2007-02-16 GnuCash IRC logs

00:38:54 *** dbr has quit IRC
01:15:15 *** Wilddev has quit IRC
01:35:51 *** prock_ has joined #gnucash
01:44:23 *** prock has quit IRC
01:47:01 *** hampton is now known as hampton|away
02:32:39 *** ErKa has joined #gnucash
03:02:54 *** ErKa has quit IRC
03:04:36 *** px79 has joined #gnucash
03:05:06 <px79> Hi!
03:05:47 <px79> Some stupid kid messed up the GNUCash Wiki. :(
03:06:29 <px79> Examples: http://wiki.gnucash.org/wiki/FAQ
03:06:42 <px79> or: http://wiki.gnucash.org/wiki/Special:Recentchanges
03:08:25 <px79> I also have posted to: gnucash-de@gnucash.org
03:11:44 <px79> Hope someone has a backup or can fix it any other way.
03:12:28 <px79> Sad thing! Since GNUCash is a great project!
03:12:28 <px79> BTW: Thx for this application.
03:14:28 <px79> Bye!
03:14:36 *** px79 has left #gnucash
03:15:03 *** warlord-afk has quit IRC
04:17:15 *** cstim has joined #gnucash
04:17:15 *** gncbot sets mode: +o cstim
04:24:47 <cstim> warlord or jsled: If you can change the wiki configuration, please switch off the permission to "move pages" for non-admin users.
04:24:47 <gncbot> cstim: Sent 1 day, 17 hours, and 26 minutes ago: <prock> you mentioned that my 'macintel' patch caused the win32 build to fail. Where/how did it fail?
04:25:41 <cstim> @tell prock: re macintel patch: I've written the failure in detail in the attachment at http://lists.gnucash.org/pipermail/gnucash-devel/2007-February/019894.html
04:25:41 <gncbot> cstim: The operation succeeded.
05:28:56 *** ErKa has joined #gnucash
05:31:09 *** exper_ has joined #gnucash
05:39:01 *** exper has quit IRC
05:54:32 *** BC^bd has quit IRC
06:05:47 *** exper_ is now known as exper
06:17:12 *** egsavage has joined #gnucash
06:42:52 *** ErKa has quit IRC
06:58:57 *** twunder has joined #gnucash
07:14:00 *** egsavage has quit IRC
07:17:22 *** egsavage has joined #gnucash
07:22:50 *** twunder has quit IRC
07:58:01 *** twunder has joined #gnucash
08:25:44 *** ErKa has joined #gnucash
08:26:05 *** ErKa has joined #gnucash
08:28:28 <prock_> .
08:28:32 *** prock_ is now known as prock
08:28:32 <prock> .
08:32:03 *** jpeach has joined #gnucash
08:58:04 <chris> g'morning prock.
08:58:37 <cstim> good morning everyone.
08:59:31 <chris> prock: I must admit that I found it somewhat amusing that you got completely different answers to your modularization question - and from people that know what they're talking about.
09:00:06 <chris> cstim: Morgen.
09:01:06 *** jpeach has left #gnucash
09:01:07 <chris> prock: I think the problem was that there's an implicit alternative to your efforts that different people infer differently.
09:02:15 <chris> E.g. If I had to choose between what we have now and a completely uniformed shared object system, I'd side with jsled and say the latter.
09:03:23 <chris> OTOH, If I get to choose between what we have now and a more narrowly defined and better designed run-time GModule system, I'd side with warlord and hampton for the latter.
09:04:29 <chris> Really, either one is an improvement and one downside to the GModule direction is that those interfaces have really never been designed.
09:05:33 *** IanL has joined #gnucash
09:06:05 <chris> So, it's more a question of doing the easy thing that's still an improvement (all shared objects) or the harder thing that's nicer still (design a real run-time GModule/plugin interface)
09:06:41 * prock nods
09:07:58 <prock> following some irc conversations with warlord/hampton, jsled has agreed with turning the modules into actual modules.
09:07:59 <chris> As for the particulars, I think you'll find quite a few opinions. Personally, I'd rather the modules were only used for specific types of functionality - importer/exports, backends
09:09:38 <chris> Two dimensions that are currently "modular" that I don't think should be are "gui" and "business". Those should be compile-time links, IMO.
09:09:56 <chris> But I'm pretty sure warlord would disagree on both counts.
09:10:53 <prock> I'm very happily trying to stay far away from the "what should be a module and what should be a library" discussion. You are all much more qualified than I to decide such things.
09:11:42 <chris> The problem with modules as they are is that the interface is practically no more rigid than a .so link anyway, so it can be used for anything, and it ended up being used for everything.
