2017-09-22 GnuCash IRC logs
00:44:47 *** gncbot has joined #gnucash
00:48:48 *** gncbot has joined #gnucash
01:30:26 *** jotrago has quit IRC
01:30:32 *** jotrago has joined #gnucash
01:39:58 *** frakturfreak1 has joined #gnucash
01:41:51 *** frakturfreak has quit IRC
01:49:13 *** Mechtilde has joined #gnucash
01:55:14 *** Mechtilde has quit IRC
02:35:13 *** Mechtilde has joined #gnucash
02:59:06 *** gjanssens has joined #gnucash
02:59:07 *** ChanServ sets mode: +o gjanssens
03:01:06 <gjanssens> .
03:04:00 *** mrklintscher1111 has quit IRC
03:18:53 *** mrklintscher1111 has joined #gnucash
03:23:59 *** Mechtilde has quit IRC
03:29:07 *** christopherlam has joined #gnucash
03:29:32 <christopherlam> gjanssens: just wondering if you know who's the wizard for things such as QOFQuery
03:30:34 <christopherlam> trying to figure out how it works when the sorting is left as default. does 'register-order' mean 'order by date of entry' ???
03:41:03 *** Groan has quit IRC
03:41:12 *** Groan has joined #gnucash
03:44:22 *** bertbob has quit IRC
03:47:06 *** fabior has joined #gnucash
03:47:18 *** mreich has joined #gnucash
03:47:43 *** bertbob has joined #gnucash
03:48:09 *** Groan1 has joined #gnucash
03:48:43 *** bertbob has quit IRC
03:48:48 *** Groan has quit IRC
03:49:04 *** christopherlam has quit IRC
03:57:08 *** Mechtilde has joined #gnucash
03:57:17 *** Groan1 has quit IRC
04:02:10 *** bertbob has joined #gnucash
04:05:07 *** gncbot has joined #gnucash
04:06:06 *** hoijui has joined #gnucash
04:10:19 *** hoijui has quit IRC
04:10:22 *** g5pw has joined #gnucash
04:20:05 *** bertbob has joined #gnucash
04:24:33 *** gncbot has joined #gnucash
04:36:13 *** hoijui has joined #gnucash
05:17:28 *** User has joined #gnucash
06:01:19 *** pilotauto has quit IRC
06:39:21 *** fekepp has joined #gnucash
06:40:57 *** fekepp has quit IRC
06:46:16 *** Jimraehl1 has joined #gnucash
06:47:04 *** Jimraehl1 has quit IRC
06:58:01 *** fbruetting has joined #gnucash
06:59:14 *** fbruetting has quit IRC
07:07:11 *** fbruetting has joined #gnucash
07:13:45 *** fabior has quit IRC
07:28:13 *** rickoehn has joined #gnucash
07:30:58 *** fabior has joined #gnucash
07:31:00 *** jotrago has quit IRC
07:42:01 *** fabior has quit IRC
07:53:47 *** hoijui has quit IRC
07:53:56 *** Jacques has joined #gnucash
07:58:39 *** fbruetting has quit IRC
08:17:10 *** Jacques has quit IRC
08:55:48 *** gncbot has joined #gnucash
09:25:40 *** hoijui has joined #gnucash
09:30:18 *** fabior has joined #gnucash
10:01:35 *** Mechtilde has quit IRC
10:04:46 *** kael has joined #gnucash
10:06:38 *** Mechtilde has joined #gnucash
10:09:21 *** fbruetting has joined #gnucash
10:20:03 *** kael has quit IRC
10:39:58 *** ArtGravity has joined #gnucash
11:15:04 *** storyjesse has quit IRC
11:18:05 *** jotrago has joined #gnucash
11:29:38 *** mrklintscher1111 has quit IRC
11:41:40 *** jotrago has quit IRC
12:25:57 *** codesmythe has joined #gnucash
12:33:50 *** mreich has quit IRC
12:59:52 *** jotrago has joined #gnucash
13:03:00 *** hoijui has quit IRC
13:11:29 *** shoonya has joined #gnucash
13:25:51 *** Mechtilde has quit IRC
13:36:26 <jralls> @tell christopherlam If anyone still active is an expert on QOFQuery it would be warlord.
13:36:26 <gncbot> jralls: The operation succeeded.
