2007-02-09 GnuCash IRC logs
00:00:44 *** minDscrm has quit IRC
00:42:12 <jsled> Hee... [[[
00:42:12 <jsled> * 2007-02-09T00:41:36 MESSG <gnc.engine> failed to load gnc-backend-postgres from /opt/gnc/unstable/lib/gnucash
00:42:13 <jsled> * 2007-02-09T00:41:37 INFO <gnc.gui> [gnc_html_graph_gog_init] init gog graphing
00:42:13 <jsled> * 2007-02-09T00:41:55 MESSG <gnc.bin> loading system configuration
00:42:13 <jsled> ]]]
01:38:59 *** dbr has joined #gnucash
01:58:01 *** dbr has quit IRC
02:21:06 *** sishen has joined #gnucash
02:52:38 *** Esaj has joined #gnucash
03:05:40 *** minDscrm has joined #gnucash
03:06:58 *** Esaj has quit IRC
03:51:28 *** RallyU has joined #gnucash
03:54:09 *** ErKa has joined #gnucash
04:26:06 *** |gunni| has joined #gnucash
04:47:28 *** cstim has joined #gnucash
04:47:29 *** gncbot sets mode: +o cstim
04:47:37 <cstim> Does anyone want to tell me something?
04:47:37 <gncbot> cstim: Sent 11 hours and 28 minutes ago: <andi5> WRT aqbanking/gwenhywfars stdcall/cdecl issues...i have built both from a clean svn checkout now and _everything_ seems to work: make check in doc/inherit, testlib, cancelling the aqbanking setup druid.... i suppose i must have something old/weird lying around, but i just do not know what..... will start a _complete_ recompile tomorrow
04:47:51 <cstim> @tell gncbot Thanks a lot for this info.
04:47:51 <gncbot> cstim: Error: I can't send notes to myself.
04:48:54 <cstim> @tell andi5 Ok, thanks. In install.sh I think I'll add a "make check" to gwenhywfar because it's rather cheap.
04:48:54 <gncbot> cstim: The operation succeeded.
04:49:34 <cstim> @tell andi5 re indent: This was added because it was being called by g-wrap. No more g-wrap, no more indent needed.
04:49:34 <gncbot> cstim: The operation succeeded.
04:54:13 <cstim> :-( http://svn.gnome.org/viewcvs/ seems to be down.
04:56:02 <cstim> :-(( not even http://svn.gnome.org/svn/glib/trunk seems to work.
05:00:13 *** RallyU has quit IRC
05:24:43 * cstim buys some safe_strcmp.
05:24:52 <cstim> http://bugzilla.gnome.org/show_bug.cgi?id=405851
05:38:07 *** ceplma has joined #gnucash
06:03:15 *** ceplma has quit IRC
06:03:44 *** ErKa has quit IRC
06:19:16 *** sishen has quit IRC
06:58:40 *** ceplma has joined #gnucash
07:03:12 *** bash has joined #gnucash
07:25:33 *** twunder has joined #gnucash
07:33:00 *** bash has left #gnucash
07:34:38 *** twunder has quit IRC
07:37:06 *** spel has joined #gnucash
07:43:51 <spel> Hello. I ran a spell cheching using KBabel on fr.po, and for all msgid or msgstr starting with an empty line the first line was truncated. Is that OK?
07:44:27 <cstim> spel: no, unfortunately.
07:46:15 <spel> Those lines start like msgid "" " Should I change the first two " for some escape character like \n ?
07:51:24 <cstim> oh, you mean those.
07:51:48 <cstim> *those* starting lines have no extra meaning.
07:51:57 <cstim> msgid ""
07:52:00 <cstim> "foo bar"
07:52:07 <cstim> those two lines above are identical to
07:52:11 <cstim> msgid "foo bar"
07:52:20 <cstim> so that can indeed be ignored and it is OK.
07:52:47 <spel> Yes quite a few start with msgid or msgstr followed by three quoting marks
07:53:12 <cstim> only if you have msgid "\nfoo bar" then this must not be changed to msgid "foo bar".
