2019-10-05 GnuCash IRC logs
01:14:39 *** Aussie_matt has quit IRC
01:31:49 *** fell has joined #gnucash
01:31:49 *** ChanServ sets mode: +o fell
01:33:59 *** fell_laptop has quit IRC
01:38:09 *** Mechtilde has joined #gnucash
03:03:50 *** bertbob has quit IRC
03:09:36 *** bertbob has joined #gnucash
03:09:36 *** ChanServ sets mode: +v bertbob
03:09:53 *** fell has quit IRC
03:46:45 *** marusich has quit IRC
03:48:46 *** sbluhm has joined #gnucash
03:51:51 *** Mechtilde has quit IRC
03:57:45 *** Mechtilde has joined #gnucash
04:00:16 *** sbluhm has quit IRC
04:14:22 *** gjanssens has joined #gnucash
04:14:22 *** ChanServ sets mode: +o gjanssens
04:14:30 <gjanssens> .
04:38:27 *** o01eg has quit IRC
04:53:22 *** fabior has joined #gnucash
04:59:06 *** fabior has quit IRC
06:29:19 *** Jimraehl1 has joined #gnucash
06:29:29 *** sbluhm has joined #gnucash
06:29:29 *** ChanServ sets mode: +v sbluhm
06:29:52 *** Jimraehl1 has left #gnucash
06:50:58 *** fell has joined #gnucash
06:50:58 *** ChanServ sets mode: +o fell
06:56:54 *** sbluhm has quit IRC
06:59:01 *** fell has quit IRC
07:16:36 *** FoxT has quit IRC
07:22:44 *** FoxT has joined #gnucash
07:23:05 *** oozer has joined #gnucash
07:32:05 *** lagertonne has quit IRC
07:32:16 *** lagertonne has joined #gnucash
07:35:04 *** lagertonne has joined #gnucash
07:35:29 *** lagertonne has quit IRC
07:35:45 *** lagertonne has joined #gnucash
07:36:55 *** lagertonne has joined #gnucash
08:56:33 *** CDB-Man_ has joined #gnucash
08:56:33 *** ChanServ sets mode: +v CDB-Man_
08:58:29 *** jervin has joined #gnucash
08:58:47 *** CDB-Man has quit IRC
09:01:19 *** bertbob has quit IRC
09:01:37 *** jervin has quit IRC
09:03:16 *** bertbob has joined #gnucash
09:03:16 *** ChanServ sets mode: +v bertbob
09:57:09 <Robert847> I have to reboot my computer to finish the Microsoft out-of-band update dated October 4 to see if that fixes the problem. Since I will be busy the next several hours it may be days before I get back here.
09:59:45 *** Robert847 has left #gnucash
10:57:05 *** fabior has joined #gnucash
11:27:03 *** mohave has quit IRC
11:27:45 *** mohave has joined #gnucash
11:36:47 *** phoenix has joined #gnucash
11:45:21 *** phoenix has quit IRC
11:57:53 *** tienne has joined #gnucash
11:58:32 *** oozer has quit IRC
11:58:35 *** tienne has joined #gnucash
12:11:17 <gjanssens> jralls: I'm slowly advancing my separation of guile code from libgnucash
12:11:31 <gjanssens> In the engine I have a question though
12:12:24 <gjanssens> There's https://github.com/Gnucash/gnucash/blob/master/libgnucash/engine/engine-deprecated.h
12:12:44 <gjanssens> Which deprecates C functions for use in scheme code
12:13:22 <gjanssens> It uses scm_c_issue_deprecation_warning to emit deprecation messages
12:13:51 <gjanssens> Should I also just move this code out to the bindings ?
12:14:49 <gjanssens> These are C functions really so theoretically we also want to inform C consumers of these functions about future removal
12:15:23 <gjanssens> I doubt we currently have any C consumers other than gnucash, so this is mostly hypothetical.
12:16:46 <gjanssens> If I move these two file to bindings, only guile consumers will get some deprecation guidelines and in C they would just no longer exist resulting in compile failures for such theoretical consumers
12:17:14 <gjanssens> What could be a proper way to handle this in the future ?
12:18:05 <gjanssens> In fact, how does one mark a function as deprecated in pure C ? I don't know the conventions for that.
