2013-01-30 GnuCash IRC logs
00:15:48 *** arrainey has quit IRC
00:22:01 *** TommyT has quit IRC
00:33:15 *** notam2000 has quit IRC
00:39:16 *** ErKa has joined #gnucash
00:42:56 *** ftjjk has joined #gnucash
00:44:02 *** arrainey has joined #gnucash
00:54:08 *** arrainey has quit IRC
00:54:41 *** arrainey has joined #gnucash
00:59:44 *** arrainey has quit IRC
01:00:26 *** arrainey has joined #gnucash
01:05:30 *** arrainey has quit IRC
01:06:06 *** arrainey has joined #gnucash
01:11:11 *** arrainey has quit IRC
01:11:59 *** arrainey has joined #gnucash
01:17:02 *** arrainey has quit IRC
01:17:42 *** arrainey has joined #gnucash
01:22:46 *** arrainey has quit IRC
01:23:18 *** arrainey has joined #gnucash
01:28:21 *** arrainey has quit IRC
01:29:02 *** arrainey has joined #gnucash
01:34:04 *** arrainey has quit IRC
01:34:38 *** arrainey has joined #gnucash
01:39:16 *** calbasi has joined #gnucash
01:44:27 *** arrainey has quit IRC
01:52:23 *** ftjjk has quit IRC
03:16:39 *** fell_ has joined #gnucash
03:16:39 *** gncbot sets mode: +o fell_
03:27:57 *** ErKa has quit IRC
03:39:02 *** nafg has joined #gnucash
05:23:25 *** cash has joined #gnucash
05:23:37 *** ftjjk has joined #gnucash
05:26:38 <cash> Hello, I'd like to set up gnucash and enter projected income for projects or customers that are planned but not solid yet. Can I do this in GnuCash? Thank you!
05:30:03 <cash> p.s. this is because I just want to get a handle on cashflow. So I want to enter accounts receivable, projected income (quotations and when payment can be expected), payment promises, items I need to pay (rent, etc.), and loans. If I can also make a budget for a project that would be cool too.
05:37:47 <cash> Hello?
05:48:35 *** cash has quit IRC
05:51:02 *** gjanssens has joined #gnucash
05:51:02 *** gncbot sets mode: +o gjanssens
05:53:20 *** Jimraeh1 has quit IRC
06:03:03 *** fell_ is now known as fell
06:04:41 *** fabior has joined #gnucash
06:17:17 *** ftjjk has quit IRC
06:24:40 <gjanssens> warlord: when you're available, can you ping me?
06:25:03 <gjanssens> The macro we configured yesterday for the mail generation (FULLDIFFS) doesn't seem to work
06:28:30 *** nomeata has joined #gnucash
07:00:07 *** fabior has quit IRC
07:27:52 *** Jimraeh1 has joined #gnucash
07:50:47 *** fabior has joined #gnucash
07:58:24 *** nomeata has quit IRC
08:09:15 *** benoitg has joined #gnucash
08:17:50 *** benoitg has quit IRC
08:27:01 <warlord> gjanssens: ping
08:27:11 <warlord> What happens with it?
08:28:25 <gjanssens> warlord: pong
08:28:35 <gjanssens> It's not being expanded
08:28:49 *** nafg has quit IRC
08:28:56 <gjanssens> I can see this because the mail script gives a weird error:
08:29:18 <gjanssens> remote: /var/lib/gitolite3/repositories/testing.git/hooks/post-receive-email: line 664: printf: ab70d203b281aa05a1b6921a09ff02f40fc08153: invalid number
08:29:19 <gjanssens> remote: /var/lib/gitolite3/repositories/testing.git/hooks/post-receive-email: line 664: 0,000000ULLDIFFS: command not found
08:29:36 <gjanssens> I checked the script, which on line 664 says
08:29:53 <gjanssens> eval $(printf "$custom_showrev" $onerev)
08:30:20 <gjanssens> Clearly, customrev is expanded to %FULLDIFFS here, instead of the command it was supposed to be
08:30:35 <gjanssens> So the expansion fails at the gitolite level
08:31:20 <gjanssens> In conf/gitolite.conf, I have added
08:31:22 <gjanssens> config hooks.showrev2 = %FULLDIFFS
08:31:31 <warlord> Yeah, I see that.
