2014-09-28 GnuCash IRC logs

00:21:47 *** GabrieleV_ has joined #gnucash
00:21:47 *** GabrieleV has quit IRC
00:21:48 *** GabrieleV_ is now known as GabrieleV
03:55:43 *** GabrieleV_ has joined #gnucash
03:55:55 *** GabrieleV has quit IRC
03:55:55 *** GabrieleV_ is now known as GabrieleV
04:13:03 *** StuM has joined #gnucash
04:27:51 *** fell has joined #gnucash
04:27:51 *** gncbot sets mode: +o fell
04:42:36 *** AndreeeCZ has joined #gnucash
04:51:00 *** hhn10 has quit IRC
05:00:31 *** MechtiIde has joined #gnucash
05:01:35 *** fabior has joined #gnucash
05:07:49 *** fabior has quit IRC
05:21:27 *** AndreeeCZ has quit IRC
05:57:51 *** GabrieleV_ has joined #gnucash
05:58:21 *** GabrieleV has quit IRC
05:58:21 *** GabrieleV_ is now known as GabrieleV
06:33:40 *** MechtiIde has quit IRC
06:37:23 *** GabrieleV_ has joined #gnucash
06:37:27 *** GabrieleV has quit IRC
06:37:27 *** GabrieleV_ is now known as GabrieleV
07:10:12 *** wol has joined #gnucash
07:14:24 *** gjanssens has joined #gnucash
07:14:26 *** gncbot sets mode: +o gjanssens
07:21:19 *** wol has quit IRC
07:34:00 *** andy has quit IRC
07:36:32 *** andy has joined #gnucash
07:57:37 *** himaxx has joined #gnucash
08:02:24 *** GabrieleV_ has joined #gnucash
08:02:25 *** GabrieleV has quit IRC
08:02:25 *** GabrieleV_ is now known as GabrieleV
08:11:39 *** himaxx has quit IRC
08:11:44 *** Jimraehl1 has left #gnucash
08:15:36 *** StuM has quit IRC
08:22:31 *** Jimraehl1 has joined #gnucash
08:28:21 <gjanssens> warlord: thanks for the pointer towards the properties callback
08:28:35 *** O01eg has joined #gnucash
08:28:46 <gjanssens> I found some pointers to that in the code as well
08:29:14 <gjanssens> There is the option to add a callback to gui changes
08:29:56 <gjanssens> However it comes with the advice not to change the options database during that callack
08:30:04 <gjanssens> Because a user can still hit cancel
08:30:32 <gjanssens> I'll see if I can tweak that when I will be working on that part of the code again
08:30:57 <gjanssens> jralls: you're welcome
08:31:33 <gjanssens> jralls: there were lots of commits yesterday creating branches and tags
08:31:37 <gjanssens> What was that all about ?
08:45:21 <Simon> if you remove the call to gtk_window_set_skip_taskbar_hint() in src/gnome-utils/gnc-splash.c it fixes startup detection on Ubuntu Unity
08:46:01 <Simon> but that'll mean you get a GnuCash window in the taskbar under Gnome for the splash window that then changes to a different new window when it finishes starting
08:47:30 <Simon> (the launcher is just ignoring the splash window because it's not set to display in the taskbar, so clicking it again causes it to launch again until the main window is displayed)
09:38:48 *** MechtiIde has joined #gnucash
09:49:23 <warlord> gjanssens: you're welcome. (not sure why gncbot didn't give you the messages, unless you contacted it privately)
10:02:49 <gjanssens> warlord: I get my messages in private the moment I log in
10:03:18 <gjanssens> I have configured my irc client to ident with gncbot at that moment
10:03:33 <gjanssens> So it replies with any outstanding messages
10:50:20 *** wol has joined #gnucash
11:16:48 <warlord> ah, okay
11:20:29 *** GabrieleV_ has joined #gnucash
11:21:00 *** GabrieleV has quit IRC
11:21:00 *** GabrieleV_ is now known as GabrieleV
11:29:03 *** ErKa has joined #gnucash
12:18:53 *** jralls has quit IRC
12:19:41 *** jralls has joined #gnucash
12:19:42 *** gncbot sets mode: +o jralls
12:34:35 <jralls> gjanssens: I'm not sure how those tags leaked back. It was from my Linux VM and they were left over from the archiving of old history. I think I got them both cleaned up.