12:32:41 *** mdf_ has joined #gnucash
12:34:22 *** mdf has quit IRC
12:43:52 *** lagertonne has quit IRC
12:52:10 *** mdf_ has quit IRC
12:52:31 *** mdf has joined #gnucash
12:52:31 *** ChanServ sets mode: +v mdf
13:05:37 *** oozer has joined #gnucash
13:05:46 *** guak has joined #gnucash
13:12:08 <jralls> gjanssens: We agreed last year that we weren't going to do deprecation on C functions until we had redone the C/C++ API into something sensible and coherent. But if you like we can apply __attribute__((deprecated)) in C and (once we change the C++ standard to 14 or 17) [[deprecated]] in C++ to deprecated functions.
13:13:41 *** Cuare has quit IRC
13:14:19 <jralls> Those timespec C functions aren't used at all in GnuCash, chris replaced all of the timespec instances with time64 last year. All any of them do is wrap the equivalent time64 function so by all means move them out to bindings.
13:16:55 <gjanssens> Ok
13:17:25 <gjanssens> Agreed on the agreement also btw. I seem to have forgotten that bit.
13:20:16 *** tienne has quit IRC
13:21:57 *** tienne has joined #gnucash
13:24:19 <jralls> On the marginally related subject of C++17, I tried that yesterday on MacOS and found an interesting wrinkle that I sort-of knew but hadn't really thought about:
13:26:48 <jralls> Boost needs to be compiled with std=C++17 in order for GnuCash to compile that way. Since we rely on packagers to provide boost for us (except on MacOS) we'll need to test widely before we actually commit that change.
13:28:40 *** bertbob has quit IRC
13:31:46 <jralls> Ubuntu-18.04 seems to be OK.
13:33:26 <gjanssens> What failure do you see if it's not ? Just to have a quick test metric.
13:42:26 <jralls> gnc-filepath-utils fails to link because of boost::filesystem name mangling differences.
13:45:34 <gjanssens> Ok
13:45:50 <gjanssens> I'll give it a try here as well
13:46:26 * gjanssens is currently fighting the Windows build over his most recent bindings migrations
13:49:38 <jralls> My master-branch guile has decided to pack it in today as well.
13:50:17 <jralls> I'm a bit worried about the MinGW build because we've had to hold back boost.
13:51:45 <gjanssens> Hmm, right
13:54:09 <gjanssens> Can we build boost ourselves on Windows in the worst case scenario ? Our would the DLL dependencies on older ICU make that impossible ?
13:54:49 *** fell has joined #gnucash
13:54:49 *** ChanServ sets mode: +o fell
13:55:29 <jralls> We can, though if the compiler flags are different we might have to build ICU as well.
13:55:48 <jralls> The change to CMakeLists.txt I'm using is https://gist.github.com/jralls/8a2afd5993b96fa5834ed014d0861b55
13:55:58 *** tejas has joined #gnucash
14:11:46 *** frakturfreak has joined #gnucash
14:14:23 *** fabior has quit IRC
14:15:55 <gjanssens> jralls: C++17 works fine on Fedora 30. No build errors in a clean build directory
14:15:59 *** lagertonne has joined #gnucash
14:16:36 *** lagertonne has quit IRC
14:16:40 *** lagertonne has joined #gnucash
14:17:57 <jralls> gjanssens: That's interesting, because I just tested the same and am stopped at a -Wunsed-result error in gnc-sql-backend.cpp.
14:20:24 <gjanssens> Huh, that's weird
14:21:07 <jralls> It's a legitimate error, too: the line is m_postload_commodities.empty(); and should be m_postload_commodities.clear();.
14:21:16 <gjanssens> What cmake command line did you use ?
14:21:27 <gjanssens> Mine is
14:21:29 <gjanssens> CC=clang CXX=clang++ cmake -GNinja -D WITH_PYTHON=ON -D CMAKE_INSTALL_PREFIX=/home/janssege/Development/installs/gnucash/master -DCMAKE_BUILD_TYPE=Debug /home/janssege/Development/gnucash/master
14:21:38 <gjanssens> Note I'm currently playing with clang
14:22:18 <jralls> No doubt the difference, I'm using gcc on Fedora.
14:24:20 <jralls> Nice to know, though, that boost built with g++ works with gnucash built with clang.
14:24:38 <gjanssens> Indeed
14:25:05 <gjanssens> I see the same error when building with gcc
14:28:03 <jralls> The thing about that is that that error goes back to pre-2.5.0. It seems to be tripped by gcc implementing -Wunused-result by sticking a virtual [[no-discard]] on every function. That's a C++17 feature.