13:50:14 *** Groan has joined #gnucash
13:52:26 *** fabior has quit IRC
14:03:13 *** warlord has joined #gnucash
14:04:30 *** Mechtilde has joined #gnucash
14:34:32 *** fabior has joined #gnucash
14:35:47 *** Groan has quit IRC
14:38:41 <jralls> gjanssens: A build on Debian 8 just failed to build report-selectors.scm because boost::filesystem threw. https://pastebin.com/qsFChWdL
14:41:16 *** shoonya has quit IRC
14:48:12 *** Mechtilde has quit IRC
15:06:01 <gjanssens> jralls: :( Looks like I have some additional debugging to do
15:06:45 <gjanssens> FWIW I can reproduce it locally on Fedora as well.
15:06:45 <gjanssens> So my last patch in that area was clearly wrong
15:06:58 <jralls> Ah, good. I'd start with why does report-selectors.scm need the directory to build?
15:07:27 <jralls> Well, good that you can reproduce it. Not good that your patch was messed up.
15:12:53 <gjanssens> I did that exercise before. Compilation of report-selectors.scm indirectly calls one of the variants of gnc_build_userdata_path.
15:13:18 <gjanssens> I don't remember which one, as that was irrelevant at that moment.
15:14:02 <gjanssens> All of the gnc_build_userdata_xyz variants will check whether the filepaths have been initialized. If not, they will do so.
15:15:08 <gjanssens> The initialization code potentially also creates what we used to call dotgnucashdir, but now .local/share/GnuCash on linux.
15:15:31 <gjanssens> I don't want the build to create these directories though, I only want them created at runtime, so there's code
15:15:42 <gjanssens> to distinguish between these two.
15:16:00 <gjanssens> Clearly my logic in that area is flawed.
15:16:50 <gjanssens> I wanted to stay as close as possible to the original code which also does the path checks and then creates the directories
15:17:25 <gjanssens> Except I didn't want my code to create directories during build.
15:17:30 <jralls> Well, the logic about not creating them is working, but it's not disabling the attempt to *use* them, hence the throw.
15:19:08 <gjanssens> True. This is because I wanted to be prudent. I don't know if the rest of the code handles a non-existing userdata dir. To be sure I'd have to trace each and every call to all of the variations of gnc_build_userdata_XXX
15:19:53 <gjanssens> While doing this code I thought it to be safer to always come up with some valid directory. In the worst case somewhere in tmp.
15:21:00 <gjanssens> Regardless, I'll clean it up. But I won't have time until after the weekend unfortunately.
15:21:18 <gjanssens> The workaround is to create the missing directory for now as you already found.
15:21:57 <jralls> OK. Webkit is still giving me grief on both Mac and Win and I've got the 2.6.18 release for this weekend plus an all-day event tomorrow.
15:22:13 <gjanssens> Right.
15:22:34 <gjanssens> I have been poking at the Windows build on and off in the past week
15:22:43 <jralls> I don't see a release of 2.7.0 happening until next week sometime even without this problem.
15:23:04 <gjanssens> Unfortunately I didn't even manage to run build_package.ps1 to successful completion
15:23:36 <gjanssens> I have had several errors, the last one gnucash not building because of a non-buildable aqbanking
15:24:15 <gjanssens> So while I wished otherwise, I have not been able to help you make progress there.
15:24:22 <jralls> What was the error with AQB?
15:24:51 <gjanssens> Let me check, it's my most recent build which I hadn't evaluated yet
15:29:12 <gjanssens> aqbanking is not building because a non-buildable xmlsec
15:29:56 <gjanssens> And that fails due to a failure to fork which looks like an intermittent error to me.
15:30:33 <gjanssens> Sadly each rebuild takes a long time and usually longer than the timespan I have available.
15:30:43 <gjanssens> So my progress is slow
15:30:47 <jralls> Failure to fork? I haven't seen that one before.
15:32:01 <jralls> Once jhbuild has installed something it doesn't try to rebuild it, so if you can keep poking at it until all of the dependencies are built it will take only a few minutes to rebuild GnuCash.
15:32:33 <gjanssens> https://pastbin.com/PVPe7gpC
15:33:06 <gjanssens> Well I deliberately wanted to test build_package.ps1
15:33:24 <gjanssens> It seems to rebuild at least a number of packages each time and again on my system.
15:33:55 <gjanssens> But don't spend your time trying to debug my local builds. There are more pressing matters :)
15:34:35 * gjanssens is currently debugging contacts sync on his phone, totally not gnucash related :(
15:35:23 <jralls> Looks like it got a bad path: 'D:?\D:\gcdev64\msys2\usr\bin\sh.exe'
15:38:35 <gjanssens> I wonder where it got that one from. I just started '.\build_package.ps1 -branch master' from the gnucash-on-windows-git\buildserver directory
15:39:18 <gjanssens> I have installed everything under d:\gcdev64 instead of c:\gcdev64 so the "d:" is correct. The double-so-'d:' is not.