09:12:38 <prock> right... I'm going to try to migrate the modules that definately should be modules towards an event based system where (upon module initialization) they will register for particular events.
09:13:08 * chris nods.
09:13:12 <prock> but first I need to just plain tear them apart... so there will be some symbol lookups in yucky places
09:14:02 <chris> I would pick one _type_ of module at a time.
09:14:26 <prock> the "modules" (as they are now) that should in fact be libraries, I'll just make sure they get initialized properly, remove the code that actually makes them modules and (continue) to link them at build time as libraries.
09:14:51 <prock> yes, one at a time for sure.
09:14:54 <chris> backends would probably be the easiest starting place.
09:16:24 <prock> It's my hope that this work may end up putting gnucash not far from having a plugin interface... If I fire some events off at appropriate places and properly document it.
09:16:58 <jsled> What problem would that solve?
09:17:09 <jsled> (having a plugin interface)
09:19:12 <jsled> gncbot: tell warlord I don't seem to have the 'userrights' permission, so I can't get to <http://wiki.gnucash.org/wiki/Special:Specialpages> to remove the 'move' permission as per <http://meta.wikimedia.org/wiki/Help:User_levels>.
09:19:12 <gncbot> jsled: The operation succeeded.
09:19:33 <prock> ex - file compatibility could be via modules. If GnuCash 1.8 wanted to open a GnuCash 2.2 xml file, it could just get the plugins and (upon file opening) migrate the 2.2 xml file down to 1.8 one module at a time.
09:20:16 <jsled> But then, wouldn't it be gnucash 2.2?
09:20:27 <jsled> I mean, if you can just "get the plugins", then you can compile the software.
09:22:33 <prock> so that I understand your case against modules, is it mostly represented here? http://lists.gnucash.org/logs/2007-02-15.html#T13:21:13
09:22:57 *** motin has joined #gnucash
09:23:15 <prock> are you against a plugin interface because it would be a plugin interface, or because you don't feel that gnucash is using it correctly?
09:25:49 <jsled> prock: let me drink a bit more coffee before answering. :)
09:26:17 <prock> in any event, gnucash right now is what it is. I think it should be fixed so that it does what it's trying to do correctly. Whether or not to use plugins is the primary dever's decision which I don't think actually needs to happen now.
09:27:17 <jsled> Certainly, as warlord points out, we do make use of runtime modules.
09:27:21 <prock> jsled: don't feel like you need to convince me of anything. I _really_ don't want to be trying to force ideas onto yourself or the other primary devers. I just am trying to fix what is agreed to be broken.
09:28:03 <jsled> prock: it's all good, we're just talkin' here. :)
09:28:06 <cstim> jsled: right, you're not a Bureaucrat in the wiki. Actually, only warlord is.
09:28:21 <jsled> I don't think that's the same as a "plugin interface", but I might (continue to) be confused.
09:28:53 <jsled> I don't want to build out a bunch of abstract indirection into the software in the hopes that if we build it, they will come.
09:29:27 <prock> ok, then perhaps I should avoid the word 'plugin'. But I continue to see an events (for which modules register) driven system the way to accomplish my present task.
09:29:43 <chris> See, I think of backends and import/export as "plugins", and the rest of the run-time "modules" I don't see much value in.
09:29:52 <cstim> @tell warlord In the wiki, can you give Bureaucrat permissions to {jsled,cstim} please?
09:29:52 <gncbot> cstim: The operation succeeded.
09:30:00 <jsled> And -- even to some degree -- I think we should be in-elegant in solving the first instances of modularity, rather than jumping to the modules system.
09:30:17 <jsled> prock: really? How so?
09:30:45 <prock> how so is an event driven system good?
09:31:14 <jsled> Maybe it's just because I don't have experience with highly-runtime-modular systems, so I don't know if there are well-worn patterns for doing it.
09:32:12 <chris> prock: I think "event" may be an already overloaded term. You're just talking about modules that offer up their functions as callbacks, right?
09:32:20 <prock> to be clear, the libraries should just be treated as such. The 'event system' I'm bringing up would only apply to modules.
09:32:40 <prock> chris: yes. but given the event based structure in gnucash now I thought it would be appropriate.
09:33:52 <chris> prock: you probably shouldn't be modeling anything after GnuCash's event system. It might just get replaced with GSignals.
09:33:53 <jsled> prock: event-driven systems are good because they're conceptually clear, loosely coupled and place the responsibility for handling events with the appropriate party.
09:33:56 <prock> jsled: my desire with events (or whatever they're called) as an approach to modules is mostly to avoid doing lots of symbol lookups at run time.
