2007-02-06 GnuCash IRC logs

00:25:11 *** franz has quit IRC
00:51:26 *** digdug has joined #gnucash
00:52:35 <digdug> is it easy to compile ofx support into the windows build (I've been able to compile normally using the instructions in the wiki)?
02:01:42 <digdug> well, from my own attempts to manually configure and compile opensp and libofx, I'm assuming the answer is no, or that it's currently not possible
02:34:56 *** solidago_ has quit IRC
02:36:16 *** solidago_ has joined #gnucash
04:32:05 *** conrad has quit IRC
04:42:26 *** |gunni| has joined #gnucash
04:51:54 *** cstim has joined #gnucash
04:51:55 *** gncbot sets mode: +o cstim
04:54:51 <cstim> digdug: how far did you succeed with opensp and libofx on win32?
05:39:48 *** nomeata has joined #gnucash
06:02:18 * cstim managed to compile opensp (after some Makefile manipulation) to get a libosp-5.dll.
06:10:32 *** conrad has joined #gnucash
06:18:37 <digdug> I never tried editing files, I just tried to compile them as downloaded
06:47:51 <cstim> digdug: but did opensp compile? With mingw or with MSVC?
06:56:27 <cstim> good news from the gtk front: The missing feature of per-cell tooltips in GtkTreeView will be available in next gtk, http://bugzilla.gnome.org/show_bug.cgi?id=80980
06:59:06 <cstim> Concerns this: http://svn.gnucash.org/trac/browser/gnucash/trunk/src/gnome-utils/dialog-options.c#L501
08:03:22 *** twunder has joined #gnucash
08:27:53 *** ceplma has joined #gnucash
08:55:39 *** warlord-afk is now known as warlord
08:55:47 <warlord> cstim: sweet!
08:55:56 <warlord> (about the tooltips)
08:56:03 <cstim> yes
08:58:00 <cstim> warlord: You know the function mkstemp() and how it's supposed to be used?
08:59:54 <warlord> cstim: see http://pastebin.ca/342685
09:00:28 <cstim> :-) yes, I've found that one as well. it's unavailable on win32. But that's not (too much of) an issue.
09:00:48 <cstim> rather, it's not supposed to be used like this: http://libofx.cvs.sourceforge.net/libofx/libofx/lib/ofx_preproc.cpp?view=markup scroll down to line 89
09:01:47 <cstim> a classical "NO NO" example of how this function should never be used
09:02:16 <warlord> Um, yeah! DEFINITELY wrong.
09:02:22 <warlord> That's a fd leak.
09:02:43 <cstim> yes. And a race condition.
09:03:09 <warlord> That, too.
09:03:21 <cstim> this part of the code was unchanged from the original 2002 code.
09:04:27 <cstim> makes me wonder why this never showed up as an issue.
09:05:31 <warlord> Probably because leaking the fd isn't a big issue on Linux.
09:05:45 <warlord> (and wont be noticed in most cases)
09:07:28 <cstim> anyway. I've got libofx compiled on win32, and aqbanking's ofx-directconnect parts as well.
09:08:11 <cstim> aqbanking needed some fixing, so that's currently only in its SVN, not yet released.
09:12:12 <warlord> Okay. Cool!
09:12:43 * cstim is testing to build gnucash --enable-ofx ...
09:13:47 <warlord> :)
09:19:12 <warlord> Regardless... Cool!
09:19:34 <warlord> (and if you find bugs in libofx you should contact benoit to get a new Libofc release)
09:24:32 <cstim> huh. That was easy. gnucash just compiled with --enable-ofx.
09:26:39 <warlord> Sweet!
09:26:53 <cstim> @tell andi5 Should opensp and libofx go into install.sh as well?
09:26:53 <gncbot> cstim: The operation succeeded.
09:28:46 <warlord> cstim: I think so..
09:29:00 <cstim> heh. another ~40MB on harddisk.
09:29:16 <warlord> Eh, disk is cheap.
09:29:18 *** bash has joined #gnucash
09:29:36 <warlord> How much does it increase the installer?
09:29:54 <cstim> no idea.
09:30:03 <cstim> ask me again in 8 hours from now :-)
09:30:35 <cstim> (I only built the src/import-export/ofx subdir)
09:30:53 <warlord> Oh. Okay. :)
09:37:52 *** aj has quit IRC
09:43:22 *** aj has joined #gnucash
09:47:54 <warlord> @tell andi5 FYI, make check worked for me on SVN/trunk on Linux.