07:53:15 <spel> thanks
07:53:44 <cstim> If you run "msgmerge" after this change from KBabel, then msgmerge will probably change it back.
07:53:58 <spel> I see
07:53:58 <cstim> do you know how to use msgmerge?
07:55:17 <spel> I read about. (learning process) and used it twice.
07:56:01 <cstim> ok. If you run it, it will change the msgid back to start with msgid "" lines. That is probably better because it makes it easier to see which ones are actually changed.
07:56:55 <spel> So I should send Fabrice a .diff file including those changes ?
07:57:31 <cstim> don't send a diff. Just send the file itself. But in any case it is good if "diff" can be used to see the differences in a meaningful way.
07:59:09 <spel> I was to send .diff only because this is how Fabrice and I had agreed to do, but I don't have a problem with the whole file.
07:59:56 <cstim> he agreed on that? well, then do what you agreed upon. I generally recommend to send po files itself (or maybe in addition to a diff), because diffs with po files are a bit tricky.
08:01:11 <spel> I am learning. I will send him both and make a note about your comment.
08:01:27 <cstim> ok, good.
08:03:38 <spel> May be he thought we exchange .diff files to agree on changes, and I would send you the complete file? I will ask.
08:03:47 <spel> We will talk some time later.
08:12:12 <cstim> sure, no problem.
08:23:00 *** twunder has joined #gnucash
08:25:33 * cstim just ran a aqbanking/HBCI statement download on win32 successfully.
08:25:56 <cstim> http://www.tu-harburg.de/~et2cs/gnc/gnc-screenshot13.png
08:30:07 <jsled> awesome.
08:34:08 *** hampton2 has left #gnucash
09:05:03 <jsled> Hmm. Is there a good, common, existing macro for (str == NULL ? "(null)" : str) ?
09:06:03 <cstim> not that I know of. I'd have written (str ? str : "null")
09:06:50 <cstim> and in the obfuscated C contest you could write (str && "null")
09:07:03 <cstim> err, (str || "null")
09:07:21 <cstim> (obfuscation hit again)
09:08:36 <jsled> :)
09:08:41 *** spel has left #gnucash
09:09:03 <cstim> @seen benoitg
09:09:03 <gncbot> cstim: benoitg was last seen in #gnucash 2 weeks, 2 days, 5 hours, 27 minutes, and 28 seconds ago: <benoitg> Ok, libofx will now "do the right thing" as far as gnucash is concerned.
09:09:22 *** spel has joined #gnucash
09:09:51 <cstim> @tell benoitg Undelivered Mail Returned to Sender: <benoitg@coeus.ca>: cannot update mailbox /var/mail/benoitg for user benoitg. cannot open file: File too large
09:09:51 <gncbot> cstim: The operation succeeded.
09:11:36 *** Demitar__ has quit IRC
09:12:26 <jsled> hmm. #define IS_NULL ... #define MAYBE_NULL ... #define STR_IS_NULL ... ?
09:19:39 *** spel has quit IRC
09:20:58 <cstim> as macro name?
09:21:20 <cstim> you'd need a name that implies you get a string as return value.
09:21:35 <cstim> STR_OR_NULL, I'd say.
09:22:25 <jsled> I'll just inline it. There's fewer places we're doing this than I thought.
09:23:38 *** Demitar has quit IRC
09:36:22 *** pisi has joined #gnucash
09:37:00 <pisi> Hi. Reading mailinglist posts from 2004 I understand that gnucash wants to move from xml to sqlite ?
09:37:40 <jsled> pisi: generally.
09:38:01 <pisi> Although i do understand what benefits it would bring - I'd also like to point out one drawback - currently I can keep the uncompressed datafile in subversion.
09:38:32 <jsled> pisi: that's true.
09:39:37 <jsled> pisi: why are you doing that?
09:41:14 <pisi> all company docs are in SVN. this includes a workflow of bills going in and out. and I know plenty of softwares that happen to fsck things up a bit
09:41:15 <jsled> (as an aside, note that the XML serializer doesn't guarantee a stable ordering to the XML, so there might be spurious diffs)
09:41:50 <pisi> there are - but still textual files seem good. Well, as the size gorws it becoes a bottleneck, sure..
