2014-09-19 GnuCash IRC logs
00:03:42 *** lwells has joined #gnucash
00:12:10 *** linas__ has quit IRC
00:12:23 *** linas__ has joined #gnucash
00:13:25 *** linas_ has joined #gnucash
00:13:26 *** linas__ has quit IRC
01:00:54 *** MechtiIde has joined #gnucash
01:03:44 *** comcry has joined #gnucash
01:11:10 *** O01eg has quit IRC
01:16:38 <lwells> Has anyone been sucessful importing quicken transactions from CVS file?
01:21:49 <lwells> I am getting the error that the withdrawal and the deposit column could not be understood
01:40:19 *** MechtiIde has quit IRC
01:44:56 <lwells> Well I fixed the amount columns by first making sure there was a zero in the column with no amounts. I cannot seem to figure out how to get the accounts(categories) to match up. When it imports it just places the Imbalance account
02:18:49 *** lwells has quit IRC
02:42:03 *** gjanssens has joined #gnucash
02:42:04 *** gncbot sets mode: +o gjanssens
03:42:14 *** Gbarr has quit IRC
04:44:04 <mikee> If lwells had hung around long enough I'd have said that this issue should have been fixed with Bug 721196.
05:16:54 *** GabrieleV_ has joined #gnucash
05:16:56 *** GabrieleV has quit IRC
05:16:56 *** GabrieleV_ is now known as GabrieleV
06:11:25 *** comcry has quit IRC
06:11:56 *** comcry has joined #gnucash
06:12:24 *** fell_ has joined #gnucash
06:12:25 *** gncbot sets mode: +o fell_
06:48:15 <gjanssens> So much for a nice idea... strfmon is not available on Windows/mingw :(
06:49:40 <gjanssens> My implementation was great though. It saved 170 lines of code, eliminating lots of code duplication in the process...
06:50:10 <gjanssens> Thank you Microsoft...
06:56:18 <Simon> someone must have implemented that function for Windows in a library somewhere...
07:11:23 *** Jimraehl1 has left #gnucash
07:16:07 *** Jimraehl1 has joined #gnucash
07:39:09 *** agaric has joined #gnucash
07:54:58 <gjanssens> Simon: mingw devs confirmed it's not in mingw...
07:57:19 *** t_ask has joined #gnucash
07:57:30 <t_ask> Hi,
07:58:44 <t_ask> I try to connect to my bank's fints server, but GNUcash shows the wrong Hash. Any idea why? I asked my Bank and the confirmed that they have another Hash code.
07:59:16 <t_ask> The certificate is shown ing it's "valid" while the Hash is wrong
07:59:56 <t_ask> URL looks good too.
08:00:57 <gjanssens> re strfmon - According to https://www.gnu.org/software/gnulib/manual/html_node/strfmon.html
08:01:15 <gjanssens> it's missing on several platforms: NetBSD 3.0, OpenBSD 3.8, Minix 3.1.8, IRIX 5.3, Cygwin 1.7.1, mingw, MSVC 9, BeOS
08:01:47 <gjanssens> While we don't actively support the *BSD's, I know GC compiles on it
08:02:03 <gjanssens> So it would be asking for more portability trouble
08:02:35 <gjanssens> I'll just drop the idea and go ponder another one...
08:02:58 <gjanssens> t_ask: I'm afraid I can't help you. I don't use the online banking parts of GnuCash
08:03:15 <gjanssens> If you don't get an answer here, you may want to ask on the gnucash-user mailing list as well
08:05:46 <t_ask> Maybe GnuCash doesn't show the SHA-1 hash of my bank's certificate?!
08:06:30 <t_ask> Who knows, it's weird, while all athoer cert values are looking fine
08:28:03 *** ErKa has joined #gnucash
08:50:34 *** t_ask has quit IRC
09:02:53 <warlord> gjanssens: how about writing an implementation for windows? That way the rest of the code is clean?
09:05:15 <gjanssens> warlord: that's what the mingw devs also proposed. No idea how hard that would be though.
09:05:46 <gjanssens> And then there are the other platforms like *BSD that don't support this function either
09:06:03 <gjanssens> And then I started wondering again what would be least effort
09:06:07 <warlord> Sure, but you can use confgure to test to see if it exists, and include our own impl if it does not.
09:06:44 <gjanssens> In the end we want to come to a locale independent save format
09:07:10 *** wol has joined #gnucash
09:07:11 <gjanssens> strfmon would have been a short-term fix until the rewrite
09:07:42 <gjanssens> But it looks like getting strfmon running may take as much effort as fixing the save format itself
09:08:08 <gjanssens> So I may prefer to focus on the latter instead.
09:08:57 <gjanssens> Unless I can pull of the locale trick I briefly suggested yesterday.
09:09:43 <gjanssens> The difference would be a Russian user being able to use the loan calculator starting in October vs
09:10:02 <gjanssens> a more profound solution for everybody in a couple of years.
09:10:16 <gjanssens> That's my trade-off
09:11:49 *** Gbarr has joined #gnucash
09:12:36 <gjanssens> Unrelated question: is there a maintenance page in mediawiki that lists broken links ?
09:25:24 <warlord> define "broken links"? Page references without a backing page?
09:36:22 *** lmat has joined #gnucash
09:39:49 <warlord> gjanssens: I see the inverse report, but I dont see what you're looking for.
