2017-12-23 GnuCash IRC logs

00:09:36 <chris> lmat - may I ask you to convert gnc_option_get_ui_value_date in gnome-utils/dialog-options.c to time64? I think this is the key to removing the option.scm compatibility shim
00:18:05 <chris> lmat - after converting dialog-options.c I'm *guessing * scheme can be completely timepair-free -- unless there's another hook somewhere else
00:21:10 <chris> lmat - another t64 conversion for gnc-dmy2timespec-start -middle and -end
00:29:31 *** Aussie_matt has joined #gnucash
01:14:34 *** jotrago has joined #gnucash
01:48:52 <chris> jralls - just thinking - if transaction.scm and time64 will need to be merged soon, I should create a PR for income-gst-statement.scm, then we can merge both reports, then we rebase time64-ftw onto it, I upgrade both reports to time64, then push time64-ftw.
01:56:41 *** Mechtilde has joined #gnucash
02:10:47 *** jotrago has quit IRC
03:04:49 <chris> jralls if you wouldn't mind - what's the safest way to disable scheme/python access to functions without disabling calls from C? e.g. wish to disable scheme access to gnc-print-date so that I can convert to qof-print-date
03:22:02 *** hoijui has joined #gnucash
03:52:47 *** fabior has joined #gnucash
04:00:34 *** fabior has quit IRC
04:08:56 *** gjanssens has joined #gnucash
04:08:56 *** ChanServ sets mode: +o gjanssens
04:10:45 *** chris has quit IRC
04:11:34 *** chris has joined #gnucash
04:13:08 <gjanssens> .
04:15:43 *** fabior has joined #gnucash
04:29:53 *** fabior has quit IRC
04:45:33 *** fekepp has joined #gnucash
05:56:31 *** CDB-Away_ has joined #gnucash
05:58:25 *** CDB-Man has quit IRC
05:58:51 *** CDB-Man_ has quit IRC
05:58:51 *** CDB-Man has joined #gnucash
06:11:30 *** Mechtilde has quit IRC
06:24:09 *** Mechtilde has joined #gnucash
06:44:09 *** Mechtilde has quit IRC
07:07:54 *** oozer has joined #gnucash
07:28:16 *** oozer has quit IRC
08:07:28 *** henrik has joined #gnucash
08:11:19 <henrik> Hi, I want to commit a small bugfix. I wonder for which branch I should create a pull request. Is there documentation anywhere how branches are intended to be used?
08:15:21 *** User has joined #gnucash
09:01:36 *** Aussie_matt has quit IRC
09:22:54 *** oozer has joined #gnucash
09:26:12 *** oozer has quit IRC
09:31:13 *** Jimraehl1 has joined #gnucash
09:41:05 *** Mechtilde has joined #gnucash
09:52:16 *** Mechtilde has quit IRC
10:06:23 *** yuriks has quit IRC
10:07:43 *** Mechtilde has joined #gnucash
10:09:25 *** yuriks has joined #gnucash
10:18:07 *** fekepp has quit IRC
10:18:15 *** kael has joined #gnucash
10:24:09 *** fekepp has joined #gnucash
10:57:40 *** henrik has quit IRC
11:05:06 *** oozer has joined #gnucash
11:09:01 *** Mechtilde has quit IRC
11:14:35 *** henrik has joined #gnucash
11:35:59 *** henrik has left #gnucash
11:56:09 *** oozer has quit IRC
12:28:55 *** noah has joined #gnucash
12:43:36 *** sborman has joined #gnucash
13:00:27 *** oozer has joined #gnucash
13:02:24 *** kael has quit IRC
13:04:39 <lmat> @tell henrik I'm excited to see your bugfix. Likely, you'll want to do your bug fix on the "maint" branch in github. Here's a document pertaining to the subject: https://wiki.gnucash.org/wiki/Release_Schedule
13:04:39 <gncbot> lmat: The operation succeeded.
13:06:21 *** oozer has quit IRC
13:19:25 <lmat> chris: I got the three changes done: transaction getvoidtime, gnc_option_get_ui_value_date, and added gnc_dmy2time64{,middle,neutral} (on the latter, I matched what was extant for timespec). Let me know if any of this can be improved.
