13, and 14: Wouldn't it be preferable to to fix the check instead of disabling it? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c16 --- Comment #16 from Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> --- Sadly the bug is still present after security update to libreoffice-5.4.5.1-15.1.x86_64. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c17 --- Comment #17 from Wolfgang Bauer <wbauer@tmo.at> --- (In reply to Ulrich Windl from comment #15)
On comments 12, 13, and 14: Wouldn't it be preferable to to fix the check instead of disabling it?
Please read those comments again. In particular, changing "make build-nocheck" to just "make" should actually *enable* to run the tests... AFAICT, the problem only occurs when running the tests separately from the build phase. But strange enough, only in the Update repo. That said, maybe it relates now to the new compile options in Leap:42.3:Update? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c18 --- Comment #18 from Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> --- (In reply to Wolfgang Bauer from comment #17)
Please read those comments again.
You are right: I was confused, obviously. Sorry. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1079744 http://bugzilla.opensuse.org/show_bug.cgi?id=1079744#c19 --- Comment #19 from Wolfgang Bauer <wbauer@tmo.at> --- *** Bug 1083320 has been marked as a duplicate of this bug. *** -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com