09:42:18 <pisi> and also subversion is used as a medium for sharing files. this includes bookkeeping.
09:42:59 <jsled> Well, treating the datafiles as an opaque structure is right. So, versioned XML vs. versioned sqlite is a wash.
09:45:07 *** spel has joined #gnucash
09:50:21 *** pisi has quit IRC
09:54:14 *** ErKa has joined #gnucash
09:56:12 *** warlord-afk is now known as warlord
10:14:56 *** motin_ has quit IRC
10:26:08 *** andi5 has joined #gnucash
10:26:08 *** gncbot sets mode: +o andi5
10:26:31 <andi5> cstim: congratulations :)
10:30:07 <cstim> andi5: thanks :-)
10:30:32 <cstim> however, only balance and statement retrieval works. Online transfer doesn't work, and as usual I have no idea why.
10:30:46 <andi5> hehe
10:31:12 <cstim> did you check all GWENHYWFAR_CB possibilities? __stdcall, __cdecl, nothing?
10:31:49 <andi5> i am just about to clean my whole environment...
10:32:07 <warlord> I love windows!
10:32:09 <cstim> oh, I've uploaded a new gwenhywfar just this moment.
10:32:18 <andi5> warlord: ??
10:33:05 <warlord> (i'm joking)
10:33:10 <andi5> *sigh*
10:36:04 <cstim> audits, anyone...? http://bugzilla.gnome.org/show_bug.cgi?id=405851
10:39:31 <cstim> heh. assert_one_dir is quite helpful in avoiding version mixup in install.sh.
10:40:33 <andi5> cstim: do the registry keys mean that aqbanking will not work from msys or c:\soft\gnucash\inst?
10:41:28 <cstim> andi5: err... I think it *does* work if the compile-time paths are correct at runtime.
10:41:46 <andi5> great
10:41:47 <cstim> because that's what's being used on Unix everytime.
10:43:19 *** wizkid238 has quit IRC
10:49:26 <warlord> cstim: I'll audit this weekend.
10:49:53 <warlord> cstim: I wonder if we could use assert_one_dir to effect the 'reset.sh' functionality?
10:50:04 *** wizkid238 has joined #gnucash
10:50:39 <andi5> assert_zero_dir? :)
10:50:43 <andi5> s
10:52:12 <warlord> Heheh
10:52:35 *** Jocke has joined #gnucash
10:52:36 <warlord> No, I was more thinking that if we changed the version of a package we could remove the 'installed' version.
10:53:35 <andi5> i am not sure whether i would want install.sh to automagically recompile stuff, unless we definitely need that shiny new version...
10:54:41 <andi5> suppose i install from another directory which is an svn checkout... would that mean my installation would not survive install.sh?
10:54:42 * cstim agreed with andi5
10:54:57 <cstim> s/agreed/agrees/
10:57:11 *** keith has joined #gnucash
10:57:35 <keith> Hi guys, gnucash is that like a opensource version of sage?
10:57:38 * warlord shrugs.
10:57:55 <warlord> I always just use install.sh, and I like to keep up with the joneses.
10:58:05 <warlord> keith: sage? never heard of sage/
10:58:53 *** ErKa has quit IRC
10:59:38 <keith> warlord: its an account package.
10:59:50 <keith> i guess like quick books or w.e m$ call it
11:00:27 <warlord> keith: well, it's more like quicken/quickbooks. Not sure about sage. But I've never used any of those..
11:01:51 <keith> well we run a small IT company IT Support contracts / broadband / webhosting would gnucash work for me ?
11:02:10 <cstim> keith: on windows? It is still experimental on windows.
11:03:01 <cstim> keith: or on which operating system?
11:03:48 <warlord> well... gnucash doesn't have repeating invoices.. So if you have LOTS of invoices it might be a pain to enter them all int.
11:22:28 *** dbr has joined #gnucash
11:23:22 <dbr> cstim: line 734 of install.sh is still hardcoded for the default Qt3win install location (libdir)
11:27:07 <keith> warlord: ah we need repeating invoices tbh
11:27:10 * jsled has a pretty large commit coming soon. Hopefully people will be able to give it a run when it lands...