09:34:12 <jsled> (ah, self-referential definition ...)
09:34:21 <jsled> s/handling events/handling changes/
09:34:44 <jsled> Oh, maybe we have different definitions of "events", then.
09:34:54 <jsled> I'm not at all sure what you just said. :)
09:35:19 <prock> yes! mission accomplish! ;)
09:35:44 <chris> prock: you can't avoid symbol lookups at "module" init-time (which is run-time).
09:35:52 <jsled> heh.
09:36:42 <prock> chris: yes, looking up "module_init" and "module_destroy" and such is to be expected and is fine.
09:37:06 <chris> jsled: I think prock just means the "modules" should register callbacks upon initialization.
09:37:42 <jsled> I see. Hooks and signals and event-handlers oh my. :)
09:38:01 <prock> but if a module is to say at init "I have this function and this function and this function..."... I don't like that.
09:38:35 <chris> prock: yeah, that's where the "designing the interface" comes in.
09:38:37 <prock> in that case why not just make it a library?
09:38:44 <chris> prock: precisely.
09:39:56 <prock> for future reference what would you consider to be an appropriate term to refer to this as? A "callback-driven interface"?
09:40:12 <prock> (biab)
09:41:04 <cstim> "callback-registering interface"?
09:42:09 <chris> hey, look. I already converted the backend from GncModule to GModule. I forgot about that!
09:42:41 <cstim> which changeset?
09:46:52 <chris> http://svn.gnucash.org/trac/changeset/14198#file16
09:49:07 <chris> actually, looking more closely, perhaps they already were GModules and I just made it more obvious.
09:51:35 <chris> and IIRC, that's what enabled warlord to do this: http://svn.gnucash.org/trac/changeset/14296
09:52:17 <chris> which was actaully probably prompted by something that my first change subtly broke.
09:53:21 <chris> anyway, the backends at least do follow that pattern of registering callback upon init.
09:53:22 <cstim> that, and by the various linker problems that showed up on windows only.
09:54:36 <chris> oh that could be. Wow. I didn't realize this win32 effort started more than 9 months ago!
09:55:21 <cstim> really?
09:56:15 <cstim> indeed. initial notes in the wiki were added on 2006-03-03.
09:56:24 <cstim> Almost a year by now.
09:57:04 <cstim> That explains how I've spent my last year :-/
09:57:43 <chris> and it seems that despite being not yet complete, we already have some happy win32 users.
09:58:01 <prock> cstim: sounds fine to me.
09:59:20 <cstim> prock: except that I should have better spent this year on completing my PhD thesis.
10:00:10 *** hampton|away is now known as hampton
10:00:23 <chris> wb, hampton
10:00:27 <hampton> cstim: you can't integrate financial accounting software into your thesis somehow?
10:00:34 <jsled> heh
10:00:41 <hampton> hi chris. hi all.
10:01:30 <jsled> "a 'double-entry' method of RF communication between independent devices", or something? :)
10:02:16 <cstim> heh.
10:02:21 <prock> cstim: .. I actually meant "callback-registering interface" sounds fine. But I must say that gnucash sounds more fun (and fine) than writing a thesis.
10:16:03 *** IanL has quit IRC
10:25:31 *** warlord has joined #gnucash
10:25:32 *** gncbot sets mode: +o warlord
10:29:21 *** motin has quit IRC
10:31:47 <cstim> hi warlord
10:32:05 <warlord> Good morning
10:32:05 <gncbot> warlord: Sent 1 hour and 12 minutes ago: <jsled> I don't seem to have the 'userrights' permission, so I can't get to <http://wiki.gnucash.org/wiki/Special:Specialpages> to remove the 'move' permission as per <http://meta.wikimedia.org/wiki/Help:User_levels>.
10:32:06 <gncbot> warlord: Sent 1 hour and 2 minutes ago: <cstim> In the wiki, can you give Bureaucrat permissions to {jsled,cstim} please?
10:33:20 <jsled> warlord: don't know if you saw http://lists.gnucash.org/logs/2007-02-16.html#T03:05:06 or not. :(
10:33:25 <warlord> Done.
10:33:32 <warlord> jsled: No..
10:36:34 <cstim> warlord: have you been able to remove the "move page" permission for normal "editors"?
10:36:51 <warlord> I haven't tried.. Not sure how to do it.
10:37:24 <warlord> I think I need to do it in the actual configuration...
10:37:34 <cstim> yes, I think so. Please try to do it.
10:37:58 <cstim> The page vandalism itself is reverted in a matter of seconds, but the moves made it a bit more tricky.
10:38:05 <warlord> Okay..