09:47:54 <gncbot> warlord: The operation succeeded.
09:49:38 <cstim> gncbot: check succeeded? That's very good.
09:49:38 <gncbot> cstim: Error: "check" is not a valid command.
09:51:01 <warlord> cstim: on Linux, yes. It still fails on win32 in src/bin/test because gnucash-build-env hasn't been updated.
09:55:39 <cstim> sounds okay.
09:57:00 *** rlaager has quit IRC
10:10:27 <cstim> warlord: re missing GUILE_MODULE_AVAILABLE macro: I clearly recall running into the same issue after I upgraded to opensuse10.2 2 weeks ago. I'll have to look whether I needed any particular package installed. Will report back later today.
10:10:53 <warlord> cstim: Okay..
10:10:59 <warlord> I'm wondering if it's a bug in autogen.sh..
10:11:19 <warlord> (and I'm REALLY beginning to lean towards using autoreconf -siv instead of our current autogen.sh)
10:17:29 <hampton> LOL. So get this. I bought a new Sony television, and I'm looking through all the paperwork that come with it and what do I find?
10:18:01 <hampton> I find the "Required Public Statement for GPL/LGPL Licensed Software Used In This Television."!!!!!
10:18:37 <hampton> That contains a list of software and a full copy of the GPL.
10:18:53 <jsled> how do you get the source? URL? DVD in the mail via request?
10:19:29 <hampton> url. http://www.sony.com/linux
10:20:06 <hampton> The gpl code is the Monta Vista kernel, busybox, and insmod. lgpl is a list of 7 libraries.
10:20:27 <cstim> heh
10:22:35 <warlord> Sweeit
10:34:00 <warlord> jsled: FYI, creating accounts in the QIF importer is pretty easy (right now).. You just say the GNC Account name and the importer will do the accounts. It doesn't pop up the New Account Dialog like OFX/HBCI.
10:35:44 <jsled> warlord: right; I was trying to say something along those lines. More that it's not the best way to go about it.
10:38:42 <warlord> It's why I pointed out the File -> New File
10:38:46 <warlord> druid
10:39:03 <jsled> As did I.
10:56:41 <warlord> *nods*
11:19:32 <cstim> There we are. gnucash/win32 with ofx import.
11:26:57 *** prock_ has joined #gnucash
11:35:22 *** prock has quit IRC
11:38:07 *** prock_ is now known as prock
11:56:30 *** cstim has quit IRC
12:13:54 *** ceplma has quit IRC
12:42:01 *** esodan has joined #gnucash
12:59:40 *** ceplma has joined #gnucash
13:07:53 *** bonez39 has quit IRC
13:22:18 *** cstim has joined #gnucash
13:22:19 *** gncbot sets mode: +o cstim
13:28:48 <warlord> cstim: nice job of libofx
13:34:02 <prock> jsled: have you fixed 397108 yet (sx editor window resizes poorly)?
13:35:13 <jsled> prock: kinda. Now that the plugin page is a split pane, it's better ... but the logic for smartly setting up the position of the split itself could be improved.
13:35:29 <jsled> I guess it's fine to close out that bug, at this poitn.
13:45:48 <jsled> Though I still don't have a good solution for the initial size issue.
13:46:36 <jsled> I guess min(height-of-sx-list, plugin-page.allocated.height * .5) is about right.
13:46:54 <jsled> Though the dense-cal has a minimum height as well...
13:48:46 <prock> I also have the problem with "339003 - sx editor too large for screen". 1280x800 resolution running FVWM WM. You mentioned in the ticket you want a screenshot. Do you want me to upload one?
13:49:10 <jsled> prock: sure.
13:53:20 *** franz has joined #gnucash
13:59:42 <cstim> warlord: I think it's not yet done. On my winXP box the build commands won't build OpenSP. Still working on it...
14:00:14 <warlord> Oh, okay.
14:03:40 *** andi5 has joined #gnucash
14:03:41 *** gncbot sets mode: +o andi5
14:04:38 <jsled> prock: the SX editor really just needs to be re-engineered to be more reasonable.
