2007-01-10 GnuCash IRC logs
00:01:06 <warlord> You get locale not supported with fr_FR?
00:02:26 <contactium> warlord : I am with gnome in French version
00:03:23 <warlord> So you got it working?
00:07:55 <contactium> warlord : Sorry, I doesn't understand "got it working"
00:10:47 <warlord> contactium: if fr_FR doesn't put gnucash into French mode, then you probably don't have a french locale properly installed. You should contact your distribution vendor for more information.
00:11:47 <warlord> anyways, bedtime for me. good night.
00:11:49 *** warlord is now known as warlord-afk
00:12:01 <warlord-afk> (and for the record, "LANG=fr_FR gnucash" worked for me)
00:12:41 <contactium> OK, I look at that, thanks warlord
00:24:05 *** contactium has quit IRC
02:20:55 *** MacGyver has quit IRC
02:24:32 *** warlord-afk has quit IRC
02:28:09 *** MacGyver has joined #gnucash
02:32:59 *** benoitg has joined #gnucash
02:43:33 *** benoitg has left #gnucash
03:03:26 *** rjwiii has joined #gnucash
03:08:12 *** MacGyver has quit IRC
03:09:45 *** mr-rich has quit IRC
03:17:50 *** MacGyver has joined #gnucash
03:44:05 *** ErKa has joined #gnucash
03:55:02 *** MacGyver has quit IRC
04:31:56 *** pacman has quit IRC
04:31:56 *** motin has quit IRC
04:31:56 *** BC^bd has quit IRC
04:31:56 *** elb_ has quit IRC
04:31:56 *** piku has quit IRC
04:31:56 *** dwery has quit IRC
04:31:56 *** Def has quit IRC
04:31:56 *** jsled has quit IRC
04:31:56 *** conrad has quit IRC
04:31:56 *** rauch has quit IRC
04:31:56 *** wizkid238_ has quit IRC
04:31:56 *** slicslak has quit IRC
04:31:56 *** ceplma has quit IRC
04:31:58 *** wizkid238 has joined #gnucash
04:32:09 *** conrad has joined #gnucash
04:32:09 *** gncbot sets mode: +o conrad
04:32:18 *** jsled has joined #gnucash
04:32:18 *** gncbot sets mode: +o jsled
04:32:30 *** rauch has joined #gnucash
04:32:39 *** slicslak has joined #gnucash
04:32:40 *** Def has joined #gnucash
04:32:43 *** dwery has joined #gnucash
04:32:43 *** piku has joined #gnucash
04:32:43 *** elb_ has joined #gnucash
04:32:43 *** BC^bd has joined #gnucash
04:32:43 *** motin has joined #gnucash
04:32:43 *** pacman has joined #gnucash
07:21:36 *** ErKa has quit IRC
07:46:39 *** ErKa has joined #gnucash
07:46:40 *** Rolf has joined #gnucash
07:57:21 *** twunder has joined #gnucash
08:16:21 *** Demitar_ has joined #gnucash
08:16:36 *** cstim has joined #gnucash
08:16:36 *** gncbot sets mode: +o cstim
08:25:38 *** Demitar has quit IRC
08:30:34 *** Rolf has quit IRC
09:23:57 *** benoitg has joined #gnucash
09:27:16 *** warlord has joined #gnucash
09:27:17 *** gncbot sets mode: +o warlord
09:29:53 <warlord> hmm, I see no /topic
09:30:17 <elb_> possibly related to the server reboots yesterday
09:30:19 *** jsled changes topic to " Welcome! Don't meta-ask, ASK and WAIT. 2.0.4 released. FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
09:34:05 <warlord> could be
09:38:28 <cstim> good morning everyone
09:39:36 <warlord> hiya cstim
09:39:59 * chris waves good morning
09:40:45 <cstim> regarding 2.2 talk yesterday: as andi5 already said, I'd also appreciate a gnucash-2.2.0 version in the near future solely because we can announce the working windows port.
09:42:15 <cstim> of course if and only if the windows port turns out to be stable in a few months.
09:42:53 <chris> what are the outstanding win32 issues? zlib? ...?
09:44:39 <cstim> most importantly zlib, yes.
09:45:23 <chris> does price-updating work?
09:45:32 <cstim> http://wiki.gnucash.org/wiki/Windows#Todo : also Finance::Quote
09:45:41 <cstim> heh. no, not yet.
09:46:38 <chris> so IPC is the weak link.
09:46:48 <warlord> I think all of the 'fork' stuff is missing on win32
09:47:08 <warlord> We had a plan to use the g_spawn APIs, but I don't think that's happened, yet.
09:47:44 <chris> cstim: nice TODO summary - very clear.
09:48:05 <cstim> fork() itself is missing, but glib provides g_spawn which might be a usable replacement.
09:48:17 <cstim> chris: actually the credit for this goes to andi5 :-)
09:48:54 <chris> @tell andi5 Nice win32 TODO summary - very clear!
09:48:54 <gncbot> chris: The operation succeeded.
09:49:36 <chris> I think that randomness issue is a non-issue.
09:50:55 *** tats has joined #gnucash
09:51:39 <chris> we should just use g_random and be happy.
09:52:01 <cstim> Do you think 900 bytes of randomness are enough? If yes, then THRESHOLD in lib/libqof/qof/guid.c can be lowered accordingly.
09:54:06 <chris> I think 128 bytes are enough, but lowering the threshold is a band-aid.
09:55:39 <chris> is the F:Q issue a problem of dependencies or of IPC?
