2013-10-05 GnuCash IRC logs

03:37:52 *** fell_ has joined #gnucash
03:37:52 *** gncbot sets mode: +o fell_
04:10:00 *** john has quit IRC
04:24:16 *** DavidTurner has joined #gnucash
04:25:34 <DavidTurner> Hi all, I'm using GNUCash 2.4.13 on Windows 7 (64-bit) and recently it's stopped working - the splash screen appears and then it says 'gnucash.exe' has stopped working. Can anyone help?
04:26:40 <DavidTurner> I've tried moving all my data elsewhere, in case it was a corrupt file, but that didn't help; I've also tried un/reinstalling and that didn't help either. Nor did installing the latest unstable release.
04:28:27 <DavidTurner> The crash seems to be an access violation somewhere deep in a call stack in libguile, but I know almost no Scheme and have no development environment on this machine so can't dig any deeper
04:36:44 *** TradeBorG has joined #gnucash
04:53:15 *** aqua___ has joined #gnucash
05:04:19 *** aqua___ has quit IRC
05:44:26 *** gjanssens-afk has quit IRC
05:44:51 *** gjanssens-afk has joined #gnucash
05:45:05 *** gjanssens-afk is now known as gjanssens
05:45:48 *** gncbot sets mode: +o gjanssens
06:20:32 *** Simon has quit IRC
06:43:06 *** Simon has joined #gnucash
06:48:35 *** Simon has joined #gnucash
06:58:34 *** Jimraehl1 has left #gnucash
07:01:30 *** Jimraehl1 has joined #gnucash
07:51:19 *** DavidTurner has quit IRC
08:13:22 *** warlord has joined #gnucash
08:13:23 *** gncbot sets mode: +o warlord
08:37:55 *** ErKa has joined #gnucash
08:56:14 *** Bodhi-Baum has joined #gnucash
09:14:54 *** benoitg has joined #gnucash
09:28:59 *** nomeata has joined #gnucash
09:48:25 *** ErKa has quit IRC
09:56:51 *** Bodhi-Baum has quit IRC
10:03:49 *** Bodhi-Baum has joined #gnucash
10:59:38 *** kpreid has quit IRC
11:54:28 *** fuzzybunny69y has quit IRC
12:50:05 *** ErKa has joined #gnucash
13:05:16 *** Bodhi-Baum has quit IRC
13:51:08 *** john has joined #gnucash
13:51:09 *** gncbot sets mode: +o john
14:22:01 *** Ardonik has joined #gnucash
14:59:09 <gjanssens> John, I'm currently working on the settings conversion
14:59:20 <gjanssens> With a little luck I can get it done tomorrow.
14:59:31 <gjanssens> The xsl transform is almost done
14:59:34 <john> Good. That would be nice.
15:00:10 <gjanssens> I still need to write the C code to actually call the transform
15:00:19 <gjanssens> and some guile glue for it.
15:00:48 <gjanssens> Not *all* preferences can be migrated though, but the important ones are
15:01:01 <john> What can't be migrated?
15:01:11 <gjanssens> What will be left out is mostly saved window position and sizes
15:01:58 <gjanssens> And the warning dialogs cache (the "don't ask me again") questions
15:02:24 <gjanssens> And maybe I won't have time to do the custom check format parameters
15:02:29 <gjanssens> which bothers me most
15:02:59 <gjanssens> But it's fairly difficult transform, so that transform may have to wait for 2.5.7
15:03:51 <john> If it's a time issue, that's OK. This is still alpha, so good enough for 2.5.6. I would only be worried if it was a "too hard" issue.
15:04:18 <gjanssens> To get that done, I have to dabble with GVariant types from within guile
15:04:24 <gjanssens> I'm not sure that is easy though
15:05:22 <gjanssens> Or I write a few wrapper functions in C that take basic type parameters and do the conversion to gvariant instead
15:05:29 <gjanssens> That's probably easier
15:05:34 <john> I thought Scheme was untyped. Why can't you unwrap the GVariant/GValue on one side, pass it with a type-name and value through Guile, and wrap it back up on the other side?
15:06:28 <gjanssens> I'm having difficulties to grok that last sentence :(
15:06:34 <gjanssens> I'm rather tired right now
15:07:37 <gjanssens> Anyway, for this particular conversion, the xml file has strings in this form [13, 25.0]
15:07:53 <gjanssens> Being a pair of floats
15:07:59 <john> I was thinking that you had GVariants that needed to pass through Guile. But that's not true: You have strings that you get from XML; you just need to figure out what they mean (which I would think would be the guile part), and call the new setter with the result.