11:27:22 * andi5 ducks
11:27:35 <keith> cstim: I run an ubuntu desktop.
11:27:44 <keith> a windows desktop and linux servers.
11:28:20 <cstim> dbr: oh, indeed I missed that one. Thanks.
11:29:13 *** Jocke has quit IRC
11:29:49 <cstim> dbr: r15535
11:29:58 <warlord> keith: there's a feature request, but it hasn't been implemented. I think even a "copy invoice" would be useful.
11:29:59 <andi5> cstim: do you know a way to let msys treat backslashes as .... backslashes?
11:30:37 <cstim> instead of what?
11:30:48 <warlord> andi5: double them: \\
11:30:48 <andi5> instead of converting them to slashes
11:31:02 <andi5> warlord: does not help, i get two slashes then
11:31:03 <cstim> it's always shell quoting. So no way around \\
11:31:14 <cstim> huh?
11:31:40 <andi5> i am inside ' quotes, not ".... otherwise i have to use \\\\ for //, of course
11:31:45 <cstim> andi5: can you give an example (with context)?
11:32:20 <andi5> /c/soft/downloads/msysDTK-1.0.1.exe //SP- //DIR=f:\\\\gabb
11:32:57 <andi5> add //SILENT once it works :)
11:34:05 <cstim> $ echo 'c:\foo'
11:34:06 <cstim> c:\foo
11:35:16 <andi5> i guess the shell internally does not change slashes and backslashes
11:36:24 <warlord> Sounds right
11:39:29 <dbr> cstim: thanks for the extra branch in install.sh. It looks like I'm finally past errors in my aqbanking build. I did have to set both QTDIR and the qt/bin path manually. Maybe if you don't use the default Qt4 install it doesn't set them.
11:40:51 *** Demitar has joined #gnucash
11:43:49 <cstim> dbr: how far does it build currently?
11:46:22 <jsled> Are you all trying to resolve a problem on trunk currently, or can I land a large change? While it touches a lot of files, it shouldn't be too disruptive...
11:46:39 <dbr> cstim: looks like I'm in the middle of making qbanking
11:47:21 *** ceplma has quit IRC
11:47:26 <dbr> jsled: mine's just a packaging/win32 thing
11:48:51 <cstim> jsled: I'm not working on trunk ATM.
11:50:31 <warlord> jsled: thunk it in!
11:51:34 * jsled does a distclean rebuild just to be sure... just a bit more.
11:53:01 <dbr> cstim: oops. Somehow I still have a default qt path in there. I'll clean up, update, and try again tonight.
11:53:41 <cstim> what do you mean by "somehow"?
11:54:41 <dbr> I edited the hard coded libdir to match my qt install in install.sh, and ran install.sh
11:55:08 <dbr> but I still get an error that /c/Qt/4.4.2/blah doesn't exist
11:58:37 <dbr> hmm. looks like not doing anything to clean out the prior aqbanking configure is what interfered this time.
11:59:26 <cstim> see ya
11:59:29 *** cstim is now known as cstim_away
12:01:49 <dbr> gotta go. thanks for the help
12:01:52 *** dbr has quit IRC
12:07:18 <andi5> hasabbll.... why the heck does inno complain about slashes if it can handle them perfectly once run //SILENTly.... *grrr*
12:09:03 <jsled> Oh, pfft. Of course I should have `svn update`d before doing the last full rebuild. Oh well, need to go to the store anyways...
12:10:15 <warlord> Hehehe
12:11:17 <andi5> yeah... goodbye user interaction
12:17:13 <warlord> That's probably okay.
12:28:39 *** MrN has joined #gnucash
13:25:12 <jsled> warlord: to verify, you're saying with those two log-levels, that g_info("gnc.foo.bar") should be logged?
13:26:03 <warlord> Let's say that I say "gnc" -> debug, then yes, g_info(gnc.foo.bar, ...) should be logged.