09:56:20 <warlord> chris: I think IPC
09:56:21 <cstim> I think it's a problem of guile because F:Q would be forked from guile. At least that's where the error message comes from.
09:57:14 *** ceplma has joined #gnucash
09:57:55 <chris> So, using g_spawn doesn't help there unless we were to rewrite price-quotes.scm into C.
09:58:39 <cstim> re F:Q, see quote on wiki page: Chris wrote: "That's process.scm which is run during price-quote retrieval. Perhaps guile's (primitive-fork) doesn't work on win32."
09:58:55 <cstim> chris: yes, apparently.
09:59:05 <warlord> chris: that's not true.. we could just swig-wrap the g_spawn API and use that instead of (primitive-fork)..
10:01:13 *** ceplma has quit IRC
10:01:36 <chris> warlord: that's an interesting idea.
10:24:01 *** Sound has joined #gnucash
10:25:26 <Sound> Hi all, I'm new to Gnucash. I have a question: why do vendor reports include the string "payment, thank you"? Is there a way to change it?
10:28:27 *** minDscrm has joined #gnucash
10:29:22 <cstim> I think it's in the FAQ
10:29:30 <jsled> Sound: probably because the report author wanted it to say that. If it's not in the report options, you probably need to change the report file itself.
10:29:49 <cstim> http://wiki.gnucash.org/wiki/FAQ#Using_Business_Features
10:30:39 <cstim> oh, I see it is not directly in the FAQ
10:31:46 <Sound> Thank you. I didn't think about the developer being funny :) However, that will require me to modify the report format on my local machine.
10:32:01 <cstim> I think to change that you need to modify the invoice.scm file manually.
10:32:09 <cstim> at least that's what the German FAQ says about this.
10:32:22 <jsled> "funny"?
10:33:00 <rauch> here it is /usr/share/gnucash/guile-modules/gnucash/report/*.scm, 4 files contain this string.
10:33:19 <Sound> jsled: I asked why reports include that string, and you didn't answer completely
10:33:32 <warlord> Sound: because a "vendor report" is generally not something you send out; it's for your own use.. And the code was written with Customer Invoices in mind, even though it's usable from Vendor Bills, too.
10:33:51 <Sound> warlord: thank you, that explains the thing to me.
10:34:13 <rauch> you may want to look at easy-invoice.scm, fancy-invoice.scm, invoice.scm & owner-report.scm
10:35:05 <Sound> Now, as a newbie I've got another question. Forgive me and help me get started with Gnucash :) Why is a payment address mandatory for vendors?
10:35:52 *** mnoir has joined #gnucash
10:36:08 <warlord> Sound: Where else are you going to mail the payment?
10:36:22 <Sound> warlord: I'm not going to use Gnucash to mail payments
10:36:38 <warlord> Then fill it in with garbage.
10:36:45 <warlord> It actually only requires 1 line of entry.
10:37:36 <Sound> warlord: I noticed that, but I was trying to understand why someone decided it's mandatory information
10:37:55 <warlord> Why not?
10:38:02 <warlord> It made the code simpler.
10:38:32 <Sound> warlord: that's an answer.
10:40:15 <Sound> Last question (be patient): is there a way to open the payable invoice by, say, double-clicking the transaction in the account register? of course, double-clicking didn't work
10:48:28 <warlord> Sound: no. There's no way to reach the Invoice from the register.
10:50:48 <Sound> warlord: thanks. I'll look for it in the bugzilla and then, if it wasn't already done, I'll submit a feature request.
10:51:45 <warlord> Sound: you're welcome to submit that, but architecturally doing that would be very challenging...
10:51:56 <warlord> the register exists outside the business features.
10:52:15 <warlord> ... and doesn't have the hooks necessary to plug in a menu item like that.
10:52:46 <Sound> warlord: isn't the registered already queried to determine the "paid/not paid" status for bills?
10:53:00 <warlord> Nope.
10:53:50 <Sound> warlord: however, if I delete a transaction related to a bill payment, the bill is instantly reported as not paid
10:54:44 <Sound> this made me think that there's some register-to-business mechanism
10:55:18 <warlord> That's because there are events...
10:55:24 <warlord> When you make changes to data.
10:56:35 <warlord> So the business features see the change events and can react. But that's not the same as a UI event.
10:57:20 *** esodan has joined #gnucash
10:58:32 <Sound> warlord: interesting. maybe one day I'll hack a bit on gnucash. Also, the Italian templates make gnucash fail with multi-byte errors (after digging I found that it's because they were probably created with old versions).
10:58:52 <warlord> Could be..