14:32:19 *** bertbob has joined #gnucash
14:32:19 *** ChanServ sets mode: +v bertbob
14:42:24 <jralls> Same error on Arch, no surprise.
14:58:55 <jralls> Fixed in maint, merged up to master.
15:02:39 <gjanssens> Saw your commits pass by
15:07:58 <jralls> So that leaves, what? Windows, OpenSuSE, and the BSDs?
15:10:46 <jralls> I'll do Windows as soon as guile finishes rebuilding. At least I hope that rebuilding will get guile to behave enough to build GnuCash. I guess I should set up VMs for the other two.
15:21:47 *** frakturfreak has quit IRC
15:24:14 *** frakturfreak has joined #gnucash
15:24:14 *** ChanServ sets mode: +v frakturfreak
15:39:48 *** TommyT has joined #gnucash
15:39:48 *** ChanServ sets mode: +v TommyT
15:42:35 *** guak has quit IRC
15:53:26 *** fell has quit IRC
15:58:52 *** gjanssens has quit IRC
15:59:30 *** sbluhm has joined #gnucash
15:59:30 *** ChanServ sets mode: +v sbluhm
16:02:35 *** Mechtilde has quit IRC
16:17:08 <jralls> @tell gjanssens Windows is OK with C++17.
16:17:08 <gncbot> jralls: The operation succeeded.
16:19:38 *** Gerd has joined #gnucash
16:27:00 *** sbluhm has quit IRC
16:28:29 *** Robert847 has joined #gnucash
16:38:09 <Gerd> Hello, I upgraded from 2.6.19 to 3.7-2019-09-23-git-3.7-60-gee260d8e3+ to follow the PSD2 regulations but now I am not able to use the mouse. Gnucash ignores completely every mouse action. I am working with Win 10.0.17763.737. Could anybody give me a hint how to solve this ?
16:41:59 <Robert847> Gerd, Have you applied the October 4 (I think) out-of-band fix for their recent problems?
16:42:19 <Robert847> Windows 10
16:43:09 <Robert847> I understand that you need to re-boot afterwrds, too
16:44:17 <Gerd> Thank you for your answer, but I am not so experienced. What do you mean with October 4 out of band fix?
16:44:47 <Gerd> Yes, I rebooted after installing the software.
17:14:16 *** Gerd has quit IRC
17:23:29 *** Gerd has joined #gnucash
17:27:41 <Gerd> Hello, I upgraded to the 3.7-2019-10-01-git-3.7-91-g34bb47e23+ to follow the PSD2 regulations, but with all versions above 2.6.19 Gnucash will not respond on any mouse action.
17:49:04 *** tienne has quit IRC
17:49:22 *** tienne has joined #gnucash
18:22:28 <jralls> Gerd: Does that mean that you've tried the regular GnuCash 3.7 release to test the mouse? (Won't work with your bank, of course...)
18:27:44 *** frakturfreak1 has joined #gnucash
18:27:49 *** frakturfreak has quit IRC
18:50:28 *** Gerd has quit IRC
19:01:03 <Robert847> Hi, after installing the Windows 7 fix, GnuCash 2.6.19 still crashes in Win 7 when opening a certain data file that it had previously opened on September 11. I also went back to Ubuntu 18.04 where I have the same version (almost) of GnuCash which does open almost the same data file except that it has about 3 weeks' more data in it.
19:04:38 *** frakturfreak1 has quit IRC
19:05:28 <Robert847> I think the Windows GnuCash data file dated September 11, 2019 has oodles of reminders, but I think that Gnucash crashed on something else. This time the last two items in the trace file are:
19:05:29 <Robert847> * 15:32:52 CRIT <GLib-GObject> g_object_ref: assertion 'G_IS_OBJECT (object)' failed
19:05:29 <Robert847> * 15:32:52 CRIT <Gdk> gdk_drawable_get_size: assertion 'GDK_IS_DRAWABLE (drawable)' failed
19:05:34 <jralls> Robert847: What Windows 7 fix is that?
19:07:07 <jralls> The trace file looks like it's hitting an assertion in Gdk (the backend of Gtk that adapts to the Operating System/Window Manager).
19:08:41 <Robert847> That is the October 3, 2019—KB4524157 (Monthly Rollup)
19:09:53 <Robert847> I now see that microsoft has issued anither fix. I'll get the details
19:09:58 <jralls> So a Microsoft security patch. That might be a problem.