12:39:12 <gjanssens> jrall: ok
12:39:33 <gjanssens> git can do surprising things sometimes...
12:41:25 <gjanssens> jralls: fyi the windows build server is currently working on boost even though we don't use it for 2.6.x
12:41:44 <gjanssens> That's a flaw in our windows build scripts we should probably look into
12:42:15 <gjanssens> Or we can just live with it for the 2.6 series. It will be mandatory as of 2.8 anyway
12:43:49 <jralls> gjanssens: We'd better fix it so that maint and master have different dependency lists; otherwise it's going to bite us as the dependencies for master continue to diverge.
12:44:24 *** wol has quit IRC
12:44:26 <jralls> OTOH, Boost is near the end, so it's almost done, which is good.
12:44:33 <gjanssens> yep
13:10:18 <warlord> Why is maint pulling in master deps? Or is this becuase we pulled out the packaging/win32 into its own repo?
13:10:48 <warlord> (maybe we need to branch/tag the packaging repo similarly to the main repo?)
13:13:16 *** wol has joined #gnucash
13:20:51 <gjanssens> that's indeed the reason and the proper solution
13:25:21 *** MechtiIde has quit IRC
13:37:24 <warlord> which begs the question: why did we separate it in the first place?
13:38:26 <gjanssens> because I want it to evolve on its own
13:38:38 <gjanssens> it's rare we add dependencies like we did now with boost
13:39:15 <warlord> There are many times that we have differing dependencies between maint and master.
13:39:47 <gjanssens> Those are mostly artificial
13:40:22 <gjanssens> Look at linux distributions: the same gnucash can be built with several version of its dependencies not just one
13:41:17 <gjanssens> The cases where we really *need* different versions on maint and master are infrequent
13:41:56 <warlord> I guess I'm more thinking about being able to move back in time and "rebuild" a release in the same way it was originally built.
13:42:48 <gjanssens> Hmm, ponders...
13:43:45 <warlord> I think it's fine to keep upgrading the deps, but we should do it explicitly, not implicitly.
13:43:45 <gjanssens> Both Fedora 19 and 20 ship gnucash 2.6.3, however not with the same versions of the dependencies
13:43:47 <warlord> (IMHO)
13:44:39 <gjanssens> We could add release tags in the gnucash-on-windows repository on the commit used to build the versions
13:45:23 <gjanssens> One advantage of separating the repo is that these things now explicitly become clear
13:45:27 <jralls> We could, but it's not really necessary. You can easily find the right commit with git log.
13:46:03 <gjanssens> And IMO we should treat the gnucash-on-osx repo the same as the windows repo
13:46:42 <gjanssens> jralls is about the only one actively using it, but it has to deal with the same issues
13:47:06 <gjanssens> I know it uses jhbuild which sets up different targets for the various supported branches
13:47:18 <gjanssens> We could go for a similar scheme in gnucash-on-windows as well
13:47:26 <jralls> I don't tag, but since I build releases off of tarballs it's easy to find the right commit for each one.
13:47:26 <gjanssens> And then we can stick to one single branch
13:48:03 <jralls> I think that's easier than separate branches.
13:48:11 <gjanssens> jralls: how can we find the right commit with git log ?
13:48:17 <gjanssens> In gnucash-on-windows, I mean
13:48:32 <jralls> Just look at the dates of the commits.
13:48:44 <gjanssens> Oh right.
13:49:33 <gjanssens> Another small improvement I had once thought of but forgot so far could be to add the gnucash-on-windows revision somewhere in the about box on Windows
13:49:40 <jralls> Though frankly I think it would be better to build releases from tarballs instead of git, if only to get rid of the git commit number in the About dialog.
13:49:45 <gjanssens> Just like the gnucash revision is there.
13:50:09 <gjanssens> Heh, obviously I was thinking the other way around...
13:50:28 <warlord> jralls: how would we do that programatically? And how do we handle different deps in master vs maint if there's only a single gnucash-on-xxx branch?
13:50:55 *** wol has quit IRC
13:51:28 <warlord> jralls: yes, that would be nice... except right now IIRC we can't run "make dist" on windows. But arguably, yes, we should "make dist" and then build the binaries from that.
