GNOME Bugzilla – Bug 766028
crash on manual addition of price entry in Price Editor
Last modified: 2018-06-29 23:48:58 UTC
This problem is new in 2.6.12, both Windows and Linux(Ubuntu) OSes. The Price Editor works fine in terms of updating fund price data with F::Q and saving the data. Data entry in manual updates of fund prices works fine and can be "applied" so that another entry is possible. However, the program crashes when pressing "OK" to close the manual entry window.
Can't replicate on OSX, and away from other OSes until next week. Can you run under gdb on Linux and post a stack trace? http://wiki.gnucash.org/wiki/Stack_Trace.
Here is the windows report Problem signature: Problem Event Name: APPCRASH Application Name: gnucash.exe Application Version: 0.0.0.0 Application Timestamp: 56f898df Fault Module Name: msvcrt.dll Fault Module Version: 7.0.9600.17415 Fault Module Timestamp: 54504b2e Exception Code: c0000005 Exception Offset: 0001bf8f OS Version: 6.3.9600.2.0.0.768.101 Locale ID: 1033 Additional Information 1: 5861 Additional Information 2: 5861822e1919d7c014bbb064c64908b2 Additional Information 3: 1318 Additional Information 4: 13181ae637ac2592fac0ca2dd662cdcc Read our privacy statement online: http://go.microsoft.com/fwlink/?linkid=280262 If the online privacy statement is not available, please read our privacy statement offline: C:\WINDOWS\system32\en-US\erofflps.txt
Correction to the original report: The reported problem is specific for windows - Linux works ok (my apologies for incorrectly reporting a problem in Linux. Ulrich
Unfortunately we use MinGW and so M$ crash reports aren't helpful. This will have to wait until I get home next week and can try to replicate it on a Windows VM unless you can generate a stack trace in the meantime.
I use Linux ubuntu 16:04 I get a segfault every time I try to add a new price using the price editor. Here is the gdb backtrace: Thread 1 "gnucash" received signal SIGSEGV, Segmentation fault. 0x00007ffff72cbfa3 in gnc_price_ref (p=0x1) at gnc-pricedb.c:288 288 p->refcount++; (gdb) bt
+ Trace 236234
My experience (windows) was that the "OK" button worked as long as the "apply" button wasn't selected ... meaning only 1 price addition could be applied at a time.
(In reply to David Wilson from comment #5) > I use Linux ubuntu 16:04 > > I get a segfault every time I try to add a new price using the price editor. > > Here is the gdb backtrace: > > > Thread 1 "gnucash" received signal SIGSEGV, Segmentation fault. > 0x00007ffff72cbfa3 in gnc_price_ref (p=0x1) at gnc-pricedb.c:288 > 288 p->refcount++; > (gdb) bt As reported by Donald McAdams the crash (on Linux) occurs on clicking the Apply button, but seems to correctly add the price with the "OK" button. Also I should have stated that I am using GnuCash 2.6.12, built from rev ed11f6d+ on 2016-05-07.
In reply to Donald McAdams with windows: Thanks for letting me know about the problem with the "apply" button. The Price Editor works fine with just 1 manual edition and the pressing "ok" without using "apply". In a previous version (2.4), the "apply" button worked fine and several manual editions could be made at a time.
I've pushed a fix that will be in tomorrow's nightly maint build, please test. Work around is to use "Cancel" instead of "OK" if the Apply button is greyed out.
John, I am a newbie here. Will the nightly maint build be on github? Ulrich
(In reply to Ulrich Hopfer from comment #10) > John, > > I am a newbie here. Will the nightly maint build be on github? Ulrich No, Win32 will be in http://code.gnucash.org/builds/win32/maint/ and an rpm should be at https://copr.fedorainfracloud.org/coprs/gjanssens/gnucash-maint/builds/ but that seems to be broken--the last one was over 3 weeks ago.
The rpm builds were broken because my login token had expired. Tried to fix that last week but failed. I just now realized I tried to fix it on the wrong server... I have manually started the rpm builds a minute ago from the most recent maint and master branch. They should be available in an hour or so.
John, the latest build in http://code.gnucash.org/builds/win32/maint/ is from 4/17/2016. I do not see your build from 5/10/16.
It would be the 5/11 build which failed from a change in the documentation. I'll look at it tomorrow, which means the nightly will be Friday or Saturday.
(In reply to John Ralls from comment #9) > I've pushed a fix that will be in tomorrow's nightly maint build, please > test. > > Work around is to use "Cancel" instead of "OK" if the Apply button is greyed > out. I think there were a few related things going on here as I had a different symptom. To test I built master + cherry pick your fix [1] on linux and it is behaving as I expect. Be nice to have a Win test as and when so other people can check if their symptom is fixed too but I think it is right now. [1] a first for me, I still find chunks of git confusing
The symptoms do vary, apparently depending upon how memory was allocated and freed; that will in turn depend on the versions of (g)libc and glib. There was no corruption at all on OS X. Why did you cherry pick to master instead of just building maint? BTW, for those waiting for a Window nightly build, that build is still broken due to a completely unrelated problem with building the documents. I'm working on it but the excruciating slowness of the Win32 build VM makes it an excessively long task.
(In reply to John Ralls from comment #16) > The symptoms do vary, apparently depending upon how memory was allocated and > freed; that will in turn depend on the versions of (g)libc and glib. There > was no corruption at all on OS X. That makes sense, I had differing symptoms with my main book and a tiny test book with half a dozen prices. > Why did you cherry pick to master instead of just building maint? 1. no (or known) crap if I used it to change my live book 2. it was a good chance for me to think some more about git rather than just using it to get and build (and I know that get isn't the git word) > BTW, for those waiting for a Window nightly build, that build is still > broken due to a completely unrelated problem with building the documents. > I'm working on it but the excruciating slowness of the Win32 build VM makes > it an excessively long task. Coincidentally and tangentially I'm having another go at building under Win64 Cygwin because of Win32 VM slowness (and Win64 possibly not working at all). Will report in dev as and when.
I thought that I would test this for you but I can only find the official release release build (2016-03-27) right now. Both the nightly Windows builds are still stuck in April. For me in Windows 7 the official build crashed when I opened the price editor, selected an existing price for an existing security, clicked on Add then in the Add dialog, clicked OK without making any other changes in that dialog.
Created attachment 328053 [details] Windows Trace from reported crash
We finally have a Windows nightly build to test on: http://code.gnucash.org/builds/win32/maint/gnucash-2.6.12-2016-05-17-git-c1ad615+-setup.exe Fedora RPMs are at https://copr-be.cloud.fedoraproject.org/results/gjanssens/gnucash-maint/fedora-23-x86_64/00257328-gnucash/
John, thanks. The new windows nightly build solved the problem of a crash on manual addition of price entry in Price Editor.
Yay. It will of course also be in the next normal release.
*** Bug 764732 has been marked as a duplicate of this bug. ***
*** Bug 790777 has been marked as a duplicate of this bug. ***
GnuCash bug tracking has moved to a new Bugzilla host. This bug has been copied to https://bugs.gnucash.org/show_bug.cgi?id=766028. Please update any external references or bookmarks.