2017-11-20 GnuCash IRC logs

00:31:06 *** storyjesse has quit IRC
00:45:28 *** bebo has quit IRC
00:57:16 *** bertbob has joined #gnucash
01:00:30 *** bertbob has quit IRC
01:09:23 *** Mechtilde has joined #gnucash
01:15:17 *** bertbob has joined #gnucash
01:26:42 *** frakturfreak has quit IRC
01:40:41 *** fell has quit IRC
01:41:29 *** frakturfreak has joined #gnucash
01:41:48 *** fell has joined #gnucash
01:52:30 *** Mechtilde has quit IRC
02:20:28 <CDB-Away> Pace left too fast for me to reply
02:20:41 <CDB-Away> Was going to suggest a USD wrapper account
02:26:58 *** Simon has quit IRC
02:27:08 *** Simon has joined #gnucash
03:04:24 *** fekepp has joined #gnucash
03:09:24 *** Simon- has joined #gnucash
03:11:12 *** Simon has quit IRC
03:11:29 *** Simon- is now known as Simon
03:27:38 *** Gene has quit IRC
03:30:12 *** fabior has joined #gnucash
03:30:18 *** feuGene has joined #gnucash
04:23:19 *** storyjesse has joined #gnucash
04:30:32 *** fabior has quit IRC
04:45:15 *** fabior has joined #gnucash
04:48:54 *** pilotauto has quit IRC
04:53:11 *** fekepp has quit IRC
04:53:16 *** fekepp has joined #gnucash
05:26:58 *** fekepp has quit IRC
05:27:01 *** fekepp has joined #gnucash
05:54:41 *** fabior has quit IRC
05:55:21 *** fabior has joined #gnucash
06:03:47 *** fabior has quit IRC
06:04:53 *** fabior has joined #gnucash
06:26:52 *** Robert has joined #gnucash
06:32:42 *** Robert has left #gnucash
07:03:56 *** jotrago1 has joined #gnucash
07:04:23 *** jotrago has quit IRC
07:04:24 *** jotrago1 is now known as jotrago
07:08:30 *** jotrago has quit IRC
07:09:11 *** jotrago has joined #gnucash
07:11:10 *** stefano_ has quit IRC
07:28:30 *** fabior has quit IRC
08:24:09 *** User has joined #gnucash
08:26:32 *** storyjesse has quit IRC
08:39:02 *** User has quit IRC
09:21:39 *** Jimraehl1 has joined #gnucash
09:24:26 *** Jimraehl1 has left #gnucash
09:37:43 *** jotrago has quit IRC
09:41:53 *** Jacques has joined #gnucash
09:56:53 *** Cuare has joined #gnucash
09:57:34 *** kael has joined #gnucash
09:58:20 *** ArtGravity has joined #gnucash
10:08:21 <lmat> chris: I'm happy to help with building maint. Full disclosure: I am unable to build maint locally -- I use a docker container.
10:08:42 *** Jacques has quit IRC
10:11:01 <chris> lmat quick chat before heading off...
10:12:22 *** fekepp has quit IRC
10:13:28 <chris> I'm slightly skeptical of the reward/risk of rewriting all .scms in 1 go - I'm comfortable testing small timepair->time64 changes for individual reports, (and fixing bugs as they arise), I'll get lost in rewriting and upgrading API at the same time
10:16:11 <chris> so not sure how to proceed - although i've tested an upgraded transaction.scm which accesses read-only APIs, and is working, it has achieved this via a shim timepair->time64 converter in gnc:make-date-option.
10:16:47 <chris> In otherwords the scm changes haven't accessed any upgraded underlying C++ code
10:19:27 <chris> Case in point - the options GUI toolkit still expects timepairs therefore will barf upon reading the time64 stored in a date option, and default back to 01/01/70. and I don't understand the option C++ code.
10:27:17 <chris> perhaps I'll need to restart and change all gnc:transaction-get-date-posted (produces timepair) to xaccTransGetDate (produces time64) in unstable/*.scm, which should be safe and doesn't require SWIG changes, and later on attempt to upgrade the options GUI glue
10:28:20 <chris> I could use a CMake/Ninja copy/paste script tho
10:28:54 <chris> (software dev is hard!)
