2007-02-17 GnuCash IRC logs
00:10:48 *** dbr has quit IRC
00:30:28 *** rhowe has quit IRC
02:15:47 *** hampton is now known as hampton|away
02:30:01 *** motin has joined #gnucash
02:30:58 *** motin_ has joined #gnucash
02:58:18 *** motin_ has quit IRC
02:58:21 *** motin has quit IRC
03:16:32 *** gander has joined #gnucash
05:08:32 *** rhowe has joined #gnucash
05:36:05 *** gander has left #gnucash
06:05:10 *** ErKa has joined #gnucash
06:24:26 *** |gunni| has joined #gnucash
06:25:50 *** |gunni| has joined #gnucash
06:36:47 *** IanL has joined #gnucash
06:58:36 *** prock has joined #gnucash
09:36:27 <prock> I'd appreciate a cr for the demodulification of the register-core, register-gnome, and ledger-core modules: http://paste.debian.net/22095
09:37:03 <prock> it adds one FIXME... if someone has input where this initialization should go I'd appreciate the feedback.
09:53:28 *** ErKa has quit IRC
09:55:00 *** ErKa has joined #gnucash
09:55:28 <jsled> prock: my auto-fu is weak; what's the overview of the change to demodularize something?
09:56:55 <jsled> prock: certainly it's not just removing the string 'mod' from the library name... though maybe it's ... adding the link lines and explicit init (gnc_register_init()) into the gnucash executable?
09:57:15 <prock> that's essentially what it is.
09:57:36 <jsled> Which is kinda a wash anyways, because we were treating them as shared libs already...
09:57:57 <jsled> (e.g., a wash in src/bin/Makefile.am)
09:58:25 <prock> sorry... what do you mean by "wash", and what is "auto-fu"?
09:58:27 <prock> :)
09:59:06 <jsled> auto-fu ... like kung-fu, but for autotools. :)
09:59:17 <prock> ah, very nice :)
10:00:26 <jsled> A wash ... no real change. http://www.answers.com/wash .. noun defn #13
10:01:18 <prock> yeah. The "module" -> library conversion will be mostly inobtrusive so I figured I'd tackle that first.
10:02:08 <prock> if the way in which I've done this one is adequate then I'll do the rest and then move onto the "module" -> _module_ transition.
10:02:30 <jsled> Yeah. I'm wondering if it should happen on a branch.
10:03:26 <prock> I'll be removing/adding files and such which would make it easier for me... but I know that you aren't in the habit of giving out svn accounts so...
10:03:52 <prock> I'm pretty indifferent. I'm versioning on my side anyway so whatever the dev team sayeth.
10:08:45 <jsled> What are you doing for versioning on your side? git? svk?
10:08:54 <prock> svn
10:09:02 *** motin has joined #gnucash
10:09:02 *** motin_ has joined #gnucash
10:09:10 <jsled> Oh? Just a completely separate tree?
10:09:47 <prock> yeah, gnucash trunk is an external.
10:09:56 <prock> versioning the diffs.
10:10:11 <prock> keeping notes for adds/removes
10:10:50 <prock> I've never looked at svk, would I be able to mirror the gnucash svn repository with it?
10:17:02 <jsled> yeah, it's a distributed layer on top of svn. you can make commits to a local copy of the repo, then push them back to the source.
10:17:15 <jsled> I've not looked at it, but warlord is (was) using it for a while.
10:17:32 <jsled> I've not looked at git-svn, but I understand it's a backend for git that works in a similar way.
10:20:17 *** motin has quit IRC
10:32:50 <jsled> prock: re: gnc_register_gnome_init. It should probably be after engine_init. gnucash-bin is presently as good a place as any, but we'll eventually want to factor all that higher-level app init into some function, maybe in app-utils.
10:33:22 <jsled> I guess the register is pretty specific to the gui version of the app (as opposed to a gnucash-cli or gnucash-httpd)
11:25:46 *** warlord-afk is now known as warlord
11:25:57 <warlord> I'm still using SVK for all my gnucash (and work) development..
11:26:15 <warlord> prock: what's "a cr"?
11:27:34 <chris> code review
11:27:48 <warlord> Ahh.. Thanks.
11:27:55 <warlord> (never heard that shorthand before)
11:29:19 <warlord> prock: what's the orginal source for gnc-register-gnome? I dont see that file in my tree, so it makes it hard to analyze that patch.. Could you perform a diff from before the "svn add" of gnc-register-gnome.c ?