13:26:42 <warlord> (even if your config says gnc.foo.bar -> warn)
13:27:15 <warlord> because, IMHO, I should be able to log as much as possible when I want to. "gee, I want to see ALL the debug messages from 'gnc'"
13:27:21 <jsled> No, it doesn't behave that way presently even.
13:27:36 <jsled> I don't believe it should.
13:28:02 <jsled> At least, log4j doens't behave that way, and we actively make good use of that fact.
13:28:27 <warlord> where do we make use of that fact?
13:28:37 <jsled> Sorry. "we" = "me (at work)"
13:29:18 <warlord> Oh, okay, so how do you at work actively make use of the fact that it's not really a true hierarchy?
13:29:24 <jsled> The example is running the overall system at info, but quieting down a noisy/stable sub-module by raising it to warn.
13:30:28 <warlord> Hmm.. I guess by default it inherits from the parent but if you specify it explicitly then it doesnt.
13:30:45 <jsled> Yes.
13:30:56 <warlord> I guess that makes sense..
13:36:16 *** rlaager has joined #gnucash
13:40:13 *** ErKa has joined #gnucash
13:51:44 *** Esaj has joined #gnucash
14:02:08 *** ErKa has quit IRC
14:43:43 *** jpeac1 has joined #gnucash
14:44:41 *** sjc has joined #gnucash
14:50:10 *** jpeac1 has left #gnucash
15:41:01 *** ErKa has joined #gnucash
16:01:22 <jsled> christ enter/leave isn't a good idea without structured exception handling.
16:01:46 <warlord> heh!
16:02:14 <jsled> I'd started to go through doing penance for break it, by trying to balance out enter/leave pairs.
16:02:35 <jsled> After a solid hour, I'm 36% way through the find. :/
16:02:48 <warlord> :(
16:03:05 <jsled> And for the first 20%, I didn't really deal with g_return_{,val_}if_fail(...) in a good way.
16:03:24 <jsled> We could have a gnc_leave_return{,_val}_if_fail ... more 'tilling.
16:03:30 <warlord> Ahh, for the love of C++ c'tor/d'tor
16:04:07 <jsled> I guess that could be used, too. But who wants to work in c++? :/
16:04:20 <jsled> I guess the existing codebase, heh.
16:04:26 * warlord raises his hand
16:07:29 <warlord> I think a LOT of things we do in GnuCash would be MUCH easier if we used C++
16:15:49 <ErKa> time to go with longjump()... :-)
16:16:45 <warlord> :-P
16:18:29 <ErKa> Just hide them in some try/catch/throw macros and it rocks :-)
16:19:03 <warlord> Um...
16:48:58 *** _gunni_ has joined #gnucash
16:57:39 *** |gunni| has quit IRC
16:59:30 <andi5> jsled: gnucash crashes on windows... shall i first rebuild or try to nail it down?
16:59:57 <andi5> (crashing in some strlen... do we printf some NULL somewhere? :))
17:00:54 <jsled> we might, though we shouldn't. I'd love a stack trace, of course.
17:01:11 <jsled> (though I need to run an ~20 minute errand very shortly...)
17:01:53 <andi5> oh, no problem
17:06:50 *** twunder has quit IRC
17:17:28 <andi5> completely undebuggable so far :( ... jsled, what have you done ;-)
17:18:26 <andi5> frame 0-20: http://pastebin.ca/raw/348087
17:24:43 <warlord> looks like the stack's been scribbled.
17:25:41 <andi5> in a nutshell, how do i see debug (enter/leave) messages in r15538? ... is `gnucash --logto c:\whatever` supposed to work?
17:27:53 <warlord> andi5: try a filename in CWD.
17:28:35 <andi5> well... gnucash indeed opens a file like c:\whatever.3BXHNT, but it is empty
17:32:53 <warlord> Hmm
17:33:20 <andi5> well, will get it anyway...
17:33:40 <jsled> `gnucash --logto stderr --log gnc=debug`
17:33:57 <jsled> (or to a file, as you see)
17:34:03 <andi5> jsled: there is no stderr on windows
17:34:31 <jsled> 'stdout', then? What does stdlib (stdio?) define the symbol 'stderr' as?
17:34:38 <andi5> ok, so i will set --log gnc=debug... do i need --debug?