13:51:57 <gjanssens> what's the problem with having the git revision number in the About dialog ?
13:52:38 <jralls> warlord: With jhbuild it's easy, just have different modules for the different versions. With the windows script we'd use different urls depending on what we're building. They'd go in defaults.sh.
13:53:11 <jralls> As for make dist, no, just get the tarballs from SF. I upload them at the same time I push the tag to code.
13:53:30 *** ErKa has quit IRC
13:53:55 <jralls> gjanssens: It makes it look like it's not an official release.
13:55:41 <gjanssens> ok. that's fair
14:24:17 <jralls> So our Debian packager is whining because uploading the tarballs to SourceForge before they're ready for public release "confuses" him. Think we could set up a place on code that we can aim the build-scripts at so he can't snoop?
14:27:43 <warlord> jralls: Sure, but would you want them to be autogenerated on code?
14:27:52 <warlord> Or are you asking for an ftp site?
14:28:17 <warlord> And why would that confuse him? How would he even KNOW that they're up there if you don't send out the announcement
14:30:27 <jralls> warlord: He looked at the release schedule in the Wiki and then went and looked when he got up this morning. It's on the Dev list as "Gnucash 2.6.4 in Debian". He's already built the Debian package and pushed it to their unstable repo.
14:32:48 <jralls> Anyway, yeah, it would need to be ftp or http so that the scripts work for non-devs (i.e., people who don't have keys on code) can get to it.
14:48:50 <jralls> Hmm. I just experimented with the "stage this folder for 3 days" feature. It turns out I can turn it on and off at will, so as long as curl can still find it, that will suffice.
14:52:26 <jralls> And it does, so that's good. I'll just add a note in the wiki to remind me to use it.
14:52:38 <warlord> Awesome.
14:52:47 <warlord> And you should tell him to wait until you announce..
14:52:58 <warlord> Unless you want me to blast him.
14:53:28 <warlord> I understand he's trying to be proactive, but schedules often slip.
14:53:39 <warlord> ... or testing shows failures that require rebuilding the release.
14:54:01 <jralls> Yeah. Especially the latter.
15:07:58 <warlord> Right.
15:08:20 <warlord> I can't get to the email thread right now -- in a nutshell what was his complaint?
15:08:32 <warlord> Ahh, I can get to your most recent reply.. I'll read that.
15:41:34 <warlord> Gotta run. Later all.
15:42:21 *** warlord has quit IRC
16:07:36 *** GabrieleV_ has joined #gnucash
16:08:03 *** GabrieleV has quit IRC
16:08:04 *** GabrieleV_ is now known as GabrieleV
16:27:06 *** GabrieleV_ has joined #gnucash
16:27:09 *** GabrieleV has quit IRC
16:27:10 *** GabrieleV_ is now known as GabrieleV
16:50:42 *** fell has quit IRC
16:53:02 *** gjanssens has quit IRC
17:05:28 *** himaxx has joined #gnucash
17:08:57 *** himaxx has quit IRC
17:25:09 *** mlncn has joined #gnucash
17:25:28 *** andy has quit IRC
17:46:38 *** GabrieleV_ has joined #gnucash
17:47:09 *** GabrieleV has quit IRC
17:47:10 *** GabrieleV_ is now known as GabrieleV
17:57:50 *** mlncn has quit IRC
18:20:53 *** ErKa has joined #gnucash
18:27:10 *** StuM has joined #gnucash
18:36:11 *** ErKa has quit IRC
18:43:40 *** lmat has joined #gnucash
18:45:10 <jralls> Hey, lmat, what was the problem you were having with compiling your KvpValue patch?