11:30:40 <prock> gnc-register-gnome.c is renamed from gncmod-register-gnome.c
11:30:46 <prock> it's just the init funciton.
11:30:50 <warlord> prock: also, what's gnc-register-gnome.h?
11:31:17 <prock> the header for the register init function. I didn't see an other appropriate header file to put it in.
11:31:22 <warlord> prock: I realize that, the patch posted seems to assume that the file already exists. Since it's a new file I'd like to see the patch as an add.
11:31:35 <warlord> That's fine, I just didn't see it in the patch you asked to review..
11:31:56 <warlord> It feels like you did: svn add; svn commit; <change stuff>; svn diff.
11:32:12 <warlord> (or 'svn move' instead of 'add')
11:32:15 <prock> no, only svn mv and svn st
11:32:22 <prock> sorry, svn mv and svn di
11:32:54 <prock> I'll try to break up the mv into rm and then diff and then add if you like though.
11:33:58 <warlord> Hmm.. I thought that if you svn move and then edit the file, svn diff will show the resulting file, not a diff against the original form of the file... Unless you performed an "svn mv" in the repository instead of in your WC?
11:34:37 <prock> I have no repository access... you are correct about the svn mv and changes not showing up in a diff though. There are open bugs against SVN for that issue.
11:35:52 <prock> s/no repository access/read-only repository acess/g
11:36:05 <warlord> I thought you were working in your own repository?
11:36:33 <warlord> Also, I think that gnc-register-gnome.h might want to get "installed" so that a third-party app can initialize the register.
11:36:49 *** twunder has quit IRC
11:38:53 *** egsavage has joined #gnucash
11:39:49 <prock> I'm versioning what I'm doing via svn diff and making notes for adds/deletes. GnuCash trunk is pulled in via svn:externals property.
11:40:04 <warlord> Oh, I see.
11:40:35 <prock> re: installing gnc-register-gnome.h I didn't see any other headers being installed anywhere (this would be done via inst_HEADERS, correct?)
11:41:43 <warlord> see src/engine/Makefile.am for a sample.
11:41:52 <warlord> Perhaps we dont install register APIs.
11:41:58 <warlord> I suppose we could leave that for later..
11:42:55 <prock> so gncincludedir = ${GNC_INCLUDE_DIR} and then list gncinclude_HEADERS(?)
11:43:13 <warlord> Yeah.
11:43:34 <prock> ok so noted.
11:47:29 <jsled> Do we have any "third-party apps" that need gnucash headers? I thought we'd agreed to stop doing that, if we were.
11:48:17 <warlord> right now I dont think so.
11:48:25 <warlord> OTOH we do install the engine headers..
11:49:23 <chris> That's kind of silly. And it may give the wrong impression. (if anyone actually noticed.)
11:49:42 <jsled> lol. http://daringfireball.net/2007/02/macrovision_translation
11:50:09 <jsled> also: http://xkcd.com/c224.html
11:53:14 <warlord> LOL. Very cute.
11:54:41 *** IanL has quit IRC
11:57:00 <jsled> warlord: I forgot to CC chris on the mail, but he's in favor. I can't imagine anyone who hasn't yet spoken will object. Would you like to do the deed, or should I?
11:58:22 <warlord> You're welcome to do it if you like.
12:02:31 <jsled> Cool. Hey, prock, we'd like to extend svn commit privs to you, for a branch for the de-modularization work.
12:02:53 <prock> that would definately make the cr easier :)
12:03:17 <jsled> Unless you're not interested ... we'll need the following things: - an ssh public-key, your preferred username, and a name for the branch.
12:03:39 <jsled> Please mail me the pubkey
12:03:57 *** exper_ has joined #gnucash
12:04:22 <jsled> branches/de-modularize/ ?
12:04:36 <prock> either that or module-cleanup
12:04:44 <jsled> module-cleanup is better.
12:06:06 <prock> prock is fine for username and public key is on its way.
12:06:34 <jsled> is copying trunk@HEAD alirght?
12:06:59 <prock> let me check BASE:HEAD logs first...
12:07:56 <prock> yeah trunk@HEAD is ifne.
12:08:00 <prock> s/ifne/fine/g
12:08:56 *** exper has quit IRC
12:17:56 <jsled> prock: right. welcome message coming, but you should be good to go.