13:22:31 <gjanssens> lmat: re henrik's bugfix: chris had already mentioned this fix and jralls committed it a few days back. And the branch in this case should have been unstable as it was a bug introduced on that branch only.
13:35:50 *** Mechtilde has joined #gnucash
13:50:15 *** fiddlerwoaroof has joined #gnucash
13:57:24 *** fabior has joined #gnucash
14:06:15 *** frakturfreak has joined #gnucash
14:16:37 *** hoijui has quit IRC
14:55:35 <jralls> chris: You have to disable the function in swig. Find "%include <gnc-date.h> and add "%ignore gnc_print_date;" immediately after it.
14:58:21 <jralls> chris: I think it would make more sense and be less work to merge time64-ftw ASAP and rebase transaction.scm and income-gst-statement.scm on the result and fix the API changes for both timepair->time64 and gnc:numeric->number.
15:00:31 *** fabior has quit IRC
15:01:54 *** CDB-Man is now known as CDB-Man_
15:11:05 <jralls> lmat: On the topic of time64, do you have a plan for replacing timespec in KVP?
15:41:45 *** fekepp has quit IRC
15:46:39 *** oozer has joined #gnucash
15:50:03 *** oozer has quit IRC
15:53:11 *** fell has joined #gnucash
15:56:31 <fell> Good evening! jralls, gjannsens, can you op gncbot?
15:56:58 <fell> gjanssens
16:05:38 *** jralls sets mode: +o gncbot
16:08:38 *** gncbot sets mode: +o fell
16:14:31 *** fabior has joined #gnucash
16:16:19 *** sborman has quit IRC
16:16:25 *** sborman has joined #gnucash
16:16:47 *** shakes808 has joined #gnucash
16:26:09 *** shakes808 has quit IRC
16:33:32 *** noah has quit IRC
16:34:18 *** Mechtilde has quit IRC
16:35:15 *** CDB-Man_ has quit IRC
16:35:33 *** CDB-Man has joined #gnucash
16:39:50 <lmat> jralls: I don't think I'll get it done very soon. I'm currently working on getting the engine to compile as C++.
16:40:06 <lmat> jralls: I mean, I think a 2.7 release is coming up soon, but I won't get it done by then.
16:42:54 <jralls> lmat: Yes, I want to release 2.7 by the end of the week. Converting the rest of engine to C++ isn't very high on my priority list. Why is it so high on yours?
16:52:24 *** Cuare has joined #gnucash
16:57:24 *** Cuare has quit IRC
16:57:49 *** Cuare has joined #gnucash
16:57:54 <jralls> gjanssens: While working on https://bugzilla.gnome.org/show_bug.cgi?id=791848 I found some instances where the return value of gnc_gmtime and gnc_localtime aren't checked before trying to deref the struct tm*, and they both return NULL if the supplied time64 is outside the valid range.
17:02:51 <gjanssens> jralls: oops, those need fixing...
17:03:27 <jralls> gjanssens: Of course. My question is about the swig wrappers that also don't check.
17:04:23 <gjanssens> Ooh. I believe those wrappers are in our own code so we should be able to fix those, right ?
17:04:38 <jralls> I can't find them.
17:06:04 *** sborman has quit IRC
17:06:28 <gjanssens> Hmm, not obvious indeed :(
17:06:29 <jralls> They make an scm_vector and populate it. My question is should I return the vector of SCM_UNDEFINED or should I fill it with 0?
17:07:26 <gjanssens> I can't answer that just like that. I'd have to study the code some more, but for that it's too late here.
17:07:33 <gjanssens> Can I get back to this tomorrow ?
17:07:51 <jralls> Sure.
17:08:01 <gjanssens> (probably by mail though as tomorrow is Xmass eve)
17:08:27 <gjanssens> I'll be with family starting 17h CET
17:08:37 <jralls> Yeah, I'll be cooking most of the day so it could wait till Monday or Tuesday if you like.
17:08:44 <gjanssens> Ok.
17:09:14 <gjanssens> Can you just give me a hint already where the scm_vector is created ? That will simplify my search
17:10:36 * gjanssens is signing off...