11:03:30 *** twunder has quit IRC
11:13:38 *** ErKa has quit IRC
11:17:04 *** MacGyver has joined #gnucash
11:26:03 *** twunder has joined #gnucash
11:31:08 *** motin has quit IRC
11:37:40 *** esodan has quit IRC
11:38:57 *** Sound has quit IRC
11:42:14 *** tats has quit IRC
12:02:15 *** ceplma has joined #gnucash
12:16:54 *** cstim has quit IRC
12:21:37 *** Rolf has joined #gnucash
12:35:53 *** andi5 has joined #gnucash
12:35:53 *** gncbot sets mode: +o andi5
13:05:23 *** benoitg has left #gnucash
13:10:36 *** esodan has joined #gnucash
13:31:35 *** andi5 has left #gnucash
13:40:31 *** ceplma has quit IRC
13:51:33 *** ceplma has joined #gnucash
13:52:09 *** sjc has joined #gnucash
13:55:19 *** |gunni| has joined #gnucash
14:05:39 *** motin has joined #gnucash
14:08:24 *** benoitg has joined #gnucash
14:43:26 *** andi5 has joined #gnucash
14:43:26 *** gncbot sets mode: +o andi5
14:44:32 *** Rolf has quit IRC
14:53:30 *** tats has joined #gnucash
14:56:33 *** sjc_ has joined #gnucash
15:05:09 *** ErKa has joined #gnucash
15:05:22 *** sjc has quit IRC
15:08:05 *** andi5 has left #gnucash
15:36:39 *** magenpie has quit IRC
15:57:45 *** dwery has quit IRC
16:06:06 *** twunder has quit IRC
16:11:45 *** dwery has joined #gnucash
16:25:34 *** magenpie has joined #gnucash
16:25:35 *** dwery has quit IRC
16:40:37 *** dwery has joined #gnucash
17:16:09 *** esodan has quit IRC
17:18:13 *** warlord is now known as warlord-afk
17:39:42 *** bonez39 has joined #gnucash
17:59:44 *** andi5 has joined #gnucash
17:59:45 *** gncbot sets mode: +o andi5
18:03:48 <andi5> hi.... i wonder whether we save the filenames in our data structures in either UTF-8 or locale encoding... i cannot see anything in the code taking this difference into account
18:18:17 *** Geot has joined #gnucash
18:22:27 *** benoitg has left #gnucash
18:51:30 *** ErKa has quit IRC
18:59:03 *** |gunni| has quit IRC
19:09:23 *** sjc__ has joined #gnucash
19:09:58 *** tats has quit IRC
19:17:23 *** sjc_ has quit IRC
19:30:35 *** esodan has joined #gnucash
19:30:52 <esodan> worlord: jsled: chris: (all one) please review my e-mail where I had coded some example in how to implement GObject directly in the engine...
19:42:40 *** esodan has quit IRC
19:58:37 *** bespin has joined #gnucash
20:00:39 *** warlord-afk is now known as warlord
20:01:01 <warlord> andi5: That's a good question.
20:01:07 *** magenpie has quit IRC
20:01:44 <andi5> hi warlord
20:02:08 *** tats has joined #gnucash
20:02:39 <warlord> hi
20:03:11 *** Demitar_ is now known as Demitar
20:03:24 <andi5> well, we should differentiate somewhere, i.e. translate with the help of g_filename_to_utf8 and g_filename_from_utf8 ... or otherwise umlauts and whatever look strange either in the windows explorer or the file chooser
20:04:23 *** slicslak has quit IRC
20:05:01 *** bespin has left #gnucash
20:08:22 <warlord> andi5: sounds fair tome
20:08:54 <andi5> but where... :)
20:11:40 *** mnoir has quit IRC
20:12:49 <warlord> I dont know...
20:12:55 *** magenpie has joined #gnucash
20:12:59 <warlord> As close to the file operations as possible, I'd say.
20:13:27 *** tats has quit IRC
20:15:21 <andi5> yes... sometimes i think the other way around might be good too.... but i guess everything except fopen, readdir, stat and whatever could use utf-8 (i.e. all filenames printed/saved anywhere)... right?
20:15:41 <warlord> Yep
20:21:32 <andi5> i am searching for a name i can use to help with this differentiation in function arguments... something like "localepath" "lpath", "opsys_path"... i do not know
20:25:06 <warlord> localepath or lpath works.
20:26:56 <jsled> "localepath", please. Character futures are down in afternoon trading. :)
20:27:12 <andi5> actually we are not talking about the locale encoding but the encoding glib uses for file names, something that can be influenced by the G_FILENAME_ENCODING env variable.... but i guess this will not hurt
20:40:35 *** sjc__ has quit IRC
20:42:50 <andi5> Q: can we depend on gtk2.6 on trunk, please? (thinking about g_stat right now)
20:44:25 <andi5> glib2.6, i mean (sorry)
20:44:28 <jsled> g_stat's in gtk, or glib?
20:44:29 <jsled> heh
20:45:36 <jsled> I think that's fine, but I'm fine with very current dependencies for trunk.
20:46:18 <jsled> I suppose that if trunk depends on swig-what.ever, then all other deps that were released about the same time (or before, obviously) as that version of swig are probably fine as well.
20:46:53 <andi5> yeah, but users building from tarballs will need glib2.6 then
20:48:03 <andi5> jsled: but hey, i suggested it (again), so i am fine with it too ;-)
20:49:18 <jsled> at the earliest possible point we coudl release (spring), distributions from 2 years ago will have (at least) glib-2.6.
20:50:03 <jsled> Even FC4 has glib-2.6.
20:56:57 <warlord> jsled: the difference is that the swig dependency is only required for building the dist tarball (or building from SVN). There's no swig dependency at all to build from a trunk tarball.
20:58:21 <warlord> I'd like to wait to hear from hampton... Right now we only need glib/gtk 2.4, and all the 2.6isms that we'd require are in core-utils. HOWEVER I have no objection of changing the base dep to 2.6
20:59:28 * andi5 puts g_stat on his wishlist
21:21:12 *** hampton has joined #gnucash
21:22:11 *** gncbot sets mode: +o hampton
21:22:42 <warlord> hampton: you should go read the logs.
21:22:47 *** warlord is now known as warlord-afk
21:23:45 * hampton waves from his new diggs
21:23:48 <hampton> ok
21:24:21 <hampton> got a url?
