Created attachment 101546 [details] Crash report, 4.3.0.1 x86-64 I am running OS X 10.9.3. I was originally going to check if the linear solver was available (it wasn't in 4.2.5.2 for example), but now when opening the solver (Tools -> Solver), even for an empty spreadsheet and regardless of enabling the JRE or not, it crashes. Attached is the crash report. Note that I used x86-64 "for 10.8 or higher".
I fail to reproduce, tested using Mac OSX 10.9 with LibreOffice Version: 4.2.5.2 Build ID: 61cb170a04bb1f12e77c884eab9192be736ec5f5 also the x86-64 version. Kind regards, Joren
I can reproduce using Mac OSX 10.9 with LibreOffice 4.3.0.2 r2. Kind regards, Joren
Can't reproduce it on Linux 4-3 and master.
can reproduce it with Windows 8.1 on 4.3.0.2 when Java is disabled. I get the notification that Java should be enabled - and then crash After next start I enabled java and Solver starts and ends without crash
I cannot reproduce it anymore with 4.3.0.3 (Windows 8.1)
Still present as originally described with 4.3.0.3 (RC3), OS X x86-64 (running 10.9.4)
It's also present in 4.3.0.4. Same details (10.9.4, x86-64)
*** Bug 82086 has been marked as a duplicate of this bug. ***
Do the people who can report this issue run daily builds? The check for conformance according to the exception specification should only be enabled in developer builds (surely not in release builds).
(In reply to comment #9) > Do the people who can report this issue run daily builds? The check for > conformance according to the exception specification should only be enabled > in developer builds (surely not in release builds). I see this too, with the release build for 4.3.1.2, running on Win 7 64-bit.
(In reply to comment #9) > Do the people who can report this issue run daily builds? The check for > conformance according to the exception specification should only be enabled > in developer builds (surely not in release builds). I have tested again with master_nightlies~2014-09-25_11.30.29_LibreOfficeDev_4.4.0.0.alpha0_MacOS_x86-64 on OS X 10.9.5, and the issue persists. I don't know what you're referring to or where to find it, and whether it will appear in the diagnostic report generated by OS X or somewhere else (there is nothing printed to the terminal by running LibreOfficeDev.app/Contents/MacOS/soffice). (I'm not a developer yet, sorry.) Again, going off what was the case when this was reported, there is probably a regression somewhere between 4.2.5.2 (maybe even 4.2.6--c.f. bug 82086) and 4.3.0.1. I can't complete the bibisect anytime soon, someone else can go right ahead.
Created attachment 106883 [details] Crash report, 4.4.0.0alpha0 x86-64
Unable to create a bibisect on Linux, sorry. Cannot replicate the crash.
Still occurs on OSX with LO 4.3.4.1 and recent 4.5 master. There are older bibisect repositories for Mac but unfortunately they stop before this bug was introduced, so setting this to Whiteboard:notBibisectable for now.
Created attachment 112904 [details] Fragment of SAL_LOG=yes output The attached fragment of SAL_LOG=yes output suggests that we are trying to load libCbc.3.dylib from the wrong place when Solver is invoked
*** Bug 84723 has been marked as a duplicate of this bug. ***
fixed on master towards LO 4.5 with <http://cgit.freedesktop.org/libreoffice/core/commit/?id=f694ab5198587dcacd8c40c9095e7bc9435cbff2> "tdf#80370: Fix Mac OS X install names of external/{coinmp,lpsolve}"
Migrating Whiteboard tags to Keywords: (notBibisectable) [NinjaEdit]