17:34:44 <jsled> no
17:35:14 <andi5> jsled: the point is that we set the subsystem to GUI... stdout and stderr have to be considered closed (or something like that)
17:35:36 <jsled> Ah, I recall that. noted.
17:36:34 <jsled> So, there's two places that call strlen: the "qof_log_check", as it steps through the log-domain string; it explicitly handles a NULL log-domain.
17:36:44 <jsled> And qof_log_prettify(), which doesn't.
17:37:12 <jsled> er, which does. :/
17:38:10 <jsled> you could also add "--log qof=debug", depending on what detail you're after.
17:39:36 <andi5> so i can use --log qof=debug --log gnc=debug ?
17:39:54 <jsled> yup.
17:41:26 <jsled> qof_log_check does some not-awesome string munging ... (gchar*)dot_pointer++ and {*dot_pointer = '.'} That might contribute.
17:47:55 <andi5> jsled: so i really have log files now :) .... meanwhile i am isolating the bug
17:57:03 <andi5> http://pastebin.ca/raw/348116 ... the crash happens inside gtk_action_activate.... i guess i will have to recompile gtk :(
17:58:13 *** prock_ has joined #gnucash
18:02:21 <jsled> andi5: might be interesting to comment out the DEBUG in gnc_main_window_restore_default_state and see if the problem moves.
18:02:56 <andi5> nice, my winxp explorer just crashed heavily
18:03:23 <warlord> Nice!
18:03:26 <andi5> well, it passes the DEBUG, do you think there is a difference?
18:04:31 <jsled> Ah, I see.
18:07:11 *** prock has quit IRC
18:07:29 *** spel has quit IRC
18:07:36 <andi5> i have found my swiss army knife, self-made pre-compiled gnome packages with debugging symbols :)
18:08:37 <andi5> oh... they even make nice stacktraces :)
18:09:04 <warlord> nice
18:09:26 <andi5> http://pastebin.ca/raw/348129
18:10:44 <andi5> yeah, that cannot work
18:10:47 <jsled> Oh, that's pretty cool.
18:10:59 <andi5> it is beautiful, is not it? :)
18:11:27 *** ErKa has quit IRC
18:13:50 <warlord> nice stacktrace
18:18:23 <andi5> hm.... newbie question: what is wrong about "LEAVE("account %s(%p)", xaccAccountGetName(account) || "(null)", account);" ? ... it complains about "format argument is not a pointer (arg 5)" ... hmm
18:20:15 <warlord> the compiler is probably being confused about the "||" construct.
18:20:30 <warlord> Does it work if you use the ( ? : ) construct?
18:20:56 <warlord> Besides, unless account is NULL xaccAccountGetName() will always return non-NULL
18:21:16 <andi5> yes, it works
18:23:23 <andi5> gnucash also starts up just fine
18:25:40 <andi5> LEAVE("account %s(%p)", (gchar*) (xaccAccountGetName(account) || "(null)"), account); works too... may i use that?
18:25:44 *** warlord is now known as warlord-afk
18:26:17 <jsled> I think the ( str != null ? : ) is more clear than the "||" and "&&" forms.
18:27:20 <andi5> jsled: is xaccAccountGetName called anyway or only if the message is to be logged?
18:28:00 <jsled> called anyways, I guess.
18:31:39 <andi5> well, i will introduce a new variable then...
18:48:19 *** prock_ is now known as prock
18:51:30 *** _gunni_ has quit IRC
18:59:43 *** andi5 has quit IRC
19:10:10 *** twunder has joined #gnucash
19:13:21 *** twunder has quit IRC
19:17:10 *** twunder has joined #gnucash
19:21:09 *** IanL has joined #gnucash
19:22:24 <jsled> Strangest gnucash reference I've seen, [off] to prevent leaking pagerank:
19:23:16 <jsled> Hmm. We should have a way to indicate a url that should be linked with @rel="nofollow"...
19:44:13 *** MrN has quit IRC
19:45:20 <IanL> morning/evening all,
19:46:34 <jsled> hello
19:46:45 <IanL> I'd like to try to hack around with the gnucash source but I already have gnucash (to manage my personal finances) installed.