21:24:40 <andi5> see the topic :)
21:24:52 <hampton> Just got my internet this afternoon and still have to set up my computers. On my laptop at the moment.
21:25:01 <hampton> andi5: thanks. (he says sheepishly)
21:26:27 <andi5> hehe :) .... hampton, what about remove-group, do you think you can merge that into trunk?
21:27:24 <hampton> needs some more testing, but I should be able to get to it in the next week or two. iirc, 'make check' passed successfully before I got sidetracked
21:27:24 <andi5> sorry, somehow i have got the impression that the branch is ready... correct me if i am wrong :)
21:27:43 <andi5> cool :)
21:31:32 <hampton> iirc the only reason we had partial glib26 support built in (lib/glib26) in was to support fc3/rhel?
21:32:00 <hampton> I think it was warlord that wanted that because of the mit upgrade cycle
21:32:38 <andi5> Q: there is no reason for searchpaths and xaccUserPathPathGenerator to stay, right? .... another Q: what is this "slash->comma" patched path good for? (engine/gnc-filepath-utils.c:242)
21:35:36 * hampton agrees with jsled that our swig dependency is the gating requirement at this point, and everything else can be brought forward to the same time frame.
21:37:48 * andi5 agrees with warlord that we cannot use swig as comparison as it is only an SVN build dependency.... maybe some other time frame?
21:38:53 <hampton> as far as file name encoding, I think all the g_xxx file access functions require the name in the system locale encoding (whatever that is), but all of gtk requires text to be in utf8. Gnucash should use utf8 internally, and only convert to the locale encoding immediately before calling a g_xxx file function.
21:39:18 <andi5> ohhh... i do not think so
21:39:38 <andi5> to me it looks like g_mkdir and friends all use utf-8 paths
21:40:02 * andi5 takes a 3rd look
21:40:16 <hampton> you're right on that one
21:40:50 <hampton> so, gnucash should only convert from utf8 to whatever immediately before calling a g_xxx function that requires the file name to already be in the system encoding. :-)
21:41:50 <andi5> oh man... i was totally wrong *bows down*
21:42:14 <andi5> maybe kneeing is better?
21:42:14 <hampton> although, looking at devhelp I'm seeing that for numerous g_xxx function calls the docs explicitly state that on windows the filename should be in utf8
21:43:35 <hampton> from devhelp:
21:43:52 <hampton> On Windows the GLib file name encoding is UTF-8. Note that the Microsoft C library does not use UTF-8, but has separate APIs for current system code page and wide characters (UTF-16). The GLib wrappers call the wide character API if present (on modern Windows systems), otherwise convert to/from the system code page.
21:44:33 <andi5> yeah, so g_filename_from_utf8 probably does nothing on windows, right?
21:44:53 <andi5> and my problem is the missing utf-8<->utf-16 conversion
21:45:21 <hampton> probably not. Its not a bad idea to call it anyway for other systems that might not have utf8 as the required encoding.
21:46:33 <andi5> yes... but it also means to avoid 'stat' and friends if we do not want to call g_utf8_to_utf16 all the time
21:48:25 <hampton> why would gnucash ever explicitly call a utf16 function? The only two states gnucash should know about are utf8 and the glib encoding. The whole point of glib is that it handles the underlying system specific crap and provides a common ui to applications.
21:48:53 <hampton> yes, I know that's the high level api and you're dealing with a real world problem.
21:49:13 <hampton> I'd just hate to see system specific code creep in if generic code can solve the problem.
21:49:33 <andi5> actually, i want to remove system specific code ;-)
21:49:41 <hampton> where are you getting the utf16 name from?
21:55:17 <andi5> no, i was wrong again (too late for me)... we use wide character support nowhere in the gnucash source, so i had a utf-8 string that got feed to 'fopen' directly.... i do not know how windows handles non-ascii characters
21:56:27 <andi5> well, i will try to make a patch over the weekend.... would love it if you could take a look at it then, hampton
21:59:45 * andi5 zzzz....
21:59:48 *** andi5 has quit IRC
22:07:32 <hampton> gncbot: tell andi5 fopen is bad. g_fopen is good. :-)
22:07:32 <gncbot> hampton: The operation succeeded.
22:08:40 <hampton> ahh, that's where the request for glib2.6 came from.....
22:17:11 <chris> hi hampton.
22:18:02 <chris> did you catch what andi5 and warlord said about swig (non-)dependency? (jsled should take note, too)
22:19:24 <chris> I think it's important to note, because I actually want our dev-only swig dep to be much newer than our minumum target deps.
22:19:38 <jsled> it is a good point.
22:20:24 <jsled> But glib-2.4 is "ancient" and glib-2.6 is still "really old"; glib-2.6 is fine.
22:20:44 <chris> But, on the orginal topic, I'm fine w/ glib-2.6 for gnucash 2.2.
22:24:44 <hampton> chris: I'm not ignoring your email from the other day, I just haven't had time to do anything with it.
22:24:58 *** warlord-afk is now known as warlord
22:25:13 <warlord> w.r.t glib/gtk-2.4.. It was for FC3 and RHEL4.
22:25:17 <warlord> Has there been an RHEL5 yet?
22:28:01 <jsled> who cares?
22:29:34 <hampton> rhel5 beta 2 was released november 16th
22:30:00 *** jpeach has joined #gnucash
22:30:04 *** jpeach has left #gnucash
22:30:46 <warlord> *nods*
22:31:01 <warlord> I have no objection to trunk depending on something newer
22:45:47 *** tats has joined #gnucash
22:57:17 *** tats has quit IRC
00:01:06 <warlord> You get locale not supported with fr_FR?
