2017-12-19 GnuCash IRC logs

00:06:31 *** CDB-Man_ has quit IRC
00:07:24 *** CDB-Away_ has quit IRC
01:11:29 *** Mechtilde has joined #gnucash
01:29:45 *** CDB-Man has joined #gnucash
01:36:28 *** CDB-Man_ has joined #gnucash
01:43:15 *** bertbob has quit IRC
01:50:20 *** Mechtilde has quit IRC
01:59:58 *** bertbob has joined #gnucash
02:05:05 *** bertbob has quit IRC
02:07:02 *** CDB-Man_ has quit IRC
02:07:11 *** CDB-Man has quit IRC
02:10:24 *** bertbob has joined #gnucash
02:13:42 *** bertbob has quit IRC
02:17:23 *** CDB-Man has joined #gnucash
02:17:41 *** CDB-Man_ has joined #gnucash
02:27:32 *** sbluhm has joined #gnucash
02:29:43 *** bertbob has joined #gnucash
02:30:57 *** sbluhm has quit IRC
02:57:24 *** fabior has joined #gnucash
03:05:49 *** CDB-Man_ has quit IRC
03:07:42 *** gjanssens has joined #gnucash
03:07:42 *** ChanServ sets mode: +o gjanssens
03:07:52 *** CDB-Man has quit IRC
03:08:07 <gjanssens> .
03:12:03 *** CDB-Man has joined #gnucash
03:12:16 *** CDB-Man_ has joined #gnucash
03:49:55 *** pilotauto has quit IRC
03:59:05 *** hoijui has joined #gnucash
04:19:43 *** Aussie_matt has joined #gnucash
04:38:03 *** fabior has quit IRC
04:51:58 *** fabior has joined #gnucash
05:09:48 *** storyjesse has joined #gnucash
05:17:10 *** fabior has quit IRC
05:33:00 *** User has joined #gnucash
05:45:31 *** storyjesse has quit IRC
06:08:50 *** fekepp has joined #gnucash
06:14:57 *** lachmanfrantisek has joined #gnucash
07:36:21 *** Aussie_matt has quit IRC
08:09:21 *** User has quit IRC
08:35:52 *** fekepp has quit IRC
08:36:20 *** jotrago1 has joined #gnucash
08:36:40 *** jotrago has quit IRC
08:36:41 *** jotrago1 is now known as jotrago
08:59:32 <chris> lmat - so here's my game plan - I think this is legit - your branch will periodically rebase to upstream/unstable and will contain only C commits, and my branch will rebase to upstream/unstable and contain .scm commits only - will this work? I'm needing to rewrite history
09:01:13 *** bastianilso_ has joined #gnucash
09:10:32 <chris> bingo - got a segfault/crash on 2.7.2 as released on win10 - Report / Business / Customer Report
09:10:49 <chris> jralls_afk bingo - got a segfault/crash on 2.7.2 as released on win10 - Report / Business / Customer Report
09:10:55 <chris> ditto linux
09:13:01 <lmat> chris: hmm, if you'r branch with scheme changes is on upstream/unstable, it won't be able to use any C changes I make, right?
09:13:40 <chris> oops typo absolutely I will rebase to your lmat/time64 from time to time
09:13:54 <chris> brain/hands interface broken
09:13:55 <lmat> chris: Okay, that makes good sense.
09:14:37 *** bastianilso_ has quit IRC
09:15:19 <lmat> chris: ^_^ Right about the reformatting change and difficulties merging. I've gotten more and more gunshy about making editorial changes for that reason. If I change a line, I better have meant it because I may have to pay considerably for it in merge conflicts!
09:15:37 <lmat> chris: And I may have to pay for it over and over as I rebase to stay current.
09:15:53 <lmat> chris: I'll gather my commits onto unstable (should we be targeting master?).
09:16:34 <chris> I suspect upstream/unstable as a moving target is the one for now?
09:18:08 *** jotrago has quit IRC
09:18:12 *** jotrago has joined #gnucash
09:20:01 <lmat> chris: Is there a goal to get this change into 2.7, or are we aiming for 2.8? I don't think it matters because the changes should be rather tangential and easy enough to rebase if needed.
09:21:32 <lmat> That was easy. I've rebased onto unstable and will run tests and push soon.
09:21:36 <lmat> not rebased. cherry-picked.
09:25:18 <lmat> warlord: Why was much of the core business logic rewritten from scheme to C?
09:30:51 <chris> lmat I have no idea, converting tp to t64 was your idea :)
09:30:57 <chris> to me it's good practice
09:31:57 <chris> my guess is scheme is great for mathematical logic and integrity but C has incredible compatibility with GUI/other logic
09:32:20 <lmat> chris: haha, okay! I found that commit you mentioned (e39464)! Pushed.
09:34:18 <warlord> lmat: what do you mean? What business logic was in scheme?
09:34:40 <lmat> warlord: My impression is that gnucash used to be a scheme project (completely), and that at some point large tracts of it were rewritten in C.