14:05:19 <jsled> Probably by supressing all the frequency selection (the GncFrequency control, maybe including the calendar) unless the user is actively editing it.
14:05:32 <jsled> Then the options can move into that space, and the height of the dialog can decrease.
14:06:24 <jsled> One could think about moving it all into individual tabs. Maybe {"Overview/Options", "Frequency", "Template Transaction"}
14:06:42 <jsled> (oh, phone)
14:08:45 <prock> that's a good idea.
14:09:13 <jsled> A lower prio task I have is to "model"ize the sx editor. That'd probably be a good time to re-do the ui.
14:09:21 <prock> it may be a good idea to have a frequency widget
14:09:39 <jsled> There is.
14:09:46 <prock> oh, well there you go.
14:10:09 <prock> what would be the sx modules?
14:10:21 <jsled> hmm?
14:10:46 <prock> oh. "model"ize, not "module"ize
14:59:46 *** esodan has quit IRC
15:00:21 *** nomeata has quit IRC
15:02:11 *** sjc has joined #gnucash
15:06:54 *** nomeata has joined #gnucash
15:19:52 <andi5> cstim: opensp fails for me too... in nsgmls, printing "can't be auto-imported" stuff.... do you need details?
15:20:03 <cstim> andi5: no, I have the same issue here
15:20:32 <andi5> cstim: i have finally found the source of my problems, see http://bugzilla.gnome.org/show_bug.cgi?id=404832 :)
15:20:58 <cstim> andi5: here's the thing: On my windows2000 box, nsgmls/ and the other program subdirectories built just fine. Here at my winXP home box, none of them builds. However, we only need the library and not the programs.
15:21:10 <cstim> andi5: cool. Congrats for tracking this one down!
15:21:39 <cstim> For opensp I'm inclined to simply run "make -i" and "make -i install" because that will give us the library and we don't need anything else.
15:22:03 <andi5> yes, please do so then :)
15:23:16 <cstim> really weird - on my win2000 box it just built fine, although I *should* have the same gcc and autotools (libtool in particular) installed
15:23:33 <cstim> I even tried libtool-1.5.23a instead of the normal 1.5.22, but this didn't change anything.
15:23:42 *** MrN has joined #gnucash
15:23:51 *** bash has left #gnucash
15:23:56 *** mnoir has joined #gnucash
15:24:03 <MrN> hi
15:24:37 *** bash has joined #gnucash
15:29:27 *** bash has left #gnucash
15:35:49 <cstim> andi5: committed.
15:44:00 <andi5> cstim: libosp-5.dll has landed :)
15:44:29 *** kris has joined #gnucash
15:44:33 <kris> cstim: around?
15:45:14 <andi5> he used to...
15:45:42 <warlord> kris: he was around 20 minutes ago.
15:46:04 <cstim> yes
15:47:24 <kris> and now he's back ;)
15:47:41 <kris> cstim: do you happen to have a reproducible testcase for that weird filter/sort model bug already?
15:48:22 <cstim> kris: no, unfortunately.
15:48:41 <andi5> does it need accessibility turned on, btw?
15:48:45 <cstim> http://bugzilla.gnome.org/show_bug.cgi?id=364946 you mean?
15:48:54 <kris> cstim: yes
15:49:16 <cstim> kris: I never observed one of those crashes myself. I only happen to be our current bugzilla manager...
15:49:24 <kris> oh, awesome, i can reproduce 398779 at least
15:50:06 <kris> but only, with accessibility enabled
15:50:13 <kris> andi5: I suspect so, yes ...
15:50:39 <cstim> kris: out of the few responses we got, some said accessibility is enabled, but at least one person said it isn't and it still crashes.
15:50:57 <kris> cstim: I saw, so I am still afraid there are different bugs
15:50:58 <andi5> maybe he did not logged out? ;-)
15:51:26 <cstim> kris: yeah, I figured that as well.
15:53:38 <cstim> andi5: I was referring to http://bugzilla.gnome.org/show_bug.cgi?id=364946#c39
15:56:33 <cstim> andi5: sh...ooot. I was trying to setup the aqbanking test account by aqhbci-tool on that winXP box, but aqhbci-tool always crashes with "assertion failed in banking.c"... :-(
15:57:27 <andi5> cstim: what did you execute?