17:10:41 <gjanssens> Night all!
17:10:45 <jralls> base-typemaps.i
17:10:50 <gjanssens> Thansk
17:10:54 <gjanssens> Thanks*
17:10:57 <jralls> Good night, and merry Christmas!
17:11:04 <gjanssens> You too !
17:11:09 *** gjanssens has quit IRC
18:01:14 <chris> jralls- thank you - disabling in swig is mainly for testing that tp is truly gone. Is it desirable to do it? Python etc may suffer.
18:01:47 <chris> jralls time64-ftw good progress, may be complete, will need thorough testing which may not be easy during next few days!
18:09:40 <chris> jralls disabling timepairs in swig is meant to be for testing; I assume we don't want to disable API for now
18:12:19 <chris> so I guess order of operations is: (1) complete time64-ftw and merge, (2) merge transaction.scm and update income-gst-statement.scm (3) ensure gnc-numeric is also banished in a new PR
18:16:59 <chris> lmat thanks for all changes - they work well!
18:29:57 *** fabior has quit IRC
18:33:02 <jralls> chris: Timepairs are Scheme-only, it's called a Timespec in C. It appears from lmat's repo that there's still a bit of work to do on the C side.
18:35:41 <jralls> chris: But that's independent of Scheme. You can %ignore API in engine.i for testing without committing the change if that helps.
18:51:17 *** sborman has joined #gnucash
18:51:46 *** User has quit IRC
19:10:13 <lmat> jralls: I'm eager to rewrite business objects in C++ (Account, etc.), and it's *much* easier if the rest of engine is compiling in C++.
19:12:11 <lmat> How much longer are we supporting autotools?
19:12:12 <jralls> lmat: Sure, but that's for the next dev cycle. We don't want wholesale rewrites into C++ on unstable this late in the game. OTOH, we *do* want to get rid of Timespec and we do want to get to code freeze soon.
19:12:42 <jralls> lmat: Codesmythe is working on a PR to remove it. When he's ready, it's gone.
19:15:38 *** sborman has joined #gnucash
19:16:58 <lmat> jralls: Certainly; I'm not targeting unstable. When are you thinking about an unstable code freeze?
19:17:50 <jralls> I was hoping to do it with the release next weekend, but if timespec isn't dead it will have to slip.
19:18:16 <lmat> jralls: I don't see how I could get timespec out of the C code in a week, I would think it would be difficult to get it done in 4 weeks.
19:19:14 <jralls> Have you done only what chris needed for the scheme code?
19:19:20 <lmat> jralls: not much more than that.
19:19:35 <lmat> jralls: I'm happy to get on it and see how things go; I might surprise myself ;-)
19:23:29 <jralls> OK. Maybe we can divide things up. I'm going to be afk most of tomorrow and Monday. Go ahead and start and I'll pull your branch and attack somewhere you haven't yet.
19:23:51 <jralls> Tuesday.
19:25:27 <lmat> jralls: Sounds like a plan.
19:25:28 *** frakturfreak has quit IRC
20:54:01 *** User has joined #gnucash
20:55:56 *** sborman has quit IRC
20:59:19 *** chris has quit IRC
21:03:05 *** chris has joined #gnucash
21:04:08 <chris> jralls lmat from all my testing time64-ftw is now timepair-free and I'd hope the reports are intact. from my testing they seem unchanged. if users report buggy report I'll hunt down anything that I've overlooked.
21:29:47 <lmat> chris: Rock on!
21:31:00 *** User has quit IRC
21:31:34 *** kael has joined #gnucash
21:37:50 <chris> jralls I'm sure you know there are still remnants of timepair in .scm that can't be removed yet but they're not in reports. e.g. gnc_copy_trans will call "gnc:transaction->transaction-scm" which calls (gnc-transaction-get-date-posted) which returns a tp... I'm wary of modifying these because they don't have visual feedback!
21:39:04 <chris> hmm I guess if ts/tp will disappear soon I'll need to work on these too
21:47:52 <chris> so I guess more work on the table. when timepairs are deleted in C if there are swig-type errors I'll handle the scm part ;)
22:46:17 *** sborman has joined #gnucash