00:02:26 <contactium> warlord : I am with gnome in French version
00:03:23 <warlord> So you got it working?
00:07:55 <contactium> warlord : Sorry, I doesn't understand "got it working"
00:10:47 <warlord> contactium: if fr_FR doesn't put gnucash into French mode, then you probably don't have a french locale properly installed. You should contact your distribution vendor for more information.
00:11:47 <warlord> anyways, bedtime for me. good night.
00:11:49 *** warlord is now known as warlord-afk
00:12:01 <warlord-afk> (and for the record, "LANG=fr_FR gnucash" worked for me)
00:12:41 <contactium> OK, I look at that, thanks warlord
00:24:05 *** contactium has quit IRC
02:20:55 *** MacGyver has quit IRC
02:24:32 *** warlord-afk has quit IRC
02:28:09 *** MacGyver has joined #gnucash
02:32:59 *** benoitg has joined #gnucash
02:43:33 *** benoitg has left #gnucash
03:03:26 *** rjwiii has joined #gnucash
03:08:12 *** MacGyver has quit IRC
03:09:45 *** mr-rich has quit IRC
03:17:50 *** MacGyver has joined #gnucash
03:44:05 *** ErKa has joined #gnucash
03:55:02 *** MacGyver has quit IRC
04:31:56 *** pacman has quit IRC
04:31:56 *** motin has quit IRC
04:31:56 *** BC^bd has quit IRC
04:31:56 *** elb_ has quit IRC
04:31:56 *** piku has quit IRC
04:31:56 *** dwery has quit IRC
04:31:56 *** Def has quit IRC
04:31:56 *** jsled has quit IRC
04:31:56 *** conrad has quit IRC
04:31:56 *** rauch has quit IRC
04:31:56 *** wizkid238_ has quit IRC
04:31:56 *** slicslak has quit IRC
04:31:56 *** ceplma has quit IRC
04:31:58 *** wizkid238 has joined #gnucash
04:32:09 *** conrad has joined #gnucash
04:32:09 *** gncbot sets mode: +o conrad
04:32:18 *** jsled has joined #gnucash
04:32:18 *** gncbot sets mode: +o jsled
04:32:30 *** rauch has joined #gnucash
04:32:39 *** slicslak has joined #gnucash
04:32:40 *** Def has joined #gnucash
04:32:43 *** dwery has joined #gnucash
04:32:43 *** piku has joined #gnucash
04:32:43 *** elb_ has joined #gnucash
04:32:43 *** BC^bd has joined #gnucash
04:32:43 *** motin has joined #gnucash
04:32:43 *** pacman has joined #gnucash
07:21:36 *** ErKa has quit IRC
07:46:39 *** ErKa has joined #gnucash
07:46:40 *** Rolf has joined #gnucash
07:57:21 *** twunder has joined #gnucash
08:16:21 *** Demitar_ has joined #gnucash
08:16:36 *** cstim has joined #gnucash
08:16:36 *** gncbot sets mode: +o cstim
08:25:38 *** Demitar has quit IRC
08:30:34 *** Rolf has quit IRC
09:23:57 *** benoitg has joined #gnucash
09:27:16 *** warlord has joined #gnucash
09:27:17 *** gncbot sets mode: +o warlord
09:29:53 <warlord> hmm, I see no /topic
09:30:17 <elb_> possibly related to the server reboots yesterday
09:30:19 *** jsled changes topic to " Welcome! Don't meta-ask, ASK and WAIT. 2.0.4 released. FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
09:34:05 <warlord> could be
09:38:28 <cstim> good morning everyone
09:39:36 <warlord> hiya cstim
09:39:59 * chris waves good morning
09:40:45 <cstim> regarding 2.2 talk yesterday: as andi5 already said, I'd also appreciate a gnucash-2.2.0 version in the near future solely because we can announce the working windows port.
09:42:15 <cstim> of course if and only if the windows port turns out to be stable in a few months.
09:42:53 <chris> what are the outstanding win32 issues? zlib? ...?
09:44:39 <cstim> most importantly zlib, yes.
09:45:23 <chris> does price-updating work?
09:45:32 <cstim> http://wiki.gnucash.org/wiki/Windows#Todo : also Finance::Quote
09:45:41 <cstim> heh. no, not yet.
09:46:38 <chris> so IPC is the weak link.
09:46:48 <warlord> I think all of the 'fork' stuff is missing on win32
09:47:08 <warlord> We had a plan to use the g_spawn APIs, but I don't think that's happened, yet.
09:47:44 <chris> cstim: nice TODO summary - very clear.
09:48:05 <cstim> fork() itself is missing, but glib provides g_spawn which might be a usable replacement.
09:48:17 <cstim> chris: actually the credit for this goes to andi5 :-)
09:48:54 <chris> @tell andi5 Nice win32 TODO summary - very clear!
09:48:54 <gncbot> chris: The operation succeeded.
09:49:36 <chris> I think that randomness issue is a non-issue.
09:50:55 *** tats has joined #gnucash
09:51:39 <chris> we should just use g_random and be happy.
09:52:01 <cstim> Do you think 900 bytes of randomness are enough? If yes, then THRESHOLD in lib/libqof/qof/guid.c can be lowered accordingly.
09:54:06 <chris> I think 128 bytes are enough, but lowering the threshold is a band-aid.
09:55:39 <chris> is the F:Q issue a problem of dependencies or of IPC?