15:58:34 *** twunder has quit IRC
16:00:14 <andi5> ouch.... http://codesnipers.com/?q=node/46
16:03:29 <cstim> ouch ouch
16:03:57 <cstim> andi5: oh, I discovered I had old DLLs around. Rebuilding...
16:25:27 *** cstim has quit IRC
16:28:13 *** esodan has joined #gnucash
16:30:45 *** kris has left #gnucash
16:35:30 <andi5> Q: what about qof_strftime? it will work like strftime, but return a utf8-string ...
16:36:05 <esodan> Hi all, just to tell you about my work on Qof->GObject: I'm deprecating most of the "virtual" functions in QofObject becouse they aren't used anyware, or using #define to call a new "way" to use it...
16:36:57 <jsled> esodan: you have access to the branch, eh? you should commit.
16:39:01 <jsled> esodan: what's the new way?
16:39:04 <esodan> jsled: I'll do it ASAP, but how can I tell you what I'm doing and what I'll do, becouse my current work isn't finised and I don't whant missunderstand about my work :-$
16:40:12 <jsled> esodan: committing code, (even if it doesn't work, (even if it doesn't compile,)) is going to be one of the clearest ways to communicate what you're doing.
16:40:15 <warlord> esodan: it's much easier to talk about actual code.
16:40:27 <warlord> (even if it doesn't work/compile)
16:40:43 <warlord> once you commit, we can cut-and-paste the changeset patch and comment on it directly.
16:41:00 <esodan> well no more words, I'll comit it... one question....
16:41:37 <andi5> yeah, these ain't words, it is a question ;-)
16:41:44 <esodan> I don't sync my work with the resent changes, does I need to sync before?
16:42:54 <jsled> esodan: you probably should, unless it's a crazy amount of effort. I doubt it is, though.
16:43:16 <jsled> I don't think there's been much change to the set of stuff I'd expect you to be changing.
16:45:42 <esodan> Sorry but I'm new with SVN could you tell me who to summit? :-D, don't worry I could find the way....;)
16:45:55 <jsled> but ... derek copied the branch from trunk@15459
16:46:20 <jsled> esodan: what sources have you been modifying? A checkout of SVN? A copy from a tarball?
16:46:47 <esodan> a checkout of SVN...
16:46:58 <jsled> what revision does `svn info` give you
16:46:58 <jsled> ?
16:47:15 <jsled> (also, what URL?)
16:47:16 *** _gunni_ has joined #gnucash
16:48:24 <esodan> Revision 15391
16:50:14 <jsled> and what url? <http://svn.gnucash.org/repo/gnucash/trunk>, I'd bet...?
16:50:31 <warlord> Hmm, it should be relatively easy to migrate to the new URL and update your WC.
16:51:03 <esodan> yes the trunk: <http://svn.gnucash.org/repo/gnucash/trunk>
16:51:10 <jsled> Right. So, there's two steps.
16:51:23 <jsled> svn update -r 15459 http://svn.gnucash.org/repo/gnucash/trunk
16:51:28 <jsled> That'll get you to the correct version.
16:51:34 <jsled> Then, backup you WC.
16:51:40 <jsled> Then,
16:52:00 <jsled> svn switch --relocate svn+ssh://svn.gnucash.org/repo/gnucash/trunk
16:52:05 <jsled> That might not work. I'm not sure.
16:52:23 *** twunder has joined #gnucash
16:52:29 <jsled> Then, svn switch svn+ssh://svn.gnucash.org/repo/gnucash/branches/gobject-engine-dev
16:52:39 <jsled> Then, you should be able to `svn commit`
16:53:02 <warlord> jsled: you dont want to switch to trunk.. you should switch to the branch.
16:53:08 <jsled> No, I don't think you do.
16:53:20 <jsled> I think you want to switch the protocol indepdnent from the trunk->branch switch.
16:53:22 <andi5> svn+ssh://esodan@svn.gnucash.org/repo/gnucash/branches/gobject-engine-dev , i guess
16:53:49 <jsled> At least, I'd do it that way so I know how bad it is if it breaks.
16:53:50 <jsled> :)
16:54:28 <jsled> In fact, I'm pretty sure you can't switch both protocol and directory at the same time.
16:54:29 <warlord> so, yeah... svn update -r 15459 http://svn.gnucash.org/repo/gnucash/trunk and then svn switch --relocate svn+ssh://esodan@svn.gnucash.org/repo/gnucash/branches/gobject-engine-dev
16:54:39 <warlord> Are you sure?