09:35:58 <warlord> No, it was never a completely scheme project.
09:36:11 <warlord> It was always a mixture; at no point was it ever more than 30% scheme
09:36:30 <lmat> warlord: Ahh, thanks for that correction, then.
09:44:53 *** hoijui has quit IRC
09:47:40 *** mebX has quit IRC
09:47:45 <lmat> chris: Hey, the CI passed! https://travis-ci.org/limitedAtonement/gnucash/builds/318648675 \o/
09:51:03 <chris> \m/ rock on
09:52:49 <lmat> chris: Now you get the heavy lifting job: rewriting large tracts of history :-S
09:55:09 <chris> for now it's easy - (1) add time64 API in C->guile - (2) convert scheme to use time64 only - (3) then the difficult part to decode guile->C hooks
09:56:37 <chris> I think that I can remove all tp in scheme within 2 weeks
09:56:49 <chris> (optimistic estimate)
09:57:01 *** fabior has joined #gnucash
09:58:12 <gjanssens> jralls_afk: in followup of our recent discussion on whether to drop guile 1.8 support: I have just discovered we're not testing with guile 1.8 *at all*
09:58:21 <gjanssens> jralls_afk: all of our travis runs use guile 2.0
09:59:36 <gjanssens> jralls_afk: as an experiment I tried to build maint with guile 1.8 (using cmake/ninja-build) on my F27 system. It builds but has 33 test failures
10:00:09 <gjanssens> I believe they are relatively easy to fix: with guile 1.8 gnucash can't find the unittest-support.scm module.
10:10:23 <lmat> gjanssens: In our docker script, there is run apt-get build-dep -qq gnucash; which should bring in guile. That's bringing guile 2?
10:11:13 *** fabior has quit IRC
10:12:19 *** fabior has joined #gnucash
10:12:26 <gjanssens> lmat: I presume so yes
10:16:18 <gjanssens> lmat: there are no docker scripts yet on the maint branch btw, they were implemented on master/unstable
10:16:38 <gjanssens> But apt-get build-dep -qq gnucash does pull in guile 2 indeed.
10:16:51 <gjanssens> Even on Ubuntu 14.04
10:17:22 <gjanssens> So I think it's really reasonable to retire our guile 1.8 support on unstable
10:17:41 <gjanssens> We should mention it in the release notes for the next 2.7.x release though
10:17:41 * chris ears hears music
10:28:21 *** lachmanfrantisek has quit IRC
10:39:56 *** meb has joined #gnucash
10:45:31 *** meb has quit IRC
10:55:54 *** meb has joined #gnucash
11:13:54 *** lachmanfrantisek has joined #gnucash
11:15:55 *** ArtGravity has joined #gnucash
11:17:01 *** lachmanfrantisek has quit IRC
11:17:26 *** lachmanfrantisek has joined #gnucash
11:20:41 *** lachmanfrantisek has quit IRC
11:27:30 *** lachmanfrantisek has joined #gnucash
11:36:36 *** jralls_afk is now known as jralls
11:38:49 *** lachmanfrantisek has quit IRC
11:40:57 <jralls> gjanssens: Good. There's numeric API in 2.0 that I'd like to use that's not available in 1.8.
11:41:56 <gjanssens> Ok, I'll work on getting it removed in the next couple of days.
11:42:03 <jralls> gjanssens: I wouldn't spend any time fixing testing bugs in maint for guile 1.8. It's clearly Windows only and we've never gotten testing to work in the old windows build.
11:42:30 <gjanssens> Fyi, barring one failing test case I have guile 2.2 working on unstable
11:43:03 <jralls> I do think it would be worthwhile to get unstable's tests to pass on Mingw64, but it's not currently on my priority list.
11:43:05 <gjanssens> I had no intention to work on guile 1.8 fixes on 2.6.
11:43:40 <jralls> OK. I was responding to your "they look like they'd be easy to fix".
11:45:43 <jralls> chris, lmat: I'd very much like to get the time64 and transaction.scm changes into unstable, ideally by 2.7.3 at the end of the month so that we can declare code and string freeze, send a tarball to the Translation Project, and start pushing the more adventurous users for bug reports.
11:46:49 *** lachmanfrantisek has joined #gnucash
11:47:21 <jralls> chris: What caused the segfault?
11:47:31 *** User has joined #gnucash
11:50:14 *** oozer has joined #gnucash
11:50:41 *** User has quit IRC
11:57:18 *** chf has quit IRC
12:01:56 *** KaiForce has joined #gnucash
12:04:06 *** chf has joined #gnucash
12:12:39 <lmat> Darn, can't open gnucash after the last archlinux upgrade yesterday :'( libenchant is not available or something.
12:15:30 <lmat> Now I get to build goffice and webkit. hopefully that will get me going :-)
12:33:51 *** Mechtilde has joined #gnucash
12:54:10 <jralls> gjanssens: Are you happy with PR233, Fix Bayes?
12:56:49 <gjanssens> lmat: if you're building unstable you don't need goffice any more.