09:56:20 <warlord> chris: I think IPC
09:56:21 <cstim> I think it's a problem of guile because F:Q would be forked from guile. At least that's where the error message comes from.
09:57:14 *** ceplma has joined #gnucash
09:57:55 <chris> So, using g_spawn doesn't help there unless we were to rewrite price-quotes.scm into C.
09:58:39 <cstim> re F:Q, see quote on wiki page: Chris wrote: "That's process.scm which is run during price-quote retrieval. Perhaps guile's (primitive-fork) doesn't work on win32."
09:58:55 <cstim> chris: yes, apparently.
09:59:05 <warlord> chris: that's not true.. we could just swig-wrap the g_spawn API and use that instead of (primitive-fork)..
10:01:13 *** ceplma has quit IRC
10:01:36 <chris> warlord: that's an interesting idea.
10:24:01 *** Sound has joined #gnucash
10:25:26 <Sound> Hi all, I'm new to Gnucash. I have a question: why do vendor reports include the string "payment, thank you"? Is there a way to change it?
10:28:27 *** minDscrm has joined #gnucash
10:29:22 <cstim> I think it's in the FAQ
10:29:30 <jsled> Sound: probably because the report author wanted it to say that. If it's not in the report options, you probably need to change the report file itself.
10:29:49 <cstim> http://wiki.gnucash.org/wiki/FAQ#Using_Business_Features
10:30:39 <cstim> oh, I see it is not directly in the FAQ
10:31:46 <Sound> Thank you. I didn't think about the developer being funny :) However, that will require me to modify the report format on my local machine.
10:32:01 <cstim> I think to change that you need to modify the invoice.scm file manually.
10:32:09 <cstim> at least that's what the German FAQ says about this.
10:32:22 <jsled> "funny"?
10:33:00 <rauch> here it is /usr/share/gnucash/guile-modules/gnucash/report/*.scm, 4 files contain this string.
10:33:19 <Sound> jsled: I asked why reports include that string, and you didn't answer completely
10:33:32 <warlord> Sound: because a "vendor report" is generally not something you send out; it's for your own use.. And the code was written with Customer Invoices in mind, even though it's usable from Vendor Bills, too.
10:33:51 <Sound> warlord: thank you, that explains the thing to me.
10:34:13 <rauch> you may want to look at easy-invoice.scm, fancy-invoice.scm, invoice.scm & owner-report.scm
10:35:05 <Sound> Now, as a newbie I've got another question. Forgive me and help me get started with Gnucash :) Why is a payment address mandatory for vendors?
10:35:52 *** mnoir has joined #gnucash
10:36:08 <warlord> Sound: Where else are you going to mail the payment?
10:36:22 <Sound> warlord: I'm not going to use Gnucash to mail payments
10:36:38 <warlord> Then fill it in with garbage.
10:36:45 <warlord> It actually only requires 1 line of entry.
10:37:36 <Sound> warlord: I noticed that, but I was trying to understand why someone decided it's mandatory information
10:37:55 <warlord> Why not?
10:38:02 <warlord> It made the code simpler.
10:38:32 <Sound> warlord: that's an answer.
10:40:15 <Sound> Last question (be patient): is there a way to open the payable invoice by, say, double-clicking the transaction in the account register? of course, double-clicking didn't work
10:48:28 <warlord> Sound: no. There's no way to reach the Invoice from the register.
10:50:48 <Sound> warlord: thanks. I'll look for it in the bugzilla and then, if it wasn't already done, I'll submit a feature request.
10:51:45 <warlord> Sound: you're welcome to submit that, but architecturally doing that would be very challenging...
10:51:56 <warlord> the register exists outside the business features.
10:52:15 <warlord> ... and doesn't have the hooks necessary to plug in a menu item like that.
10:52:46 <Sound> warlord: isn't the registered already queried to determine the "paid/not paid" status for bills?
10:53:00 <warlord> Nope.
10:53:50 <Sound> warlord: however, if I delete a transaction related to a bill payment, the bill is instantly reported as not paid
10:54:44 <Sound> this made me think that there's some register-to-business mechanism
10:55:18 <warlord> That's because there are events...
10:55:24 <warlord> When you make changes to data.
10:56:35 <warlord> So the business features see the change events and can react. But that's not the same as a UI event.
10:57:20 *** esodan has joined #gnucash
10:58:32 <Sound> warlord: interesting. maybe one day I'll hack a bit on gnucash. Also, the Italian templates make gnucash fail with multi-byte errors (after digging I found that it's because they were probably created with old versions).
10:58:52 <warlord> Could be..