15:08:27 <gjanssens> That correct
15:08:52 <gjanssens> But there is no setter for this particular type (Gvariant type "(dd)")
15:09:26 <gjanssens> In our C code, I create the GVariant myself and then call g_settings_set with that variant
15:09:45 <gjanssens> So there I was thinking of writing another setter that takes two floats instead
15:09:59 <gjanssens> It will only be used by the conversion script
15:10:36 <gjanssens> But that should do it.
15:10:44 <gjanssens> Not sure if I can finish that part tomorrow.
15:11:10 <gjanssens> I'll first focus on more important parameters (like default currency and stuff)
15:11:30 <john> I don't understand why you can't just extract that "create the GVariant and call g_settings_set" into a separate function and call that from guile -- and from wherever it's needed in the regular C code.
15:12:05 <gjanssens> To prevent an explosion of gsettings
15:12:10 <gjanssens> wrapper functions
15:12:44 <john> An explosion of wrapper functions is better than 500-line-long switch statements.
15:13:29 <john> After all, in C++ they can be collapsed into *one* template!
15:15:31 <gjanssens> Ok
15:16:12 <gjanssens> I think there are 2 different GVariant uses in GnuCash
15:16:17 <gjanssens> "(dd)"
15:16:22 <gjanssens> "(iiii)"
15:16:46 <john> So "explosion" means *two*? ;-)
15:16:52 <gjanssens> I guess it's not too bad if I create wrappers for these two
15:16:55 <gjanssens> That two for now
15:17:04 <gjanssens> :)
15:17:30 <warlord> Why not treat GVariants in guile the same way we do KVPs in guile?
15:17:48 <gjanssens> How's that ?
15:18:24 <john> It's very soon going to be "we don't". KVPs are going to be private to the classes that own them.
15:19:03 <gjanssens> Anyway, I think writing two more wrapper functions is fairly easy
15:19:05 <warlord> Scheme itself is untyped; swig, however, isn't.. we wrap the kvp type with an internal switch to handle the different types.
15:19:30 <gjanssens> Setting up swig looks more complicated
15:19:48 <gjanssens> Especially for a code path that essentially runs only once per user
15:20:38 <gjanssens> The rest will be for tomorrow btw
15:20:42 <gjanssens> I have to go
15:20:44 <warlord> probably...
15:20:50 <warlord> good night, gjanssens
15:20:52 <john> G'night, Geert.
15:22:58 <gjanssens> bye
15:23:03 *** gjanssens is now known as gjanssens-afk
15:29:58 *** Ard0nik has joined #gnucash
15:33:29 *** Caesar has joined #gnucash
15:37:38 *** kpreid has joined #gnucash
15:38:13 *** Ardonik has quit IRC
15:41:34 *** Ard0nik has quit IRC
15:41:50 *** fuzzybunny69y has joined #gnucash
15:47:56 *** ErKa has quit IRC
16:52:39 *** gncbot has joined #gnucash
16:52:46 *** john sets mode: +o gncbot
16:52:48 <warlord> @op
16:52:48 *** gncbot sets mode: +o warlord
16:52:49 <warlord> thanks.
16:52:58 <warlord> (nice to have lots of ops around nowadays)
16:53:01 <john> denada.
16:54:17 <john> BTW, I'm writing a generic KvpValue <-> GValue conversion routine. I'm a bit horrified at the "binary" option for KvpValue, but it doesn't look like it's actually used. Do you know if it ever was?
16:55:00 <warlord> I dont think so.
16:55:07 <warlord> I think it was included for "completeness"
16:55:38 <john> That's one of Martin Fowler's "bad smells". ;-)
16:55:52 <warlord> lol
16:56:08 *** TradeBorG has quit IRC
17:00:19 <warlord> well, it was designed to be the extensible format, so they tried to anticipate any future usage. e.g. i can imagine using it for storing imges.
17:01:13 <john> Serialized as bin64, I suppose.
17:01:42 <warlord> I dont know, honestly.
17:22:26 *** fell_ has quit IRC
17:40:00 *** Jimraehl1 has quit IRC
18:03:43 *** ErKa has joined #gnucash
18:36:39 *** nomeata has quit IRC
18:42:05 *** blathijs has quit IRC
19:17:00 *** mikee has quit IRC
19:29:06 *** mikee has joined #gnucash
19:29:07 *** gncbot sets mode: +o mikee
19:33:48 *** talso has quit IRC
19:40:21 *** mikee has quit IRC
19:43:01 *** mikee has joined #gnucash
19:43:02 *** gncbot sets mode: +o mikee
19:49:45 *** mikee has quit IRC
19:50:23 *** mikee has joined #gnucash
19:50:24 *** gncbot sets mode: +o mikee
19:51:51 <warlord> john: Qt-MySQL ;)
19:52:23 <john> As a replacement for libdbi?