15:39:28 <gjanssens> Anyway I'll pick it up later
15:39:42 <gjanssens> Is there much to do still for 2.6.18 ?
15:43:37 <jralls> I got everything more-or-less ready yesterday. The only commit since was your add of christopherlam's income-gst-statement.scm to POTFILES.skip... though it didn't seem to bother make distcheck yesterday.
15:44:38 *** CDB-Away has quit IRC
15:44:52 *** CDB-Man_ has quit IRC
15:46:22 <jralls> BTW Debian 9 has a new automake that does an out of source build in the dist dir for the check and that breaks the intltools-update check because the sources wind up in _build/sub and those are obviously not in POTFILES. I used F25 for the trial distcheck yesterday to avoid the problem.
15:50:20 *** CDB-Man has joined #gnucash
15:50:44 *** CDB-Away has joined #gnucash
15:52:03 <gjanssens> hmm, that commit was a mistake. I had reverted it, but it looks like I forgot to push that revert.
15:52:27 <gjanssens> Can I still do this ? Or do you prefer me to hold off until after release ?
15:54:25 <gjanssens> jralls: ^
15:55:55 *** CDB-Man_ has joined #gnucash
15:56:06 *** CDB-Away_ has joined #gnucash
15:56:20 *** CDB-Away has quit IRC
15:56:23 <jralls> No, go ahead. I was about to say that if you don't have anything more to commit I'll go ahead and tag the release, but then I remembered that there's one more thing I want to look at on https://bugzilla.gnome.org/show_bug.cgi?id=787572 first.
15:56:31 *** CDB-Man has quit IRC
15:57:36 <gjanssens> jralls: ok, done
16:04:17 *** hoijui has joined #gnucash
16:13:20 *** mrklintscher1111 has joined #gnucash
16:18:43 *** hoijui has quit IRC
16:24:02 *** gjanssens has quit IRC
17:14:15 *** fabior has quit IRC
17:18:01 <jralls> codesmythe: The whole automake thing in your cmake dist isn't working out very well: Debian has automake1.11 but it's broken and Fedora seems to offer only the current automake.
17:20:36 <jralls> Unfortunately in spite of it being broken and apparently everyone knowing that it's broken it's still the only old automake that Debian puts in their repo, so one must either hand-install a non-broken automake-1.11 or use an older distro that provides automake < 1.15.
17:28:07 <codesmythe> jralls: OK. Recommendations on how I should handle the fact at distcheck won't pass on 1.15?
17:28:53 <jralls> What goes wrong?
17:29:26 <jralls> IIUC you're not using automake for distcheck, just to generate configure and Makefile.in.
17:29:42 <codesmythe> The thing you mentioned earlier on Debian 9. The out of source build breaks things.
17:30:12 <codesmythe> Right. I was doing automake distcheck but then took it out.
17:30:32 <jralls> So with it out can't the check for automake go too?
17:30:33 <codesmythe> So I guess I can take the 1.15 check out also.
17:30:47 <codesmythe> yeah, probably.
17:43:29 *** User has quit IRC
17:52:55 *** codesmythe has quit IRC
18:16:09 *** codesmythe has joined #gnucash
18:17:54 <jralls> @tell codesmythe: Seems to work. I've committed and will push with the release commits.
18:17:54 <gncbot> jralls: The operation succeeded.
18:18:24 <jralls> @tell codesmythe Seems to work. I've committed and will push with the release commits.
18:18:24 <gncbot> jralls: The operation succeeded.
18:19:11 *** Juser123 has quit IRC
18:24:17 *** rickoehn has quit IRC
18:35:15 *** To7 has joined #gnucash
18:48:46 *** frakturfreak1 has quit IRC
18:52:38 *** User has joined #gnucash
19:05:25 *** fbruetting has quit IRC
19:05:42 *** fbruetting has joined #gnucash
19:56:01 *** ArtGravity has quit IRC
20:19:37 *** mrklintscher1111 has quit IRC
20:32:55 *** mrklintscher1111 has joined #gnucash
20:34:42 *** bertbob has quit IRC
20:40:46 *** To7 has quit IRC
20:49:44 *** bertbob has joined #gnucash
20:50:54 *** User has quit IRC
21:10:25 *** bertbob has quit IRC
21:24:24 *** bertbob has joined #gnucash
22:44:42 *** fbruetting has quit IRC
22:45:24 *** fbruetting has joined #gnucash
22:51:52 *** fbruetting has quit IRC
23:50:30 *** jralls has quit IRC
23:56:22 *** pilotauto has joined #gnucash