11:03:30 *** twunder has quit IRC
11:13:38 *** ErKa has quit IRC
11:17:04 *** MacGyver has joined #gnucash
11:26:03 *** twunder has joined #gnucash
11:31:08 *** motin has quit IRC
11:37:40 *** esodan has quit IRC
11:38:57 *** Sound has quit IRC
11:42:14 *** tats has quit IRC
12:02:15 *** ceplma has joined #gnucash
12:16:54 *** cstim has quit IRC
12:21:37 *** Rolf has joined #gnucash
12:35:53 *** andi5 has joined #gnucash
12:35:53 *** gncbot sets mode: +o andi5
13:05:23 *** benoitg has left #gnucash
13:10:36 *** esodan has joined #gnucash
13:31:35 *** andi5 has left #gnucash
13:40:31 *** ceplma has quit IRC
13:51:33 *** ceplma has joined #gnucash
13:52:09 *** sjc has joined #gnucash
13:55:19 *** |gunni| has joined #gnucash
14:05:39 *** motin has joined #gnucash
14:08:24 *** benoitg has joined #gnucash
14:43:26 *** andi5 has joined #gnucash
14:43:26 *** gncbot sets mode: +o andi5
14:44:32 *** Rolf has quit IRC
14:53:30 *** tats has joined #gnucash
14:56:33 *** sjc_ has joined #gnucash
15:05:09 *** ErKa has joined #gnucash
15:05:22 *** sjc has quit IRC
15:08:05 *** andi5 has left #gnucash
15:36:39 *** magenpie has quit IRC
15:57:45 *** dwery has quit IRC
16:06:06 *** twunder has quit IRC
16:11:45 *** dwery has joined #gnucash
16:25:34 *** magenpie has joined #gnucash
16:25:35 *** dwery has quit IRC
16:40:37 *** dwery has joined #gnucash
17:16:09 *** esodan has quit IRC
17:18:13 *** warlord is now known as warlord-afk
17:39:42 *** bonez39 has joined #gnucash
17:59:44 *** andi5 has joined #gnucash
17:59:45 *** gncbot sets mode: +o andi5
18:03:48 <andi5> hi.... i wonder whether we save the filenames in our data structures in either UTF-8 or locale encoding... i cannot see anything in the code taking this difference into account
18:18:17 *** Geot has joined #gnucash
18:22:27 *** benoitg has left #gnucash
18:51:30 *** ErKa has quit IRC
18:59:03 *** |gunni| has quit IRC
19:09:23 *** sjc__ has joined #gnucash
19:09:58 *** tats has quit IRC
19:17:23 *** sjc_ has quit IRC
19:30:35 *** esodan has joined #gnucash
19:30:52 <esodan> worlord: jsled: chris: (all one) please review my e-mail where I had coded some example in how to implement GObject directly in the engine...
19:42:40 *** esodan has quit IRC
19:58:37 *** bespin has joined #gnucash
20:00:39 *** warlord-afk is now known as warlord
20:01:01 <warlord> andi5: That's a good question.
20:01:07 *** magenpie has quit IRC
20:01:44 <andi5> hi warlord
20:02:08 *** tats has joined #gnucash
20:02:39 <warlord> hi
20:03:11 *** Demitar_ is now known as Demitar
20:03:24 <andi5> well, we should differentiate somewhere, i.e. translate with the help of g_filename_to_utf8 and g_filename_from_utf8 ... or otherwise umlauts and whatever look strange either in the windows explorer or the file chooser
20:04:23 *** slicslak has quit IRC
20:05:01 *** bespin has left #gnucash
20:08:22 <warlord> andi5: sounds fair tome
20:08:54 <andi5> but where... :)
20:11:40 *** mnoir has quit IRC
20:12:49 <warlord> I dont know...
20:12:55 *** magenpie has joined #gnucash
20:12:59 <warlord> As close to the file operations as possible, I'd say.
20:13:27 *** tats has quit IRC
20:15:21 <andi5> yes... sometimes i think the other way around might be good too.... but i guess everything except fopen, readdir, stat and whatever could use utf-8 (i.e. all filenames printed/saved anywhere)... right?
20:15:41 <warlord> Yep
20:21:32 <andi5> i am searching for a name i can use to help with this differentiation in function arguments... something like "localepath" "lpath", "opsys_path"... i do not know
20:25:06 <warlord> localepath or lpath works.
20:26:56 <jsled> "localepath", please. Character futures are down in afternoon trading. :)
20:27:12 <andi5> actually we are not talking about the locale encoding but the encoding glib uses for file names, something that can be influenced by the G_FILENAME_ENCODING env variable.... but i guess this will not hurt
20:40:35 *** sjc__ has quit IRC
20:42:50 <andi5> Q: can we depend on gtk2.6 on trunk, please? (thinking about g_stat right now)
20:44:25 <andi5> glib2.6, i mean (sorry)
20:44:28 <jsled> g_stat's in gtk, or glib?
20:44:29 <jsled> heh
20:45:36 <jsled> I think that's fine, but I'm fine with very current dependencies for trunk.
20:46:18 <jsled> I suppose that if trunk depends on swig-what.ever, then all other deps that were released about the same time (or before, obviously) as that version of swig are probably fine as well.
20:46:53 <andi5> yeah, but users building from tarballs will need glib2.6 then
20:48:03 <andi5> jsled: but hey, i suggested it (again), so i am fine with it too ;-)
20:49:18 <jsled> at the earliest possible point we coudl release (spring), distributions from 2 years ago will have (at least) glib-2.6.
20:50:03 <jsled> Even FC4 has glib-2.6.
20:56:57 <warlord> jsled: the difference is that the swig dependency is only required for building the dist tarball (or building from SVN). There's no swig dependency at all to build from a trunk tarball.
20:58:21 <warlord> I'd like to wait to hear from hampton... Right now we only need glib/gtk 2.4, and all the 2.6isms that we'd require are in core-utils. HOWEVER I have no objection of changing the base dep to 2.6
20:59:28 * andi5 puts g_stat on his wishlist
21:21:12 *** hampton has joined #gnucash
21:22:11 *** gncbot sets mode: +o hampton
21:22:42 <warlord> hampton: you should go read the logs.
21:22:47 *** warlord is now known as warlord-afk
21:23:45 * hampton waves from his new diggs
21:23:48 <hampton> ok
21:24:21 <hampton> got a url?