08:31:37 <warlord> And that's what the docs say to do..
08:31:45 <gjanssens> Indeed.
08:31:47 <warlord> Maybe it's a bug in gitolite?
08:32:08 <gjanssens> Did you double check quotes and stuff in the rc file ?
08:32:33 <gjanssens> And you did use FULLDIFFS, not FULLDIFS ?
08:32:42 <warlord> I copied the rc file exactly from the webpage.
08:32:47 <gjanssens> I just mean, no typos or the like
08:32:59 *** notam2000 has joined #gnucash
08:33:07 <warlord> RC has:
08:33:10 <warlord> SAFE_CONFIG =>
08:33:10 <warlord> {
08:33:10 <warlord> FULLDIFFS => "git show -C %s; echo"
08:33:10 <warlord> },
08:33:23 <gjanssens> Yes, that should be it
08:33:40 <gjanssens> Do you think we have to run gitolite setup after changing this config file ?
08:34:22 <warlord> I dont think so. But I can..
08:35:20 <warlord> Hmm, the gnucash-docs repo config has:
08:35:22 <warlord> showrev2 = %FULLDIFFS
08:35:40 <warlord> (even after I re-run gitolite setup)
08:35:59 <gjanssens> That doesn't seem right
08:36:31 <gjanssens> Can gitolite be run in a more verbose or debug mode ?
08:39:22 <warlord> Maybe..
08:40:21 <warlord> I dont see a way...
08:41:58 <gjanssens> Looking at the gitolite code, I suspect there isn't one
08:44:10 <gjanssens> The SAFE_CONFIG parameter is processed in src/triggers/post-compile/update-git-configs
08:44:23 <gjanssens> which is triggered normally in POST_COMPILE
08:45:47 <gjanssens> perhaps you have to run a gitolite compile instead
08:47:31 <warlord> I just ran 'gitolite setup' by hand, but there is no verbose mode.
08:48:17 <gjanssens> Can you also try "gitolite compile" ?
08:51:21 <warlord> I see no reference to "SAVE_CONFIG" in any of the .pm files
08:51:55 <warlord> I only see a reference to UNSAFE_PATT
08:52:09 <warlord> So maybe the docs are ahead of the version I have installed?
08:52:28 <gjanssens> It's SAFE_CONFIG, not SAVE_CONFIG
08:53:02 <gjanssens> I see it in gitolite I downloaded yesterday
08:53:41 <gjanssens> But from the gitolite master repo on github, it appears to be added only 23 days ago.
08:53:56 <gjanssens> So yes, perhaps it's not there yet in your gitolite clone
08:54:41 <warlord> Sorry, that was a typo here -- I have it correct in the config.
08:54:57 <warlord> Yeah, I am using the 3.2 release.
08:55:05 <warlord> (not master)
08:55:48 <gjanssens> Hmm, looked at the gitolite history, it's indeed a bleeding edge feature
08:56:04 <gjanssens> gitolite 3.3 was released end of December
08:56:18 <warlord> Was this feature in 3.3?
08:56:50 <gjanssens> the SAFE_CONFIG only added January 7
08:56:52 <gjanssens> So no
08:57:02 <warlord> Okay, so not worth updating..
08:57:27 <gjanssens> So for now we'll have to depend on UNSAFE_PATT
08:57:31 <warlord> I guess I need to adjust UNSAFE_PATT for now, then
08:57:36 <gjanssens> Right
08:58:31 <gjanssens> You can leave the SAFE_CONFIG section in the rc, so we can use it once it's officially released
08:58:33 <warlord> Give me a sec.. Then you can pull and update the conf
08:58:38 <warlord> Yep.
08:59:15 <gjanssens> Oh, and while you're updating the rc, can you also allow git config parameter gitweb.description ?
08:59:35 <gjanssens> Our hook depends on it to determine the name of the repo
08:59:52 <gjanssens> Even though we're not using gitweb itself
09:01:39 <warlord> Okay, try now.