12:57:08 <gjanssens> jralls: I haven't had time to check yet
12:57:19 <gjanssens> I'll do so one of the following days
13:04:02 <Mechtilde> is Dmitry Smirnov also here
13:22:48 *** KaiForce has quit IRC
13:34:50 <jralls> Mechtilde: I don't think so.
13:36:58 *** fabior has quit IRC
13:37:42 *** lachmanfrantisek has quit IRC
13:38:18 *** fabior has joined #gnucash
13:51:53 *** MadDogMcCree has quit IRC
14:06:04 *** Mechtilde has quit IRC
14:14:00 *** fabior has quit IRC
14:27:07 *** frakturfreak has joined #gnucash
14:28:44 *** fekepp has joined #gnucash
14:36:47 *** Jimraehl1 has joined #gnucash
14:38:06 *** Jimraehl1 has quit IRC
14:44:08 *** lachmanfrantisek has joined #gnucash
15:40:15 <lmat> gjanssens: Not unstable, I'm wanting to input some bills, etc. :-) webkit is still building.
15:45:25 *** puck has quit IRC
15:51:22 *** puck has joined #gnucash
15:53:59 <gjanssens> lmat: oh, that's unfortunate...
16:10:29 *** Gulba has joined #gnucash
16:11:41 *** Gulba has quit IRC
16:22:28 <lmat> jralls: Unstable it is for time64 work.
16:28:29 *** pilotauto has joined #gnucash
16:32:36 <chris> jralls - I'm confident I can rewrite all scm to access read-only scheme time64 API calls in a couple weeks. In my mind, this change is harmless and will only get rid of timepairs in scheme's code. There'll be one or two places (eg options.scm) whereby a compatibility shim can reside.
16:32:58 <chris> jralls - I haven't written a unit test before - I can try!
16:33:11 <chris> jralls and I don't know how to diagnose a coredump - can try
16:35:52 <chris> to crash, just run Report/Business/any of Customer-Employee-Vendor Report which runs owner-report.scm
16:37:16 <lmat> jralls: Are timespecs written to the database? It looks like they're written as time64s?
16:38:25 <jralls> lmat: I think they're saved as time64s in Postgresql and as ISO date-time strings everywhere else.
16:38:51 <jralls> chris: OK, I got the crash.
16:39:13 *** frakturfreak has quit IRC
16:40:33 <jralls> lmat: It's yours: Derefing a nullptr at KvpFrameImpl::get_slot(char const*) const + 173
16:41:18 <jralls> lmat: called from qof_book_get_option + 399
16:43:44 <lmat> hm
16:44:12 <lmat> This is on chris's time64-ftw branch?
16:45:33 <jralls> No, 2.7.2 release.
16:46:11 <lmat> Ahhh
16:48:32 <lmat> I'll check out the 2.7.2 tag and diagnose from there.
16:59:08 *** warlord is now known as warlord-afk
17:11:35 <chris> quick qu - if build fails when WITH_SQL=OFF (no DBI) because the #ifdef DBI #else code is incorrect, is it worth reporting?
17:30:10 <chris> done on github
17:38:55 *** gjanssens has quit IRC
17:43:24 <jralls> chris: Did you try your suggestion?
17:44:19 <chris> yes seems to work
17:47:44 <jralls> OK. I'll apply it in a few minutes since gjanssens has gone to bed.
18:43:33 *** fekepp has quit IRC
18:53:17 <jralls> chris: Fixed.
18:53:46 <chris> jralls: yay
18:54:16 <chris> jralls: I think I made a good unit test but I dont understand CMakeLists.txt - I'll push to PR; can you enable unittest?
18:55:07 <jralls> lmat: Looks like it's not in KVP after all. Running in the debugger shows that a swig wrapper fn is calling qof_book_get_option with book=nullptr.
18:55:53 <jralls> chris: OK.
18:59:39 *** lachmanfrantisek has quit IRC
18:59:45 *** ArtGravity has quit IRC
20:20:14 *** wget has quit IRC
20:21:52 *** wget has joined #gnucash
20:42:50 *** chf has quit IRC
20:44:46 *** chf has joined #gnucash
21:11:10 *** oozer has quit IRC
21:19:39 *** jralls has quit IRC
21:31:50 *** User has joined #gnucash
21:34:43 *** O01eg has quit IRC
21:39:12 *** chf has quit IRC
21:41:06 *** chf has joined #gnucash
21:48:52 *** O01eg has joined #gnucash
21:57:53 *** meb has quit IRC
21:59:26 *** meb has joined #gnucash
22:02:51 *** meb has quit IRC
22:10:07 *** meb has joined #gnucash
22:17:12 *** O01eg has quit IRC
22:25:08 *** sborman has joined #gnucash
22:46:45 *** O01eg has joined #gnucash
23:05:37 *** chf has quit IRC
23:06:00 *** chf has joined #gnucash
23:20:28 *** bastianilso_ has joined #gnucash
23:23:46 *** bastianilso_ has quit IRC
23:31:15 *** CDB-Man has quit IRC