16:55:05 <jsled> Pretty sure, but not 100%.
16:55:54 *** |gunni| has quit IRC
16:59:38 <esodan> Please help me then?
16:59:41 <jsled> esodan: but the general idea is that you need to do 3 things: a) get updated to the point the code was branched (r15459), b) change from 'http' to 'svn+ssh', and c) move from trunk to the branch.
17:00:19 <jsled> there's another approach, but it'll be more tedious work, and probably should only be done if the above doesn't work.
17:01:22 <jsled> in any case, make a backup of what you have before going ahead, just in case.
17:02:21 <esodan> Ok step a) update: I'll use a simple svn update ./gobject-engine-dev
17:03:15 <jsled> esodan: you might want to merge [15459:HEAD] separately.
17:04:02 <warlord> esodan: you probably DO NOT want to just update to HEAD!
17:04:47 <jsled> either way. esodan might want to pull in the [15459:HEAD] changes, but it just compilcates things to do it right now. *shrug*
17:04:47 <esodan> Ok, then svn update -r 15459 ./gobject-engine-dev
17:04:47 <warlord> esodan: you should do what was suggested above. Update to r15459, then migrate to your branch. And then you can (later) worry about merging the trunk changes.
17:05:18 <jsled> Certainly there's already enough moving parts here to not need to bring in another.
17:06:49 <warlord> Yeah
17:08:15 <esodan> Can I chekout my branch, patch it localy with my changes and then commit?
17:11:23 <jsled> esodan: yeah, that's "the other way", except you'll still want to do the [15391:15459] update.
17:11:42 <jsled> And it's tediuos because you'll need to manually deal with file {add,del}s.
17:12:14 <jsled> but the patch can be done pretty easily ...: (cd existing; svn diff ) | (cd new ; patch -p0 -)
17:12:31 <jsled> s/want to do/need to do/
17:12:42 <jsled> (since the branch includes those changes)
17:14:38 <warlord> he wont have to worry about [15391:15459] if he does "svn diff | (cd .../new-checkout; patch -p0)"
17:15:10 <jsled> assuming it patches in cleanly.
17:15:32 <jsled> which it probably will.
17:16:07 <andi5> and do not forget to svn remove/add all necessary files :) ... i guess nobody has always done that ;-)
17:17:36 *** andi5 has quit IRC
17:23:41 *** bonez39 has joined #gnucash
17:26:31 <esodan> I have done a svn diff | patch -p0, and reject most of the changes in the files I have worked...
17:26:56 <jsled> "and reject"? You did, or it did/
17:26:56 <jsled> ?
17:28:01 <esodan> well when the script patch ask for 'Reversed (or previously applied) patch detected! Assume -R? [n]' I have told NO
17:28:41 <esodan> when ask Aply anyway? again NO, and I have a *.rej files...
17:28:49 <warlord> esodan: did you run the "svn diff" and "patch" in the SAME directory???
17:29:07 <warlord> You need to run the "diff" in the existing checkout and the "patch" in the "NEW" checkout.
17:29:12 <esodan> Yes :-X, don't worry I have a copy...
17:29:25 <esodan> Ok, I'll do it again...
17:30:28 *** twunder has quit IRC
17:30:31 <jsled> The command I gave is literal ... the ()s create two sub-shells, one in the existing/modified directory, and one in the new dir.
17:31:27 <jsled> Another way to do it is: cd existing; svn diff > /tmp/changes.patch; cd ../new; patch -p0 < /tmp/changes.patch
17:40:34 <warlord> esodan: have you read http://wiki.gnucash.org/wiki/Subversion ?
18:12:16 *** bash has joined #gnucash
18:29:37 *** _gunni_ has quit IRC
18:30:15 *** bash has left #gnucash
18:32:08 *** MrN has quit IRC
18:43:35 *** warlord is now known as warlord-afk
18:57:03 *** nomeata has quit IRC
18:58:26 *** AlonzoTG has joined #gnucash
18:58:43 <AlonzoTG> How do I remove a bogus line from a transaction?
19:05:22 <prock> I see references in the code to being able to have multiple books, is this truely the case?