19:47:17 <IanL> I remember there being problems with having multiple gnucash versions on one machine,
19:47:28 <IanL> from the mailing list,
19:48:07 <jsled> If you run one in /usr and one in /usr/local, that can be problematic. Devs generally run the dev/svn copy --prefix'ed into /opt/gnc/unstable (or whatever). It's easy to identify, wipe away, and runs cleanly separated.
19:48:22 <jsled> See http://wiki.gnucash.org/wiki/Building
19:48:23 <IanL> So I was wondering the best way to be able to build and test the SVN version of gnucash while still being able to the released version to manage my finances.
19:48:40 <IanL> ok,
19:48:55 <IanL> thanks,
19:49:00 <jsled> Another place to --prefix it is $HOME/installed/gnucash or whatever.
19:49:54 <jsled> If you go the /opt/ route, you might want to have a container directory (/opt/gnc) and an installed directory under that (/opt/gnc/unstable) ... you can chown /opt/gnc to your user account, and never need to su(do) ...
19:50:21 <jsled> Even if you want to blow away the whole dir, you can still rm -Rf /opt/gnc/unstable, and leave /opt/gnc chown'ed as yourself, &c.
19:50:44 <jsled> Plus, if you want to run multiple versions, you can make /opt/gnc/svn vs. /opt/gnc/gda-dev-branch or whatever.
19:50:48 <IanL> hmm, I prefixed it to a directory under my home ($HOME/src/gnucash/unstable) that way and I got an error trying to make install it.
19:51:03 <IanL> good point,
19:51:05 <jsled> Is that the same directory you're building in?
19:51:31 * jsled should add all that to the wiki... we reiterate it infrequently, but frequently enough..
19:51:39 <IanL> Well I have the source in $HOME/src/gnucash/gnucash
19:51:49 <IanL> so it's not the same,
19:52:11 <jsled> Ah, interesting. Have the error?
19:52:29 <IanL> libtool: install: error: cannot install `libgsf-gnome-1.la' to a directory not ending in /usr/local/lib/gnucash
19:52:40 <IanL> though I prefixed it to the directory under my home,
19:53:12 <IanL> I ran configure earlier without a prefix but I realized the error and ran it again with the prefix,
19:53:13 <jsled> Hmm. What commands did you use?
19:53:19 <IanL> perhaps that's the problem,
19:53:37 <IanL> ./configure --prefix
19:53:44 <IanL> oops
19:53:50 <jsled> Hmm; re-running $(./configure --prefix $whatever) will work fine...
19:54:00 <IanL> ./configure --prefix=$HOME/src/gnucash/unstable
19:54:05 <IanL> make
19:54:07 <IanL> make install
19:54:07 <jsled> Ah. Try without the '='
19:54:15 <IanL> really?
19:54:21 <jsled> I think so, yeah. :/
19:54:48 <jsled> Which would be bad, since that's the instruction on that Building page. :/
19:55:02 <IanL> hmm, always worked before with ./configure scripts for me,
19:55:07 <IanL> I'll give it a try though,
19:55:19 <jsled> Hmm. I never provide it, but if you're confident it works, I am.
19:55:57 <IanL> yah, I've never done it without,
19:55:59 <jsled> So, in the $top_srcdir/Makefile, what's "prefix"? (around line 430, here)
19:56:19 <IanL> well shit, I just overwrote it,
19:56:24 <jsled> :)
19:56:50 <IanL> whoops. I should mind the children ;)
19:57:05 <IanL> since this channel is publicly logged after all
19:57:17 <jsled> ah. heh heh.
19:57:37 <IanL> Anywho, I'll know in a sec if it worked this time,
19:59:00 <IanL> nah same problem,
20:00:47 <jsled> Hmm. So, what's prefix, then?
20:01:12 <jsled> And, maybe, can you pastebin a bit of context around the error?