19:52:53 <warlord> potentially
19:53:13 <warlord> (I know nothing about qt, honestly, so I dont know how good an "abstraction" it is)
19:53:22 <john> That implies that we migrate to the Qt ecosystem for everything else as well.
19:53:30 <warlord> Previously there was someone touting GnomeDB, but I dont know what happened there
19:53:38 <warlord> john: yes, it does. ;-)
19:54:39 <john> I thought that the point of the exercise was to confine that sort of massive dependency to the GUI layer.
19:55:25 <john> Engine and backends should depend on only the STL and Boost as much as is practical.
19:56:28 *** mikee has quit IRC
19:57:07 *** mikee has joined #gnucash
19:57:07 *** gncbot sets mode: +o mikee
19:57:30 <warlord> Yes, but the SQL Backend could add some dependency, and frankly once you include qt you have all of qt.. Rarely do you have to make sure you include some other piddly library for some minor functionality.
19:57:47 <warlord> I.e., qt dosn't drown you in a sea of dylibs
19:58:10 <john> Well, it does, but they're all Qt's. ;-)
19:58:50 <john> GnomeDB seems alive and actively maintained FWIW.
19:59:58 <warlord> I think when we were first looking at it it wasn't "production" quality, yet.
20:00:33 *** talso has joined #gnucash
20:00:40 <john> Anyway, that's more Gnome, so we perhaps don't want to go there.
20:01:15 <warlord> It's too bad libdbi doesn't consider it a bug..
20:02:53 <john> Not so much that they don't consider it a bug; more that they only want to support FreeBSD.
20:04:04 <warlord> Eh, I know people who close bugs because they can't reproduce it on Linux. ;)
20:07:09 <john> Not really the point. If we're going to depend on a library we need its developers to support MinGW, OSX, and Linux. If they don't, we have to find a different library that does. That's one of the ongoing bitches about Gtk: The non-Linux support is spotty at best.
20:12:31 *** mikee has quit IRC
20:12:55 *** mikee has joined #gnucash
20:12:55 *** gncbot sets mode: +o mikee
20:17:07 <john> This looks good: http://soci.sourceforge.net/doc.html
20:17:28 <john> Depends only on Boost + the backend libraries.
20:17:40 <john> Plus, wxWidgets
20:18:14 <john> <sigh>. Plus, wxWidgets's Vadim Zeitlin is one of the maintainers.
20:23:35 *** mikee has quit IRC
20:35:27 <warlord> LOL. And here I thought you were saying that it dependended on boost, backend libraries, and wxwidgets. ;)
20:36:27 <john> No, though I do like wx better than Qt.
20:37:12 <john> I'm just familiar with Vadim's work on wx, and I know that it's really good, so that's a recommendation for SOCI.
20:38:15 <warlord> I've honestly never really looked at Wx
20:38:24 <warlord> is it C or C++?
20:39:55 <john> C++98. Very MFC-like, but uses as many native controls as it can, so it produces a much more native look-and-feel than Qt.
20:40:31 <john> The only downside would be that it wraps Gtk for Linux. :-/
20:41:31 <warlord> hmm
20:41:42 <john> Of course the best results are from writing a separate GUI for each platform, but only the largest teams like Firefox can pull that off.
20:42:05 <warlord> :)
21:09:27 *** ErKa has quit IRC
21:19:37 *** Jimraehl1 has joined #gnucash
21:29:33 *** fuzzybunny69y has quit IRC
21:44:12 *** Jimraehl1 has quit IRC
21:44:12 *** uXus has quit IRC
21:44:12 *** Sk1d has quit IRC
21:44:12 *** johey has quit IRC
21:44:12 *** stl has quit IRC
21:44:12 *** xnox has quit IRC
21:44:12 *** ggherdov has quit IRC
21:44:14 *** Jimraehl1 has joined #gnucash
21:44:14 *** uXus has joined #gnucash
21:44:14 *** Sk1d has joined #gnucash
21:44:14 *** ggherdov has joined #gnucash
21:44:14 *** stl has joined #gnucash
21:44:14 *** xnox has joined #gnucash
21:44:14 *** johey has joined #gnucash
21:51:08 *** Ardonik has joined #gnucash
21:56:51 *** kpreid has quit IRC
22:00:07 *** kpreid has joined #gnucash
22:19:41 *** Ard0nik has joined #gnucash
22:26:22 *** Ardonik has quit IRC
22:32:10 *** todd has quit IRC
22:35:11 *** todd has joined #gnucash
23:04:38 *** Ard0nik has quit IRC
23:04:39 *** Caesar has joined #gnucash
23:09:42 *** talso has quit IRC
23:14:59 *** talso has joined #gnucash