12:18:25 <prock> jsled et al. : thanks very much.
12:19:34 <jsled> :)
12:20:06 <warlord> welcome, prock :)
12:20:17 <prock> :)
12:22:08 <jsled> Great. Now I can run errands. :) baiw.
12:22:11 <warlord> commit early, commit often.. And make your changelogs as verbose as necessary to convey what you're doing :-D
12:36:17 <warlord> :)
12:36:44 *** |gunni| has quit IRC
12:37:20 *** |gunni| has joined #gnucash
12:41:14 *** writerz_ has joined #gnucash
12:48:08 *** exper_ has quit IRC
12:49:55 <warlord> later
12:49:59 *** warlord is now known as walord-afk
12:59:04 *** Demitar has joined #gnucash
12:59:05 *** twunder has joined #gnucash
13:20:02 *** exper_ has joined #gnucash
13:20:36 *** twunder has quit IRC
13:26:24 *** writerz_ has quit IRC
13:54:27 *** MrN has joined #gnucash
13:55:06 <MrN> hi
14:02:00 *** ural0 has joined #gnucash
14:24:48 *** ErKa has quit IRC
14:25:04 *** ErKa has joined #gnucash
14:35:27 *** writerz_ has joined #gnucash
14:42:46 *** exper_ has quit IRC
15:11:10 <prock> for scm_init_sw_engine_module/scm_init_sw_report_system_module, what does the "sw" stand for? swig?
15:20:11 <prock> hmmm... when the reports are calling gnc:module-load on the various modules are they actually getting init'd? Is this even necessary for modules that are sure to already be initialized (like gnucash/engine)? Or are they pretending that the modules are in fact more modular than they are and that a report could be generated by something other than gnucash[-bin]?
15:22:00 <chris> prock: yes, "sw" stands for "swig"
15:22:43 <chris> I don't know why I used only two characters, except perhaps because it replaced "gw" for "gwrap".
15:24:50 <chris> In general, both on the guile-side with gnc:module-load (is it gnc-module-load?) and on the C-side with (gnc_module_init?) all the modules tend to init all other modules they "depend" on.
15:26:08 <chris> Part of this is a throw back to the time when gnucash was actually a guile program, so it was more likely that some C-side modules might not have been initialized if you were calling random guile functions from a guile interpreter.
15:30:37 <prock> for my purposes, am I able to assume that the root of all things gnucash shall be gnucash[-bin]?
15:34:12 <prock> i.e., do I need to run the engine initialization functions from the .scm report?
15:37:44 <chris> you can assume that the Program Entry Point will always be in gnucash-bin. and that no reports will be called before the engine has been initialized.
15:40:34 *** motin_ has quit IRC
16:23:35 *** hampton|away is now known as hampton
16:41:39 * jsled ruthlessly slashes away at the GncFrequency widget...
16:42:39 *** twunder has joined #gnucash
16:59:09 *** twunder has quit IRC
17:25:29 *** andi5 has joined #gnucash
17:25:29 *** gncbot sets mode: +o andi5
17:27:54 *** andi6 has joined #gnucash
17:27:55 *** gncbot sets mode: +o andi6
17:36:26 *** andi5 has quit IRC
17:39:49 *** andi6 is now known as andi5
17:52:09 *** ErKa has quit IRC
18:12:37 <jsled> Hmm. Is swig-1.3.31 problematic? I'm seeing [[[
18:12:37 <jsled> /opt/gnc/unstable/share/gnucash/scm/business-prefs.scm:86:1: Wrong type argument in position 1: #<collectable-swig-pointer QofIdType * 81fb868>
18:12:38 <jsled> ]]]
18:13:10 <jsled> (after getting upgraded on thursday)
18:15:18 <andi5> hm... win32 is using swig 1.3.31 for some time now
18:15:28 <jsled> Oh rly?
18:15:33 <andi5> yep
18:15:42 * jsled makes clean, &c.
18:15:59 <hampton> yeah, but that's scheme run-time checked code, not compile time checked code.
18:16:34 <jsled> I'm seeing that right during startup, so presumably it'd be seen on win32 as well.
18:16:45 <hampton> I ran into a few problems with the group-removal process where gwrap had a glist of objects swig now had a scheme list of objects.