20:01:20 *** conrad has quit IRC
20:01:25 <IanL> ok, gimme a sec
20:02:47 <IanL> $HOME/src/gnucash/unstable
20:03:01 <IanL> well it used the absolute path not $HOME
20:06:41 <IanL> pastebin is being wierd,
20:09:08 <IanL> http://paste.lisp.org/display/36649
20:10:58 <jsled> IanL: What distro?
20:11:15 <jsled> What verison of libtool?
20:13:49 <IanL> lets see,
20:14:15 <IanL> I'm using Ubuntu 6.10 "Edgy"
20:15:08 <IanL> libtool is installed from a package
20:15:10 <IanL> 1.5.22-4 is the package version
20:15:31 <jsled> sounds right. Isn't there a goffice package you can install?
20:16:03 <IanL> if there is one it's not called goffice,
20:16:09 <jsled> Or, more appropriately libgsf ...
20:16:16 <jsled> libgoffice, maybe?
20:16:19 <IanL> yah,
20:16:22 <jsled> (and still libgsf. :)
20:16:23 <IanL> ok, libgsf
20:16:38 <IanL> and libgoffice both exist
20:16:44 <jsled> installed already?
20:17:01 <IanL> nope,
20:17:14 <jsled> really? That's both disturbing and frustrating.
20:17:26 <IanL> yah,
20:17:28 <jsled> The package really should prefer them to the crap-hacked versions we distribute in-package.
20:17:39 <jsled> What versions are availble, of goffice especially...?
20:18:14 <IanL> libgoffice-gtk-0-3 0.3.0
20:19:08 <IanL> libgsf0.0-cil 0.8.1
20:19:33 <jsled> maybe a libgsf-1?
20:19:48 <IanL> libgsf-1-dev 1.14.1-2
20:20:21 <jsled> aye. Both of those should be great, along with a libgsf-gnome-1 or libgsf-1-gnome or however they want to say it.
20:20:41 <IanL> ok, I'll go ahead and install them,
20:21:12 <IanL> the cil one says it's cli bindings for libgsf. Would I need those?
20:21:42 <jsled> No, I don't believe so. IIRC it contains a couple of command-line util programs.
20:22:46 <IanL> a wait,
20:22:56 <IanL> I do have libgsf and libgoffice installed,
20:23:01 <IanL> my bad,
20:23:18 <IanL> libgsf 1.14.1-2
20:23:31 <IanL> and libgoffice 0.3.0
20:23:34 <IanL> but not the dev files,
20:24:07 <jsled> Ah.
20:24:09 <jsled> get those. :)
20:24:11 <IanL> ok,
20:24:23 <jsled> -dev[el] packages are your friends...
20:24:48 <jsled> So, what are you looking to change in gnucash...?
20:25:07 <IanL> Well, not anything in particular,
20:25:16 <IanL> maybe I would look at fixing some bugs in bugzilla,
20:26:32 <IanL> But one thing that bugs me is that if you type japanese in the register it doesn't show the characters as you type (with japanese you have to do completion before you actually enter the text)
20:27:54 <jsled> Ah.
20:29:41 <IanL> Anyway, I gotta go out,
20:29:51 <IanL> Thanks for your help,
20:30:01 <IanL> I'll try to play around with it later,
20:31:03 *** prock_ has joined #gnucash
20:31:13 *** IanL has quit IRC
20:39:59 *** prock has quit IRC
20:40:56 *** rlaager has quit IRC
20:46:45 *** prock has joined #gnucash
20:47:13 *** sjc has quit IRC
20:55:30 *** prock_ has quit IRC
21:45:33 *** dbr has joined #gnucash
21:45:57 <chris> jsled: Did you actually see that "Impossible? commiting pending" message?
21:48:18 <jsled> No. But I figured that if it's "Impossible?", it probably deserves a step up from "message".
21:48:49 <chris> ok, whew.
21:48:58 <jsled> oh, sorry. lol.
22:10:13 *** warlord-afk is now known as warlord
22:22:43 *** Esaj has quit IRC
22:26:38 *** Demitar has quit IRC
22:34:16 *** Demitar has joined #gnucash
22:41:57 <warlord> you've been busy, josh
23:33:57 <warlord> okay, bedtime for me.. Blah.
23:33:59 *** warlord is now known as warlord-afk