21:24:40 <andi5> see the topic :)
21:24:52 <hampton> Just got my internet this afternoon and still have to set up my computers. On my laptop at the moment.
21:25:01 <hampton> andi5: thanks. (he says sheepishly)
21:26:27 <andi5> hehe :) .... hampton, what about remove-group, do you think you can merge that into trunk?
21:27:24 <hampton> needs some more testing, but I should be able to get to it in the next week or two. iirc, 'make check' passed successfully before I got sidetracked
21:27:24 <andi5> sorry, somehow i have got the impression that the branch is ready... correct me if i am wrong :)
21:27:43 <andi5> cool :)
21:31:32 <hampton> iirc the only reason we had partial glib26 support built in (lib/glib26) in was to support fc3/rhel?
21:32:00 <hampton> I think it was warlord that wanted that because of the mit upgrade cycle
21:32:38 <andi5> Q: there is no reason for searchpaths and xaccUserPathPathGenerator to stay, right? .... another Q: what is this "slash->comma" patched path good for? (engine/gnc-filepath-utils.c:242)
21:35:36 * hampton agrees with jsled that our swig dependency is the gating requirement at this point, and everything else can be brought forward to the same time frame.
21:37:48 * andi5 agrees with warlord that we cannot use swig as comparison as it is only an SVN build dependency.... maybe some other time frame?
21:38:53 <hampton> as far as file name encoding, I think all the g_xxx file access functions require the name in the system locale encoding (whatever that is), but all of gtk requires text to be in utf8. Gnucash should use utf8 internally, and only convert to the locale encoding immediately before calling a g_xxx file function.
21:39:18 <andi5> ohhh... i do not think so
21:39:38 <andi5> to me it looks like g_mkdir and friends all use utf-8 paths
21:40:02 * andi5 takes a 3rd look
21:40:16 <hampton> you're right on that one
21:40:50 <hampton> so, gnucash should only convert from utf8 to whatever immediately before calling a g_xxx function that requires the file name to already be in the system encoding. :-)
21:41:50 <andi5> oh man... i was totally wrong *bows down*
21:42:14 <andi5> maybe kneeing is better?
21:42:14 <hampton> although, looking at devhelp I'm seeing that for numerous g_xxx function calls the docs explicitly state that on windows the filename should be in utf8
21:43:35 <hampton> from devhelp:
21:43:52 <hampton> On Windows the GLib file name encoding is UTF-8. Note that the Microsoft C library does not use UTF-8, but has separate APIs for current system code page and wide characters (UTF-16). The GLib wrappers call the wide character API if present (on modern Windows systems), otherwise convert to/from the system code page.
21:44:33 <andi5> yeah, so g_filename_from_utf8 probably does nothing on windows, right?
21:44:53 <andi5> and my problem is the missing utf-8<->utf-16 conversion
21:45:21 <hampton> probably not. Its not a bad idea to call it anyway for other systems that might not have utf8 as the required encoding.
21:46:33 <andi5> yes... but it also means to avoid 'stat' and friends if we do not want to call g_utf8_to_utf16 all the time
21:48:25 <hampton> why would gnucash ever explicitly call a utf16 function? The only two states gnucash should know about are utf8 and the glib encoding. The whole point of glib is that it handles the underlying system specific crap and provides a common ui to applications.
21:48:53 <hampton> yes, I know that's the high level api and you're dealing with a real world problem.
21:49:13 <hampton> I'd just hate to see system specific code creep in if generic code can solve the problem.
21:49:33 <andi5> actually, i want to remove system specific code ;-)
21:49:41 <hampton> where are you getting the utf16 name from?
21:55:17 <andi5> no, i was wrong again (too late for me)... we use wide character support nowhere in the gnucash source, so i had a utf-8 string that got feed to 'fopen' directly.... i do not know how windows handles non-ascii characters
21:56:27 <andi5> well, i will try to make a patch over the weekend.... would love it if you could take a look at it then, hampton
21:59:45 * andi5 zzzz....
21:59:48 *** andi5 has quit IRC
22:07:32 <hampton> gncbot: tell andi5 fopen is bad. g_fopen is good. :-)
22:07:32 <gncbot> hampton: The operation succeeded.
22:08:40 <hampton> ahh, that's where the request for glib2.6 came from.....
22:17:11 <chris> hi hampton.
22:18:02 <chris> did you catch what andi5 and warlord said about swig (non-)dependency? (jsled should take note, too)
22:19:24 <chris> I think it's important to note, because I actually want our dev-only swig dep to be much newer than our minumum target deps.
22:19:38 <jsled> it is a good point.
22:20:24 <jsled> But glib-2.4 is "ancient" and glib-2.6 is still "really old"; glib-2.6 is fine.
22:20:44 <chris> But, on the orginal topic, I'm fine w/ glib-2.6 for gnucash 2.2.
22:24:44 <hampton> chris: I'm not ignoring your email from the other day, I just haven't had time to do anything with it.
22:24:58 *** warlord-afk is now known as warlord
22:25:13 <warlord> w.r.t glib/gtk-2.4.. It was for FC3 and RHEL4.
22:25:17 <warlord> Has there been an RHEL5 yet?
22:28:01 <jsled> who cares?
22:29:34 <hampton> rhel5 beta 2 was released november 16th
22:30:00 *** jpeach has joined #gnucash
22:30:04 *** jpeach has left #gnucash
22:30:46 <warlord> *nods*
22:31:01 <warlord> I have no objection to trunk depending on something newer
22:45:47 *** tats has joined #gnucash
22:57:17 *** tats has quit IRC