18:16:56 <hampton> jsled: agreed
18:19:00 <andi5> jsled: FYI, http://svn.gnucash.org/trac/changeset/15213
18:21:15 <hampton> The report options dialog needs to be cleaned up according to the HIG. What a PITA, since its dynamically generated from the scheme code.
18:22:19 * andi5 wonders how to merge (at least) three major branches....
18:22:33 <hampton> lol. two at a time. :-)
18:23:33 <andi5> this will only work as long as 3 < oo ;-)
18:35:41 *** dbr has joined #gnucash
18:38:19 <dbr> andi5: I have an interesting observation on windows installs -- if Qt4 is installed in c:\Qt\4.2.2\foo (the Trolltech default), the aqbanking setup wizard does not work for aqofxconnect lib. If Qt4 is installed in c:\soft\Qt\4.2.2\foo (install.sh-style default), then the wizard does work (at least for me).
18:39:00 <andi5> dbr: will test real soon
18:39:08 <dbr> thanks.
18:40:08 <andi5> dbr: i want to spawn the qt3-wizard from our druid... i guess this should work tomorrow (hope so)
18:41:24 <dbr> andi5: good news. I've been so frustrated with the wizard not even working separately, that I'm pretty happy at the moment, even without the spawn vs. fork switch.
18:41:57 *** benoitg has joined #gnucash
18:44:01 <dbr> andi5: also the setup in gnucash works after the separate wizard run so that I can retrieve transactions in ofxdirectconnect sessions in gnucash under windows now.
18:44:08 <andi5> lol, it is qt3-wizard..exe now :)
18:44:43 <dbr> I didn't even look this afternoon...
18:44:56 <andi5> dbr: so your only issue is the one with the installation directory right now?
18:45:58 <dbr> andi5: yes. I haven't tested the cert storage much at all, but I'm not too worried about that.
18:47:35 <andi5> ok... spawning seems to work basically
18:48:42 <andi5> dbr: i have installed qt to c:\qt\4.2.2.... what do you mean by "does not work for aqofxconnect lib"?
18:49:09 <andi5> i can activate all backends, whee :)
18:50:30 <chris> jsled: IIRC, that code is run on startup every time, so it has probably worked before.
18:50:34 <dbr> andi5: when i used the default Qt install location, then build aqbanking and try to run qt3-wizard, the wizard crashes during a bank connection (my note to aqbanking-devel yesterday). If I uninstall Qt, reinstall it in /c/soft/
18:50:54 <dbr> Qt, rebuild aqbanking, then the wizard connects correctly
18:51:10 <dbr> sorry about the stray line break
18:51:24 <andi5> dbr: i should really start to read all those emails ;-)
18:52:02 *** dbr is now known as dbr-afk
18:52:07 <dbr-afk> dinner time biab
18:52:09 <andi5> @tell prock welcome to the dev gene pool :)
18:52:09 <gncbot> andi5: The operation succeeded.
18:53:06 <chris> jsled: also, the type in the error messages is really the correct type, I believe.
18:53:39 <chris> so, it may be a case of finding a gwrap version of that scm file.
19:32:16 *** egsavage has quit IRC
19:33:30 *** dbr-afk is now known as dbr
19:38:21 *** egsavage has joined #gnucash
19:43:56 *** ural0 has quit IRC
19:44:34 *** egsavage has quit IRC
19:47:17 *** IanL has joined #gnucash
19:52:00 *** egsavage has joined #gnucash
20:08:12 *** benoitg has left #gnucash
20:24:59 <jsled> The hell? <http://bugzilla.gnome.org/show_bug.cgi?id=387102#c4>
20:26:06 <andi5> huh
20:39:33 <chris> Is this something I should be worried about?
20:41:31 *** |gunni| has quit IRC
20:55:48 <andi5> dbr: you should be able to deinstall qt and install to another location without recompiling aqbanking... can you switching between works and does_not_work that way?
20:56:22 <dbr> andi5: I'll give it a try. It'll only take minutes instead of hours
21:11:34 <dbr> damn, I hate windows.
21:12:37 <andi5> lol
21:14:57 <dbr> I forgot to reboot or otherwise clear the cache, so when I reinstalled, the demo was looking in the old location for files. Not much of a controlled test.
21:15:35 <andi5> should that make me happy or not? ;-)
21:16:22 <dbr> well, you just get to wait longer for the first part of the answer...