10:39:37 *** ArtGravity has quit IRC
10:46:03 *** ArtGravity has joined #gnucash
10:51:36 *** kael has quit IRC
11:06:59 <chris> lmat: xaccQueryAddDateMatchTT needs modifying - it expects parameters: use_start?, start-time, use_end? end-time, etc --- however it cannot handle where use-start? or use-end? is False.
11:09:12 <chris> lmat: OTOH I can circumvent this limitation by sending dummy values for stt and ett where use_start or use_end is False...
11:17:23 <warlord> chris: which language?
11:46:46 *** fekepp has joined #gnucash
11:54:25 *** Mechtilde has joined #gnucash
12:24:21 *** Mechtilde has quit IRC
12:24:34 *** Mechtilde has joined #gnucash
12:48:42 *** sixwheeledbeast has quit IRC
12:50:43 *** Mechtilde has quit IRC
12:53:45 *** sixwheeledbeast has joined #gnucash
12:59:26 *** feuGene has quit IRC
12:59:26 *** feuGene has joined #gnucash
13:01:20 *** jralls has joined #gnucash
13:01:20 *** ChanServ sets mode: +o jralls
13:03:54 <jralls> chris: +1 on small timepair->time64 changes for individual reports. Once commit per file is a good idea.
13:06:12 <jralls> chris: I'd like to see gnc-date-utils.scm go away. That might be too hard to get done for unstable. Same with rewriting the options mess into just C++ with a thin scheme wrapper. The bouncing back-and-forth between C and Scheme is a b***h to maintain.
13:06:52 <jralls> chris: But that's got to wait for after 2.8.
13:12:06 *** PaceyIV has joined #gnucash
13:12:54 <jralls> chris: Why do you say xaccAddDateMatchTT can't handle use_start or use_end being false? It just returns if they're both false but otherwise generates a query phrase.
13:36:12 *** Mechtilde has joined #gnucash
13:36:34 *** User has joined #gnucash
13:46:54 *** User has quit IRC
14:06:01 *** gjanssens_afk is now known as gjanssens
14:21:26 *** kael has joined #gnucash
14:26:12 *** kael has quit IRC
14:27:44 *** fabior has joined #gnucash
14:35:46 *** carwynnelson has joined #gnucash
14:44:12 *** fekepp has quit IRC
15:08:07 *** carwynnelson has quit IRC
15:18:22 <jralls> @tell codesmythe Could you have a look at https://bugzilla.gnome.org/show_bug.cgi?id=790620? Will using a target for gschemas.compile get make/ninja in the loop so that they can apply DESTDIR?
15:18:22 <gncbot> jralls: The operation succeeded.
16:02:01 *** fabior has quit IRC
16:13:12 <chris> jralls: hmm. so far one commit per file isn't guaranteed, because I'm still experimenting through scheme and bugfixing as I go. perhaps after it's all done I'll just re-do the whole branch, picking commits one by one while comparing with the latest unstable.
16:14:49 <chris> jralls: date-utilities.scm is being called by most reports willy-nilly; after converting all I'll see how difficult it seems and get back
16:17:45 <chris> jralls: see https://github.com/Gnucash/gnucash/blob/unstable/libgnucash/engine/Query.c#L382 - this routine will check gboolean for #true before accessing sts and ets. If gboolean is #false, sts/ets is ignored.
16:18:35 *** carwynnelson has joined #gnucash
16:19:01 <jralls> chris: Bug fixes should each get their own commit and should go to the maint branch. Refactor commits should also be small, no more than one source file and the tests associated with the refactor.
16:19:41 <jralls> chris: After all that it should be more straightforward to get the time64 changes. Just remember that committing is like voting, best done early and often!
16:20:38 <chris> jralls: ok will try remember bugfixes. but I'm having to learn some scheme intricacies and learning how to work in a git team too :)
16:21:01 <jralls> chris: xaccQueryAddDateMatchTS, yes, that's the point. if use_start is #f then sts is ignored. If both use_start and use_end are #f the function is a no-op.