09:02:38 <gjanssens> Making necessary changes right now...
09:03:50 <warlord> k
09:07:12 <gjanssens> Works like a charm now
09:07:44 <gjanssens> The diff is properly generated and sent by mail
09:08:03 <gjanssens> The repo name is used properly in the mail
09:08:41 <gjanssens> I'll forward two example mails to the devel list for people to give feedback
09:09:18 <gjanssens> The default wording is fairly long-winded, so perhaps others have suggestions for a more compact output
09:10:45 <warlord> Great!!
09:15:49 <warlord> I like the 'repo @all' :)
09:16:07 <gjanssens> Yes, that makes configuration much easier
09:16:12 <warlord> Indeed.
09:16:22 <warlord> I wonder why you need that 'echo' for the second one?
09:17:18 <gjanssens> I'm not really sure, but I would guess it is to force git not to invoke a pager for long listings
09:17:59 <gjanssens> That or the output of that command doesn't always end in a newline
09:18:08 <gjanssens> Probably the second
09:21:33 <warlord> Yeah, that's what I was thinking too
09:23:42 <warlord> Granted, I don't think we want to turn on email until we turn off svn
09:23:49 <warlord> But it's good to know that it works.
09:25:49 <gjanssens> Sure.
09:26:13 <gjanssens> But for example the git-helper-scripts has no svn base, so no mails are sent for changes
09:26:23 <gjanssens> That could be enabled already for example
09:26:43 <gjanssens> Also once the website works based on git, I'd turn off svn for that repo as well
09:27:06 <gjanssens> So that will need the mails sooner than gnucash and gnucash-docs
09:27:21 <gjanssens> Luckily it's easy to configure per repo
09:28:56 <warlord> Yep
09:31:53 <warlord> Glad we seem to have it all under control.
09:32:09 <warlord> I'll note that git-helper-scripts already has an origin of github
09:32:21 <warlord> We just need to reset the other three and then push out.
09:33:03 *** benoitg has joined #gnucash
09:43:06 *** benoitg has quit IRC
09:49:10 <gjanssens> Do you think we need to reset the git-helper-scripts repo ?
09:49:26 <gjanssens> It's a clean clone from github, right ?
09:49:33 <warlord> No, we do not need to reset.
09:49:39 <warlord> gitolite is a direct clone
09:50:00 <warlord> (although I can't 'pull' from it because it is a bare clone)
09:50:07 <warlord> (... or fetch)
09:50:31 <gjanssens> Didn't know that...
09:50:54 <warlord> Maybe it is me...
09:51:10 <warlord> But so long as there have not been any new commits we should be fine with switching direction.
09:52:19 <gjanssens> Yes, that's how I understand it as well
10:28:31 *** kpreid has quit IRC
10:29:11 *** kpreid has joined #gnucash
10:29:51 *** arrainey has joined #gnucash
10:33:42 *** ses1984 has quit IRC
10:39:25 *** TommyT has joined #gnucash
10:50:22 <warlord> gjanssens: does the new git mail script have the 'BP' logic?
10:50:57 <gjanssens> Hmm, nope
10:51:03 <gjanssens> Will have to look into that
10:52:12 <gjanssens> That will be difficult to handle in this context
10:52:22 <gjanssens> BP is a per commit item
10:52:38 <gjanssens> While the mails are per push (one push can have multiple commits)
10:54:49 <warlord> oh. hmm
10:57:14 <gjanssens> I probably still can scan for a BP message
10:57:23 <gjanssens> and mark the whole push as AUDIT
10:58:03 <gjanssens> At least that suggests the push has *some* commits that were requested for backporting
10:58:48 <gjanssens> On the other hand, once svn is out we could depend more on git's branching power
10:59:11 <gjanssens> and simply define a backports branch
10:59:27 <warlord> hmm
10:59:30 <gjanssens> to which each dev pushes all his commit he wants backported
10:59:40 <gjanssens> (cherry-picks)
11:00:18 <gjanssens> git is pretty smart at merging
11:00:23 <gjanssens> much smarter than svn
11:00:44 <gjanssens> so there's less reason to avoid this in git than in svn
11:01:03 <gjanssens> but this is again a topic for the devel list as well IMO
11:07:09 *** arrainey has quit IRC
11:07:47 *** arrainey has joined #gnucash
11:14:23 *** fabior has quit IRC
11:18:27 <gjanssens> Mail sent to devel for further public discussion
11:23:38 *** arnotixe has quit IRC
11:23:47 <warlord> Okay.