21:16:54 <andi5> dbr: sometimes it really helps to open a new msys shell.... e.g. new environment variables will only be available in new terminals after some package installation
21:17:41 <dbr> dbr: except in this case I'm only dealing with windows so far -- the Qt installer
21:17:52 <dbr> now I'm talking to myself. arghhh
21:17:58 <MrN> n8
21:18:17 *** MrN has quit IRC
21:20:09 <andi5> dbr: do you know (1) a way to make qt3-wizard fail, i.e. exit with a non-zero status or (2) an ofxdirectconnect test server? (i guess cstim would know it otherwise)
21:22:36 <dbr> andi5: (1) has been totally consistent (failing) for me with Qt in the default location. (2) I don't know of any ofxdirectconnect test servers open to the public. I just use my accounts and hope that the banks don't get mad. Happily Quicken makes people connect way more often than logical so my connections aren't that big a deal.
21:26:15 *** IanL has quit IRC
21:36:47 <andi5> dbr: thanks
21:38:22 <dbr> andi5: and now I'm even more confused -- uninstalled, rebooted, installed in default location, checked for environment variables in the system control panel, and Qt Assistant still looks for files in /c/soft/Qt/doc instead of /c/Qt/4.2.2/doc. ???
21:39:36 <dbr> It's basing something on the base mingw path, I think. Oh well, I'll give aqbanking a try now, just for fun.
21:39:58 <andi5> dbr: just for fun, have you tried searching the registry?
21:40:40 <dbr> andi5: nope. I'll go poke around there, too. Anything better than regex available?
21:41:12 <andi5> regex or regedit?
21:42:42 <dbr> andi5: maybe I shouldn't be doing this tonight. tongue tied, etc.
21:43:01 <andi5> :-)
21:43:22 <dbr> andi5: anyway, the wizard now works. So much for my ability to repeat failure instead of find new ways to 'have fun'
21:53:03 <dbr> andi5: I found (at least 1) place in the registry that has things like the docpath for Qt screwed up. Maybe that's been my problem all along. ? Especially since I can't repeat the connection failure now that I've had a successful connection.
21:54:08 <andi5> dbr: i do not know whether the _doc_path is important at all, but other paths may be :)
21:54:39 <dbr> andi5: yeah. that's just the first one I found. I'll be digging.
21:56:45 <andi5> dbr: happy digging... i am tired---see you
21:57:15 <dbr> andi5: goodnight (or morning where you are, right?)
21:57:56 *** andi5 has quit IRC
22:05:31 *** Wilddev has joined #gnucash
22:05:32 *** gncbot sets mode: +o Wilddev
22:23:53 *** egsavage has quit IRC
22:49:40 *** writerz_ has quit IRC
23:07:55 *** foo has quit IRC
23:38:24 *** walord-afk is now known as warlord
23:38:49 <warlord> jsled: another potential swig issue: if you changed the version of swig you probably need to regenerate the swig-runtime.h, which is NOT automatically regenerated.
23:44:23 <Wilddev> hey warlord
23:44:33 <warlord> hiya Wilddev
23:44:46 <warlord> damn, I forgot that there are a few more patches for backport.
23:44:59 <Wilddev> oops
23:45:12 <Wilddev> yeah I'll be starting on that in the morning probably
23:45:21 <warlord> What time?
23:46:34 <Wilddev> not sure exaclty
23:46:42 <Wilddev> I'll come here b4 I start tho
23:47:05 <Wilddev> likely later rather than earlier tho
23:47:09 <warlord> Okay.. I'll try to backport.. Someone needs to audit my TRL->TRY patch.
23:47:11 <warlord> Okay..
23:47:17 <Wilddev> ok
23:49:15 *** dbr has quit IRC
23:50:54 <hampton> warlord: looks good to me
23:51:06 <warlord> hampton: thanks!
23:51:21 <warlord> I've never done a change like that before, but it looked pretty straightforward.
23:53:01 <hampton> the mapping table in gnc-commodity.c works well for a 1:1 mapping of old to new currency
23:53:34 <warlord> is there ever not a 1:1 mapping?
23:54:01 <hampton> dollar to dollar, not currency to currency
23:54:29 <warlord> Oh.
23:54:36 <hampton> i.e. for 100 old = 100 new, not for 100 old = 1 new
23:55:06 <warlord> Right.