10:38:19 *** motin has joined #gnucash
10:41:28 <cstim> I think I'll upgrade Andi5 to sysop in the wiki. Just in case he needs to block some spammers there :-)
10:44:57 <warlord> Sure.
10:47:14 <warlord> Ahh, I see how to disable it..
10:49:33 <warlord> I need to restart apache..
10:49:50 <warlord> objections?
10:50:57 <warlord> done.
10:51:05 <warlord> users should no longer be able to move pages.
10:53:32 * warlord ponders a nightly database dump for backup purposes..
10:53:43 <cstim> warlord: thank you very much
10:53:49 <cstim> yes, nightly database backup would be good.
10:53:54 *** andi5 has joined #gnucash
10:53:54 *** gncbot sets mode: +o andi5
10:54:21 <warlord> Not that would've helped here, because it was done before the backup.
10:54:36 <warlord> (it was done at 1am, the backup from cron.daily would've been at 4)
10:57:32 <cstim> yes. Reverting all the spam was actually easier than I had feared, once you've figured out which "rollback" or "revert" links can be used.
11:00:02 <warlord> *nods*
11:00:03 <warlord> Good..
11:00:08 <warlord> Glad it wasn't too bad.
11:00:24 <warlord> We should at least have an email address for that user...
11:00:29 * hampton needs to restart his nightly backup of svn. (It got lost in the move.)
11:00:40 <cstim> warlord: don't bother.
11:02:15 <warlord> Oh... we didn't block that!! I think that it DIDN'T require you to login in order to rename pages!!!
11:02:47 <warlord> He didn't have an actual editor account!
11:05:53 <warlord> He claimed to be Bobby Gannuer <bganu@spam.la>
11:06:09 <warlord> (he did have an account, the lookup is case-sensitive)
11:08:47 <warlord> jsled: do you happen to remember to mysql root password?
11:08:57 <warlord> (this is a binary question)
11:09:18 <jsled> warlord: (of course :) no, I don't believe I do.
11:09:34 <warlord> DAMN. Neither do I.
11:09:53 * jsled checks my passwd file...
11:10:19 <andi5> chris: ping
11:10:53 <warlord> Aha! I remember it now.
11:11:05 <chris> andi5: pong
11:11:15 <jsled> warlord: Ah, good. I was going to say that I don't recall knowing it, either. :)
11:11:19 <warlord> Although I dont need to root password, I need the wiki password, which is in the wiki config file.
11:11:21 *** esodan has joined #gnucash
11:12:13 <cstim> warlord: no, I don't think the spammer was working *without* account.
11:12:35 <warlord> cstim: no, he had an account.
11:12:54 <cstim> warlord: he edited pages just normal and he's also listed as (editor)
11:12:58 <andi5> chris: i need your advice as swig master :) .... for f::q i want to wrap g_spawn_async_with_pipes and the enum GSpawnFlag .... does that sound reasonable? i also wonder how to wrap the enum, ... *looking at SET_ENUM in engine.i* ...
11:13:12 <cstim> warlord: username "Bganu", he's 75th in the long user list.
11:13:35 <warlord> Yeah.. I know. I found that from the Recent Changes.
11:13:47 <cstim> warlord: so I think we set up our permissions just right, and this stupid kid happened to be the first one to test how much damage he is allowed to do.
11:13:59 <andi5> was a long night for him... maybe he met his three friends john, jack and jim?
11:14:18 <cstim> Now that page moving is disabled, this shouldn't be any similarly large problem anymore.
11:14:25 <warlord> *nods*
11:15:41 <andi5> chris: i could probably just %include <glib/gspawn.h> .... that may be the dirty way
11:16:58 <cstim> Everyone: I think I'll quit IRC for now and the coming weeks. Takes too much time and I have to concentrate on PhD work right now. So if you want to reach me, please resort back to mailing lists and email, but no longer here in IRC.
11:17:09 <chris> andi5: SET_ENUM is kind of a hack to treat enums as guile symbols instead of thunks which is the default.
11:17:16 *** cstim is now known as cstim_leaving
11:17:36 <chris> cstim: bis dann.
11:18:05 *** Demitar_ has quit IRC
11:18:40 <andi5> cstim_leaving: can you give us (ml) a short overview of the current aqbanking&win32 status? ... (maybe i just missed it, *feels flooded by emails*)
11:18:43 <chris> andi5: newer swigs actually have a feature flag to do this automatically.
11:18:47 <jsled> cstim_leaving: sorry to hear it, but well understood. Good luck with your work, and see you back soon (and in the other media, of course).
11:19:59 <hampton> cstim_leaving: Good luck with the thesis.
