2016-05-09 GnuCash IRC logs
00:45:25 *** fell_ has quit IRC
01:20:57 *** Mechtilde has joined #gnucash
01:25:10 *** Mechtilde has quit IRC
03:45:26 *** fabior has joined #gnucash
03:48:33 *** fabior has quit IRC
04:02:59 *** fabior has joined #gnucash
04:33:03 *** nomeata has joined #gnucash
04:55:28 *** aqua___ has joined #gnucash
04:58:43 *** aqua___ has quit IRC
05:12:27 *** aqua___ has joined #gnucash
05:34:29 *** raeburn has quit IRC
05:37:48 *** fell has joined #gnucash
05:56:11 *** fell_ has joined #gnucash
05:57:56 *** fell has quit IRC
06:33:31 *** olpc-ch has joined #gnucash
06:33:51 <olpc-ch> hello
06:34:10 <olpc-ch> Is anyone there ?
06:46:04 *** finster has joined #gnucash
06:47:40 <olpc-ch> hello
07:00:29 *** Jimraehl1 has joined #gnucash
07:26:51 <g5pw> olpc-ch hi! ask your question and wait 😊
07:29:13 *** finster has quit IRC
07:30:24 *** olpc-ch has quit IRC
07:33:29 *** fabior has quit IRC
07:36:23 *** mlncn has joined #gnucash
08:17:02 *** warlord has joined #gnucash
08:17:03 *** gncbot sets mode: +o warlord
08:19:19 *** mlncn has quit IRC
08:37:16 *** raeburn has joined #gnucash
09:22:14 *** mlncn has joined #gnucash
09:38:54 *** fabior has joined #gnucash
10:12:56 *** fabior has quit IRC
10:26:46 *** fabior has joined #gnucash
10:55:31 *** Muhannad_ has joined #gnucash
11:25:32 *** jimvideo has joined #gnucash
11:28:55 *** jimvideo has quit IRC
11:30:25 *** jimvideo has joined #gnucash
11:45:06 *** nomeata has quit IRC
12:34:09 *** Mechtilde has joined #gnucash
12:56:49 *** minot has joined #gnucash
13:14:14 *** Mechtilde has quit IRC
13:19:47 <jralls> lmat: re 748669, if you want smaller commits (and I do too) PRs are better than bugs. If you periodically push to feature branches on your github repo we can review and discuss along the way too.
13:23:55 <jralls> lmat: You can also do `git format-patch foo` where foo isn't HEAD^ to get a series of patches, each of which does something more bitesize like getting rid of qofsession-p.h and the old static analyzer markup. But a PR is easier to review and comment on.
13:49:45 *** Muhannad_ has quit IRC
13:52:21 *** Muhannad__ has joined #gnucash
14:13:14 *** rpg has joined #gnucash
14:13:55 *** rpg has quit IRC
14:36:56 *** raeburn has quit IRC
14:42:44 *** fabior_ has joined #gnucash
14:44:04 *** fabior has quit IRC
14:44:53 <lmat> jralls: I think PR is the right way to go; thanks.
14:45:53 *** fabior_ has quit IRC
14:45:57 *** warlord has quit IRC
14:46:07 *** gnomey has quit IRC
14:46:53 <jralls> OK. Please close the bug as obsolete. If you want I can look at your feature branch before you make a PR.
14:48:19 *** gnomey has joined #gnucash
14:48:21 *** warlord has joined #gnucash
14:48:22 *** gncbot sets mode: +o warlord
15:36:03 *** minot has quit IRC
15:38:05 *** raeburn has joined #gnucash
16:05:20 *** fabior has joined #gnucash
16:08:28 *** fabior has quit IRC
16:21:40 *** fabior has joined #gnucash
17:04:37 *** fabior has quit IRC
17:28:56 <lmat> jralls: I take it you're interested in me going ahead and being more aggressive with changesets?
17:29:47 <lmat> jralls: For instance, using RAII to own the backend and book will completely change the way src/engine uses session; I should go ahead and make that kind of change, too?
17:30:24 *** raeburn has quit IRC
17:30:33 <lmat> jralls: (I'm thinking of gnc_get_current_session which can return a session that has no backend, etc.)
17:31:19 <jralls> lmat: I think that particular change needs my c++-backend finished. That's a big part of why I went there.
17:32:21 <lmat> jralls: I can't quite tell to what you're referring by "went there". Why you created the branch, why you suggested RAII, or why you are suggesting using PRs (or something else)?