11:25:05 <warlord> I think it's more going to be about a process issue. We'd already have a branch for e.g. 2.6 -- it seems silly to me to have yet another branch to queue up changesets to backport to the 2.6 branch.
11:33:11 *** arrainey has quit IRC
11:34:08 *** arrainey has joined #gnucash
11:40:15 <gjanssens> Probably.
11:40:27 <gjanssens> Our current process is very much adapted to svn
11:40:38 <gjanssens> In particular its weak merging facilities
11:41:03 <warlord> True..
11:41:11 <gjanssens> Most projects I have come across that use git have a very different branching strategy in their process
11:41:14 <warlord> git will just make cherry-picking easier, hopefully.
11:41:40 <gjanssens> Which seem to rely heavily on powerful merging tools
11:41:56 <gjanssens> But I'm not actively involved in any such project
11:42:11 <warlord> I think having branches for stable release series still makes sense
11:42:12 <gjanssens> I think John is and probably Christian also
11:42:44 <gjanssens> I hope they can give feedback from their experience in those other projects
11:43:11 <warlord> me too
11:43:14 <gjanssens> re branches for stable release series: I agree
11:43:32 <gjanssens> I think one important difference between git and svn is
11:43:48 <gjanssens> that for git the trunk/master branch is just another branch in the bunch
11:44:01 <gjanssens> while svn seems to treat it as main path
11:44:13 <gjanssens> with branches being derived from it
11:44:26 <gjanssens> so in git all branches are treated equal
11:44:35 <warlord> Well, for svn it's just a convention. In SVN everything is a "directory" tree. trunk happens to be one such directory in the tree.
11:44:52 <warlord> There isn't really anything special about 'trunk' in svn -- it's just a layout convention.
11:44:57 <gjanssens> Yes, that's more correct
11:45:12 <warlord> One could easily have set up 'trunk' and 'branches/*' in the same place.
11:45:27 <gjanssens> but that also implies svn doesn't really have an infrastructure for branches
11:45:32 <warlord> but then 'trunk' gets hidden. Moreover, there is nothing special about the name 'trunk'
11:45:32 <gjanssens> other than a directory structure
11:45:41 <warlord> That is true.
11:53:14 *** fell_ has joined #gnucash
11:53:14 *** gncbot sets mode: +o fell_
11:58:27 *** ErKa has joined #gnucash
12:01:53 *** fell has quit IRC
12:21:01 *** calbasi has left #gnucash
12:36:29 *** TommyT has quit IRC
12:36:31 *** arrainey has quit IRC
12:37:09 *** arrainey has joined #gnucash
12:41:40 *** jmd has joined #gnucash
12:45:31 *** ecocode has quit IRC
12:49:39 <gjanssens> warlord: what is the syntax to clone the gitolite repos as a non-developer ?
12:50:37 <gjanssens> The gitolite conf shows the repos are publically readable, but I'm not sure if ssh will allow access without a public key
12:51:40 <gjanssens> I tried "git clone git@code.gnucash.org:gnucash-htdocs"
12:52:02 <gjanssens> in an account without my ssh keys configured
12:52:14 <gjanssens> ssh asked for a password in this case
12:53:34 <gjanssens> Same with "git clone ssh://git@code.gnucash.org/gnucash-htdocs"
12:56:07 *** fabior has joined #gnucash
12:56:35 <warlord> gjanssens: you cannot right now. Only ssh:// URLs are enabled.
12:56:50 <warlord> .. so only devs can access gitolite right now
12:57:06 <warlord> I don't have "git://" read only urls supported, yet.
12:57:59 <warlord> Also, if ssh asks for a password and you enter one it's possible that it will block your IP Address.