11:20:11 <chris> andi5: In my local repo I actually don't use SET_ENUM, because I have newer swig.
11:20:19 <cstim_leaving> andi5: I think I'll email this weekend.
11:20:20 <cstim_leaving> see ya
11:20:26 <andi5> cstim_leaving: thanks, see you
11:20:29 *** cstim_leaving has quit IRC
11:22:47 <andi5> chris: but there is no way to avoid duplicating the enum definition in (say) core-utils.i or %including the whole header file? ... and, what do you use in your local repository? (i have never wrapped anything and generally only like things that eatable)
11:24:15 <chris> andi5: there is base-typemaps.i
11:24:45 <chris> andi5: Instead of all those SET_ENUMs I use: %feature("constasvar", "1");
11:27:01 <chris> andi5: If there are just a few C callers with similar arguments, consider creating a C wrapper that's simpler, and then just swig-wrapping that.
11:27:15 <chris> andi5: maybe you can hide all those enum in C.
11:27:49 <andi5> actually that was my master-plan... but someone convinced me to wrap directly :)
11:28:45 <warlord> andi5: Yeah, I think you should just write a gnc wrapper in C, wrap that, and then use the wrapper..
11:35:57 *** Demitar has quit IRC
11:39:59 *** esodan has left #gnucash
11:54:36 *** franz has joined #gnucash
12:11:45 *** exper has quit IRC
12:36:08 *** wizkid238 has quit IRC
12:36:12 *** wizkid238 has joined #gnucash
12:43:33 *** ErKa has quit IRC
12:46:35 *** |gunni| has joined #gnucash
12:46:51 *** |gunni| has joined #gnucash
12:47:42 *** jpeach has joined #gnucash
12:48:28 *** jpeach has left #gnucash
12:52:24 *** wizkid238 has joined #gnucash
13:12:33 *** andi5 has left #gnucash
13:13:35 *** |gunni| has joined #gnucash
13:20:46 *** MrN has joined #gnucash
13:21:07 <MrN> hi
13:21:19 *** egsavage has quit IRC
13:36:30 *** egsavage has joined #gnucash
13:49:23 *** benoitg has joined #gnucash
13:49:37 *** benoitg has left #gnucash
13:50:32 *** sjc has joined #gnucash
15:18:27 *** warlord has quit IRC
16:11:46 <bistra> I am having trouble installing gnucash-2.0.4, in Fedora Core 6 (with all updates). GNUCASH ./configure looks for the XML::Parser in perl and fails to find it.
16:12:09 <bistra> Perl is installed on the system
16:12:35 <jsled> bistra: the error is reporting the perl module XML::Parser is not installed.
16:12:59 <jsled> bistra: why aren't you using the RPM?
16:13:21 <bistra> jsled: did not find it in the download page
16:13:42 <jsled> bistra: it should be available from Fedora directly...
16:14:28 <jsled> http://lwn.net/Articles/217638/
16:14:54 <bistra> ok: I'll try that. tks BY THE WAY I fixed the problem I reported whereby I suddenly I could not print invoices from 1.8.11, simply by opening the relevant xac file. tks for your help
16:41:09 *** rlaager has quit IRC
16:57:41 *** egsavage has quit IRC
17:06:57 *** twunder has quit IRC
17:09:29 *** motin_ has joined #gnucash
17:39:51 *** motin_ has quit IRC
17:39:55 *** motin has quit IRC
17:53:07 *** twunder has joined #gnucash
18:27:40 *** minDscrm has quit IRC
18:34:35 *** bistra has quit IRC
18:53:50 *** franz has quit IRC
18:54:16 *** warlord has joined #gnucash
18:54:17 *** gncbot sets mode: +o warlord
18:56:38 *** warlord is now known as warlord-afk
19:04:54 *** andi5 has joined #gnucash
19:04:55 *** gncbot sets mode: +o andi5
19:12:53 *** egsavage has joined #gnucash
19:17:41 *** minDscrm has joined #gnucash
20:06:22 *** exper has joined #gnucash
20:21:44 *** andi5 has quit IRC
21:02:56 <MrN> n8
21:03:42 *** MrN has quit IRC
21:29:01 *** sjc has quit IRC
21:35:07 *** prock has quit IRC
21:52:08 *** dbr has joined #gnucash
21:58:22 <dbr> egsavage: I tried printing in windows via the Adobe PDF printer driver, and it worked. Looks like the free versions are missing something -- maybe it's an oddity with gnome fonts.
21:58:53 <egsavage> Hmm, interesting
21:59:27 <egsavage> thx for the info though. I still have that "Create a PDF" option that saves me for now
23:02:01 *** egsavage has quit IRC