18:55:26 *** mlncn has joined #gnucash
19:03:59 *** lmat has quit IRC
19:21:52 *** mlncn has quit IRC
19:40:43 *** jeremy has quit IRC
19:42:09 *** GabrieleV_ has joined #gnucash
19:42:27 *** GabrieleV has quit IRC
19:42:27 *** GabrieleV_ is now known as GabrieleV
19:46:15 *** Krzysiek_K has quit IRC
19:51:10 *** mlncn has joined #gnucash
20:35:39 *** jralls has quit IRC
20:48:08 *** hhn10 has joined #gnucash
21:27:12 *** GabrieleV has quit IRC
21:27:12 *** GabrieleV_ has joined #gnucash
21:27:15 *** GabrieleV_ is now known as GabrieleV
21:49:00 *** mlncn has quit IRC
21:49:21 *** mlncn has joined #gnucash
21:54:06 *** mlncn has quit IRC
22:02:34 *** warlord has joined #gnucash
22:02:34 *** gncbot sets mode: +o warlord
22:10:58 *** andy has joined #gnucash
22:11:26 *** fell has joined #gnucash
22:11:26 *** gncbot sets mode: +o fell
22:21:38 *** jralls has joined #gnucash
22:21:38 *** gncbot sets mode: +o jralls
22:34:08 *** fell_ has joined #gnucash
22:34:09 *** gncbot sets mode: +o fell_
22:35:57 *** lmat has joined #gnucash
22:36:14 <lmat> jralls: howdy! Sorry, I hopped in and out earlier :) but I see your question in my log.
22:36:35 <lmat> let's see...about kvp_value, I'm not sure what the problem is.
22:37:53 <lmat> checking
22:38:21 <lmat> But, the reason I just signed in (and the reason I signed on earlier) is because I tried to make an earlier tag (on the maint branch I guess), and got an error about deprecated dbi function usage.
22:39:02 <lmat> I guess I upgraded my libdbi-dev and some functions are marked deprecated? This causes an error with current compilation flags. I haven't git fetch;ed recently, though, so if this has been addressed, don't bother wasting your breath :)
22:41:07 *** jralls has quit IRC
22:41:25 *** fell_ has quit IRC
22:42:38 <lmat> Oh yeah
22:43:22 *** fell has quit IRC
22:43:25 <warlord> lmat: what version did you upgrade from/to?
22:43:29 <lmat> @tell jralls I remember the problem. It has to do with a scm call. The test "Option DB Load from KVP" (test-option-util.c) fails, and I'm not sure
22:43:29 <gncbot> lmat: The operation succeeded.
22:43:32 <lmat> warlord: dunno... I can check
22:43:56 <lmat> ah, I have libdbi-dev 0.9.0-1
22:44:19 *** mlncn has joined #gnucash
22:44:22 <warlord> There are definitely issues between 0.8 and 0.9, but I *thought* we fixed them.
22:44:28 <warlord> Maybe only in master, not maint?
22:44:39 <lmat> I don't know what I used to have, or even *if* I upgraded it. I'm assuming I did upgrade it because I'm now getting errors.
22:44:52 <lmat> warlord: Good to know this is trodden soil :) Yeah, it's likely that it's fixed in master, not maint.
22:45:03 <warlord> dunno.
22:45:08 <warlord> Anyways, I need to head to bed
22:45:15 *** mlncn has quit IRC
22:45:15 <lmat> warlord: I was trying to build a stable build to show my friend today at church, and had that error, but I don't get that error when building from master.
22:45:21 <lmat> warlord: g'night!
22:45:30 <warlord> sorry ... and good luck.
22:45:36 <lmat> warlord: Better get sleep when you can! (Does the little one still wake you up?)
22:46:09 <warlord> no. but i'm away on a biz trip.. have an 8:30am meeting tomorrow.
22:46:14 <warlord> and a 9:30 demo on tues
22:46:22 <warlord> (first trip since the baby was born)
22:46:27 <lmat> ah, nice. Sleep well, then!
22:47:05 <warlord> Thanks. But yes, the baby actually sleeps; the issue at home is that the two kids' schedules are "different", making it hard for my wife to handle both.
22:47:22 <warlord> It'll stabilize to something better in a few months, I think (hope)
22:47:41 <warlord> anyways, ttyl
22:48:59 <lmat> yup, the schedules should match up sooner than later :)
23:12:07 *** jralls has joined #gnucash
23:12:08 *** gncbot sets mode: +o jralls
23:14:30 *** lmat has quit IRC
23:19:41 *** jralls has quit IRC
23:24:58 *** fell_ has joined #gnucash
23:24:58 *** gncbot sets mode: +o fell_
23:37:01 *** Coderjoe has quit IRC
23:42:44 *** Coderjoe has joined #gnucash
23:55:59 *** ErKa has joined #gnucash