12:58:19 <warlord> (I think the idea was that non-devs would pull from github)
13:05:14 *** rpg_ has joined #gnucash
13:06:43 <gjanssens> This was an experiment for Linas' webserver
13:07:02 <gjanssens> I can pull from github there as well,
13:07:17 <gjanssens> but given the trigger for the website and the push to github are in the same hook
13:07:25 <gjanssens> I fear race conditions in there
13:12:48 *** arrainey has quit IRC
13:13:21 *** arrainey has joined #gnucash
13:14:57 *** fabior has quit IRC
13:23:20 <warlord> just push to github first. that's synchronous
13:28:38 *** arrainey has quit IRC
13:29:13 *** arrainey has joined #gnucash
13:29:14 <warlord> the svn->gitolite is async, but gitolite->github is sync.
13:32:39 *** aqua___ has joined #gnucash
13:36:08 <gjanssens> ok I'll try that
13:43:59 *** fabior has joined #gnucash
13:45:18 <gjanssens> But only after the github reset
13:45:40 <gjanssens> I prefer to work with the correct hashes first time
13:49:10 *** aqua___ has quit IRC
13:51:56 <warlord> * nods *
13:52:15 <warlord> no rush on that one.. It can still use svn until after the github reset
14:00:05 *** Topcat has joined #gnucash
14:27:43 *** TommyT has joined #gnucash
14:32:06 *** Topcat has quit IRC
14:36:02 *** Bodhi-Baum has joined #gnucash
14:44:07 *** benoitg has joined #gnucash
14:54:51 *** jmd has quit IRC
15:01:31 *** moldy has quit IRC
15:07:07 *** moldy has joined #gnucash
15:41:13 *** arrainey has quit IRC
15:41:47 *** benoitg has quit IRC
15:41:50 *** arrainey has joined #gnucash
15:41:56 *** benoitg has joined #gnucash
15:54:19 *** arnotixe has joined #gnucash
16:04:45 *** nomeata has joined #gnucash
16:11:01 *** nomeata has quit IRC
16:13:12 *** rpg_ has quit IRC
16:24:39 *** notam2000 has quit IRC
16:28:14 *** rpg_ has joined #gnucash
16:37:51 *** arrainey has quit IRC
16:38:26 *** arrainey has joined #gnucash
16:42:28 *** TommyT has quit IRC
16:58:04 *** Bodhi-Baum has quit IRC
17:03:46 *** rpg_ has quit IRC
17:17:08 *** gjanssens has quit IRC
17:29:27 *** arrainey has quit IRC
17:29:54 *** notam2000 has joined #gnucash
17:29:59 *** arrainey has joined #gnucash
17:52:16 *** rpg_ has joined #gnucash
17:55:25 *** fabior has quit IRC
17:57:45 *** arnotixe has quit IRC
18:05:38 *** arrainey has quit IRC
18:06:20 *** arrainey has joined #gnucash
18:34:59 *** compywiz has quit IRC
19:05:39 *** nafg has joined #gnucash
19:31:16 *** arrainey has quit IRC
19:33:27 *** arrainey has joined #gnucash
19:42:54 *** benoitg has quit IRC
19:53:52 *** arrainey has quit IRC
19:54:27 *** arrainey has joined #gnucash
20:09:44 *** arrainey has quit IRC
20:10:18 *** arrainey has joined #gnucash
20:12:11 *** ErKa has quit IRC
21:27:58 *** arrainey has quit IRC
21:28:42 *** arrainey has joined #gnucash
21:33:28 *** TommyT has joined #gnucash
21:34:15 *** nafg has quit IRC
21:45:18 *** benoitg has joined #gnucash
21:47:38 *** TommyT has quit IRC
22:13:55 *** arrainey has quit IRC
22:14:29 *** arrainey has joined #gnucash
22:14:40 *** rpg_ has quit IRC
22:29:43 *** TommyT has joined #gnucash
22:41:53 *** arrainey has quit IRC
22:49:48 *** fell_ has quit IRC
23:42:42 *** notam2000 has quit IRC
23:43:23 *** nafg has joined #gnucash