19:11:38 <Robert847> The newest one is just the daily security essetials update
19:17:59 <Robert847> Windows 7 had several other patches since September 11, but the one that worried me was the one that supposedly fixed IE but broke printing in Windows 10, 7 and 8.1 that they issued in late September, which I thought the Oct 3 patch was supposed to fix
19:21:27 <Robert847> I am not out of business, since my Ubuntu instance is still working, but it is one less tool because that instance is now the only one that can find the current data file on my LAN's NAS.
19:22:50 <jralls> Plus Gtk and therefor GnuCash seems to be broken.
19:24:10 <jralls> Is there anything in the event log relating to the crash? I'd suspect something to do with system protection.
19:25:05 <Robert847> I will have to do a little fishing to find that. Give me a few minutes...
19:28:19 <Robert847> Faulting application name: gnucash.exe, version: 0.0.0.0, time stamp: 0x5a36e48c
19:28:20 <Robert847> Faulting module name: libgdk-win32-2.0-0.dll, version: 2.24.24.0, time stamp: 0x5456cd8a
19:28:20 <Robert847> Exception code: 0xc0000005
19:28:20 <Robert847> Fault offset: 0x000079e4
19:28:20 <Robert847> Faulting process id: 0x21a4
19:28:20 <Robert847> Faulting application start time: 0x01d57bbaa1be79f2
19:28:20 <Robert847> Faulting application path: C:\Program Files (x86)\gnucash\bin\gnucash.exe
19:28:20 <Robert847> Faulting module path: C:\Program Files (x86)\gnucash\bin\libgdk-win32-2.0-0.dll
19:28:22 <Robert847> Report Id: 4e015ce2-e7af-11e9-a33a-34f39aa0eb22
19:29:47 <Robert847> that is in the application error log
19:30:33 <jralls> That confirms that whatever went wrong in is Gdk, but it doesn't shed any light on what exactly went wrong.
19:31:36 *** mohave has joined #gnucash
19:32:33 <Robert847> Does GnuCash use the same GDK in release 3.7
19:33:38 <jralls> No, that's the big change, remember? We had to switch to Gtk3 because the Linux distros were dropping WebKit1 and WebKit2 doesn't support Gtk2.
19:35:50 <Robert847> OK. I think my Ubuntu 18.04 may still be using gtk2????
19:37:35 <Robert847> that is using 2.6.19 This copy was built from rev c1b5e6c8d+ on 2018-04-09
19:38:02 <jralls> Yeah, we released GC3.0 about a week before they released 18.04 so they went with 2.6.19.
19:38:54 *** oozer has quit IRC
19:41:00 <jralls> Debian had dropped WebKit1 from testing, but didn't release it as Buster until last Spring. Fedora and Arch dropped WebKit1 in 2017. We were off of Arch for about a year, but the Fedora Packager did a workaround where he put WebKit1 in the GnuCash rpm.
19:41:41 <jralls> So it turns out that Exception code is useful: https://social.technet.microsoft.com/Forums/windows/en-US/9e83c5fc-caa4-4d64-82a2-342bda0b8673/exception-code-0xc0000005?forum=w7itproappcompat.
19:42:02 <jralls> Which says it's a memory protection violation.
19:42:42 <jralls> Several sites suggest that it might be an AV problem.
19:44:14 <jralls> Did it put up a dialog box when it crashed or just disappear?
20:09:15 <Robert847> No, it just disappeared
20:16:47 <Robert847> I have not tried opening smaller data files other than opening GnuCash with the --nofile option, which worked. I hope to start testing release 3.7 soon in Virtualbox to see how it works for me.
20:25:09 *** tienne has quit IRC
20:28:04 <Robert847> Anyway, there won't be a fix, so if it isn't easy to stop banging my head on the wall, I will still have a headache.
20:30:35 <Robert847> Thanks for listening.
20:48:16 *** TommyT has quit IRC
20:55:02 *** Robert847 has left #gnucash
20:57:03 *** oozer has joined #gnucash
21:10:24 *** TownsendHardware1 has joined #gnucash
21:12:35 *** TownsendHardware has quit IRC
21:12:35 *** TownsendHardware1 is now known as TownsendHardware
21:31:29 *** oozer has quit IRC
21:45:54 *** TownsendHardware has quit IRC
23:55:04 *** fell has joined #gnucash
23:55:05 *** ChanServ sets mode: +o fell