16:21:24 *** carwynnelson has quit IRC
16:21:44 *** carwynnelson has joined #gnucash
16:22:13 <chris> see https://github.com/Gnucash/gnucash/blob/unstable/libgnucash/engine/Query.c#L478 - xaccQueryAddDateMatchTT will access sts/stt before calling xaccQueryAddDateMatchTS - some legacy scheme code will call xaccQueryAddDateMatchTT with sts/ets being '#false' and use_start/use_end being '#false' too -- it raises exception "expect integer, got boolean"
16:22:42 <chris> so, it's not really a 1-to-1 equivalent code - I can circumvent by sending integer-zero when use_start/use_end is #false
16:24:11 *** frakturfreak has quit IRC
16:24:46 *** carwynnelson has quit IRC
16:28:34 <chris> jralls: it's just one of the exceptions that arise that makes it slightly tricky to convert timepair->time64 - "is it something I did wrong?" but now I know how to circumvent
16:32:53 *** bunk3m has joined #gnucash
16:36:28 <jralls> It's difficult for me to believe that Scheme can detect a type mismatch inside a C function that it's called, especially when the parameter is passed by copy to the C function.
16:36:36 *** Mechtilde has quit IRC
16:38:28 *** bunk3m has left #gnucash
16:38:35 *** carwynnelson has joined #gnucash
16:41:34 *** pilotauto has joined #gnucash
16:42:32 *** carwynne_ has joined #gnucash
16:42:32 *** carwynnelson has quit IRC
16:43:24 <chris> well here's screenshot https://imgur.com/a/44v0L
16:45:41 <chris> not really a big issue in the bigger picture; I'll apply conversion along the way
16:50:06 <jralls> chris: It's not because of how the two functions are written. It's because the SWIG wrapper for TS calls gnc_timepair2timespec that silently returns {0, 0} if the type is wrong and the TT wrapper calls Guile's scm_to_int64 that raises an exception. The function internals are irrelevant.
16:54:26 <chris> ah that's why
16:56:10 <chris> my feeling is the API caller should ensure correct type then
16:59:05 <chris> warlord: I assume you were asking about xaccQueryAddDateMatchTT - the issue has been found
16:59:07 <jralls> Agree. One of the things I dislike about "scripting" languages is their uneven enforcement of type safety. That start-date-tp might be a boolean or an int depending on whim encourages poor programming.
17:02:27 <chris> Well with this small hurdle overcome I should be able to move on quickly.
17:02:36 <chris> Thanks for the guidance
17:02:46 <jralls> NP
17:05:07 *** fbruetting has joined #gnucash
17:06:55 *** carwynne_ has quit IRC
17:06:55 *** fbruetting has quit IRC
17:07:14 *** fbruetting has joined #gnucash
17:09:37 <chris> btw I'm up early because of pre-wedding blues, and will be a bit occupied in next couple weeks ;-)
17:10:42 <jralls> Yours?
17:13:36 *** carwynnelson has joined #gnucash
17:18:05 <chris> Yes ...
17:18:22 <chris> hacking code is therapeutic :)
17:18:24 *** carwynnelson has quit IRC
17:19:36 <jralls> Can be... ;-). When's the wedding?
17:20:36 <chris> dec 3rd
17:22:38 <jralls> So a bit occupied for a couple of weeks and completely occupied for a month or many after that... ;-)
17:26:41 <chris> should be ;)
17:30:31 *** Mechtilde has joined #gnucash
17:31:01 *** PaceyIV has quit IRC
17:32:19 *** DataWraith has joined #gnucash
17:32:24 *** carwynnelson has joined #gnucash
17:35:56 *** Mechtilde has quit IRC
17:37:30 *** fekepp has joined #gnucash
17:37:39 *** gjanssens has quit IRC
17:41:18 *** DataWraith has quit IRC
18:55:24 *** carwynnelson has quit IRC
19:07:46 *** carwynnelson has joined #gnucash
19:09:44 *** carwynnelson has quit IRC
19:10:05 *** carwynnelson has joined #gnucash
19:13:08 *** carwynnelson has quit IRC
19:48:12 *** Cuare has quit IRC
20:34:30 <lmat> chris: Congratulations!
20:54:37 *** fbruetting has quit IRC
21:05:30 *** storyjesse has joined #gnucash
22:14:47 *** kael has joined #gnucash
22:24:45 *** ArtGravity has quit IRC
23:07:05 *** kael has quit IRC
23:11:13 *** storyjesse has quit IRC
23:37:48 *** feuGene has left #gnucash