17:34:15 <warlord> lmat: I suspect by "went there" he meant why he startedworking on the c++-backend
17:34:19 <jralls> Why I left qof and started working on making the backend C++. A lot of classes in qof depend on it, so I figured getting it to C++ would make it easier to finish qof. Also there are some classes in qof that I want to get rid of.
17:34:37 *** Muhannad__ has quit IRC
17:35:29 <lmat> gotach
17:35:31 <warlord> I think it's fine to have a session without a backend (right now)... It's like having an un-saved Excel spreadsheet -- you can work on it for as long as you want.
17:35:57 <lmat> jralls: What classes are on your hitlist?
17:36:00 <lmat> warlord: Makes sense.
17:36:09 <warlord> However, I'd love to see the "session" globals go away and have everything require a context. I.e., I'd love to see gnc_get_current_session() to disappear.
17:36:54 <lmat> warlord: Ah yes; I think I read recently about an idea to be able to have multiple sessions simultaneously. It wasn't fleshed out, but sounded interesting.
17:38:06 <jralls> QfoClass, QofObject, QofCollection are the ones that come to mind first. QofQuery needs to be gutted and turned into a wrapper for SQL queries.
17:39:11 <warlord> jralls: that only works once we transition over to using embedded-SQL(ite) for our in-RAM storage.
17:39:42 <jralls> Warlord: Yes, I know. Hence the backend work.
17:39:48 <warlord> k
17:40:50 <lmat> jralls: What do you think about existing ORMs?
17:41:09 <lmat> jralls: I've only used Wt in C++, and I'm not sure if I like it or not (probably because I've only used that one).
17:42:43 <jralls> I looked at them last year. I don't know that it makes a lot of sense with our architecture, we have a pretty good mapping between objects and tables already.
17:43:12 <lmat> jralls: Sounds right to me.
17:44:15 <jralls> I'm more interested in the db abstraction to make it easy to use different SQL implementations.
17:44:40 <jralls> But I'm not terribly happy with libdbi.
17:46:31 <lmat> okay
17:48:59 <jralls> Back to gnc_session. It's a clumsy singleton with thread safety issues. I hope that the only place where it wouldn't have a book is in gnucash-bin.c.
17:49:40 <warlord> Yes, gncSession should always have a book, but not necessarily a backend.
17:53:19 <warlord> Anyways, gotta run to dinner. be back later.
17:53:34 *** warlord has quit IRC
17:56:40 <jralls> Hmm. gnucash-bin only calls gnc_get_current_session in inner_main_get_quotes...
17:56:50 * jralls runs gc in the debugger
18:14:00 <lmat> Supper time. Talk to you later.
18:14:13 *** mlncn has quit IRC
18:14:46 *** puck has quit IRC
18:20:37 <jralls> lmat: Yes, radically changing the engine so that book is owned by session is the right thing to do, but the only way to make that safe is for it to be a shared_ptr, so unfortunately it's going to be the one of the last changes since everything that needs a book for an argument will have to be C++. The backend is used a lot less, so it can convert sooner, but it still needs to be a C++ object.
18:24:11 *** puck has joined #gnucash
19:25:51 *** warlord has joined #gnucash
19:25:52 *** gncbot sets mode: +o warlord
19:39:17 *** warlord has quit IRC
19:45:55 *** tanderson has joined #gnucash
19:47:03 <tanderson> Q: Why does gnc-fq-update not pull in DateTime before trying to install Finance::Quote? I had to install it manually through CPAN to allow gnc-fq-update to succeed
20:12:41 *** raeburn has joined #gnucash
20:16:11 *** bertbob has quit IRC
20:23:48 *** mlncn has joined #gnucash
20:30:50 *** bertbob has joined #gnucash
20:52:44 *** raeburn has quit IRC
20:58:56 *** gnomey has quit IRC
21:01:05 *** tanderson has quit IRC
21:03:22 *** gnomey has joined #gnucash
21:09:50 *** tanderson has joined #gnucash
21:27:24 *** warlord has joined #gnucash
21:27:25 *** gncbot sets mode: +o warlord
21:57:42 *** raeburn has joined #gnucash
22:24:27 *** warlord has quit IRC
23:19:01 *** mlncn has quit IRC
23:22:06 *** jimvideo has quit IRC
23:24:05 *** jimvideo has joined #gnucash
23:59:45 *** tanderson has quit IRC