19:12:29 *** franz has quit IRC
19:26:38 *** Esaj has quit IRC
19:29:13 <jsled> AlonzoTG: just delete all the parts, then tab past the end; the Split should disappear.
20:10:49 <jsled> prock: kinda. I forget how much it's supported. Most of the app only cares about the "current" book. The XML file does not have the book as the root element, so it could contain multiple books.
20:11:33 <jsled> But IIRC a book is the "root" of a qof object graph. Which isn't to say that that book couldn't itself contain the historical books.
20:17:50 <jsled> [ot, via /.] http://blogs.zdnet.com/threatchaos/?p=311
20:20:09 *** ceplma has quit IRC
20:30:29 *** sjc has quit IRC
20:36:47 *** rlaager has joined #gnucash
20:39:29 *** ceplma has joined #gnucash
20:44:16 *** esodan has quit IRC
20:54:39 *** ceplma has quit IRC
21:05:42 <prock> jsled: I'm more thinking about some sort of "branch" of a book. If it were possible to make a branch of a book it would then be possible to apply SX's to it.
21:07:10 <jsled> Ah. Well, if you want to go down that path, you can probably just create a new session, copy the existing book into it, apply the SXes, take the results you want, and toss it.
21:07:26 <jsled> Of course, if the book is large, that copying step might be a bitch.
21:07:46 <jsled> And memory intensive, depending.
21:33:43 *** twunder has joined #gnucash
21:51:39 *** mnoir has quit IRC
22:01:39 <AlonzoTG> !!! All ebuilds that could satisfy ">=dev-scheme/guile-1.8" have been masked.
22:01:44 <AlonzoTG> (dependency required by "app-office/gnucash-2.0.4-r1" [ebuild])
22:01:57 <AlonzoTG> (and gnucash is working right now...)
22:04:04 *** digdug has left #gnucash
22:04:13 <prock> jsled: where'd you get the link to that zdnet windows security blog?
22:11:51 <jsled> prock: /.
22:12:00 <prock> ah
22:12:25 <jsled> AlonzoTG: what's req'ing guile-1.8?
22:13:35 <jsled> Ah, as you say, gnucash-2.0.4-r1.
22:14:19 <jsled> Yeah. It seems like they want to move to guile-1.8, which should work fine with gnucash.
22:14:25 <jsled> (and maybe better with slib, I don't know)
22:17:45 <AlonzoTG> -- I just got finished downgrading to 1.6 in order to run gnucash =|
22:17:51 <jsled> why?
22:17:53 <AlonzoTG> it's running now.
22:18:03 <AlonzoTG> cuz I had to see how broke I am.
22:18:19 <AlonzoTG> (Actually I got a job which is worth (salary + benefits) around 60k)
22:18:20 <jsled> No, why did you need to downgrade?
22:18:25 <AlonzoTG> so I shouldn't have much to complain about soon.
22:18:31 <AlonzoTG> because it wouldn't run.
22:18:34 <jsled> Why not?
22:18:37 <AlonzoTG> it wouldn't even compile.
22:18:41 <jsled> why not?
22:18:49 <AlonzoTG> some BS about some access violation.
22:19:00 <AlonzoTG> I'll let it upgrade and report...
22:19:53 <jsled> (`quickpkg` is probably advisable, if you're not already using it)
22:20:34 <AlonzoTG> I have a revdep rebuild running which will take up the rest of the evening.
22:20:43 <AlonzoTG> I'll work on gnucash some more tomorrow.
22:28:19 *** twunder has quit IRC
22:56:16 *** conrad has quit IRC
22:57:51 *** rlaager has quit IRC
23:01:36 *** adrigen has joined #gnucash
23:02:25 *** AlonzoTG has quit IRC
23:04:16 <adrigen> Anyone: I am setting up some accounds for my small business (in australia w/GST), just wondering if there is anything I need to do besides checking the "bill tax included" check box, to equate for gst in my books. I havent made any specifiaction of what the GST rate is anywhere...
23:09:05 *** jpeach has joined #gnucash
23:09:26 <adrigen> sorry... found the FAQ. Ill go through that
23:09:36 *** jpeach has left #gnucash
23:10:19 *** adrigen has quit IRC
23:30:08 *** wizkid238 has quit IRC
23:33:35 *** hampton has left #gnucash
23:36:03 *** wizkid238 has joined #gnucash