09:40:38 <gjanssens> warlord: page references without a backing page indeed is what I'm looking for
09:41:10 <gjanssens> I have just cleaned up the formatting of the questions in the faq (to get a uniform looking index)
09:41:33 <gjanssens> But I want to check if this change causes links that no longer match in the rest of the wiki
09:41:37 <warlord> Where do you see that? I'm on Special Pages and don't see that report.
09:41:43 *** agaric has quit IRC
09:42:40 <gjanssens> I don't see it either. That why I was asking if it exists
09:49:38 <warlord> gjanssens: http://wiki.gnucash.org/wiki/Special:WantedPages
09:54:58 <gjanssens> warlord: lovely, thanks
09:57:14 <gjanssens> So my change didn't introduce any wanted pages and the faq index no longer hurts my eyes. Double yay.
10:04:52 <warlord> :)
10:20:06 *** agaric has joined #gnucash
10:42:56 *** GabrieleV_ has joined #gnucash
10:43:01 *** GabrieleV has quit IRC
10:43:01 *** GabrieleV_ is now known as GabrieleV
10:50:15 *** wol has quit IRC
11:10:59 *** bhardwajs has joined #gnucash
11:20:12 *** O01eg has joined #gnucash
11:20:43 *** bhardwajs1 has joined #gnucash
11:27:00 *** bhardwajs has quit IRC
11:32:48 <lmat> Currently, the configuration is set in some place accessible to a daemon running on that computer, right?
11:33:22 <lmat> (Microsoft Windows Registry, and something similar (but not gconf? dconf?) on GNU/Linux and others, etc.)
11:33:39 <lmat> Why don't we use a configuration file?
11:43:43 *** linas_ has quit IRC
11:44:02 *** linas_ has joined #gnucash
11:52:28 <gjanssens> lmat: in an effort to be nice to the host OS ?
11:52:37 <gjanssens> lmat: Allow users to leverage their finger memory ?
11:52:51 <gjanssens> lmat: Because writing a config file parser has its own challenges and the excercise has been done already in gsettings ?
11:53:01 <lmat> gjanssens: I follow the last one :)
11:53:32 <gjanssens> lmat: all moot btw, because if we move away from glib, inevitably we also have to move away from gsettings
11:53:56 <gjanssens> perhaps boost has a nice alternative
11:54:52 <lmat> gjanssens: I vote for a configuration file :D
11:56:19 <gjanssens> lmat: why ?
11:56:44 <lmat> gjanssens: Because they're easier to edit (remotely, without an host, by script, etc.)
11:56:55 <lmat> gjanssens: And easier to version control. ... possible to version control.
11:57:16 <lmat> gjanssens: Because there is more than one tool for editing files (as opposed to the Microsoft Windows Registry, for instance).
11:58:14 <gjanssens> lmat: by not using the system's own preference infrastructure you take away to possibilities that come with the system for integration
11:58:16 <lmat> gjanssens: I would think it's easier to help users, too. Tell them "add 'setting=value' to the file at ..." ?
11:58:53 <gjanssens> lmat: your assumption there's only one tool to edit the Windows registry is false by the way
11:59:11 <lmat> gjanssens: That's good to know, but surely the point is a good one?
11:59:40 <gjanssens> lmat: not from my point of view
11:59:58 <gjanssens> If you're on Windows you want to be able to use windows tools to edit preferences
11:59:59 <lmat> gjanssens: Can you edit it from a separate host? Can you edit it if you only have an hard drive/file share?
12:00:31 <lmat> gjanssens: I'm not familiar with the Microsoft Windows user experience...what tools are you talking about?
12:01:48 <lmat> gjanssens: I have to head out. I don't have much time until my next meeting, and I need to get home and eat lunch. I'll read when I get back :)
12:01:49 <gjanssens> you can from a separate host IIRC: quick google gives one example http://support.microsoft.com/kb/934958
12:03:56 <gjanssens> As for other tools: our own gnucash installer adds registry keys
12:05:16 <gjanssens> guestfish allows you to edit registry entries on a harddrive
12:05:42 <gjanssens> then I found the hivex library, which allows you to edit the windows registry files directly
12:05:55 <gjanssens> it has ruby bindings and is available on linux
12:06:16 <gjanssens> Not a comprehensive list at all, but just to illustrate regedit is not the only option
12:08:19 <gjanssens> And I'm not necessarily *against* a configuration file either
12:08:30 <gjanssens> Surely for just gnucash-the-desktop-application, it doesn't really matter where the preferences are stored
12:08:41 <gjanssens> I'm thinking further than that.
12:08:59 <warlord> gjanssens: boost has an INI file parser
12:09:57 <gjanssens> What if we eventually are at a point that the gnucash engine can be used for more fancy things like a web service
12:10:19 <gjanssens> or someone wants to integrate the gnucash engine in a windows automated workflow ?
12:11:01 <gjanssens> In that case good system integration is useful
12:11:16 <warlord> I care less about windows ;)
12:11:24 <warlord> (I care more about Android or iPhone)
12:11:36 <gjanssens> (I could just as well have suggested integration in a gnome environment, in which case dconf would have been the central integration point)
12:12:10 <gjanssens> warlord: Actually so do I :)
12:13:08 <gjanssens> Anyway I am not strongly defending a specific point of view here
12:13:24 <gjanssens> Just bringing another view to the table to evaluate all options
12:13:46 <gjanssens> And in general I like good integration into a native environment
12:14:16 <gjanssens> Whether the way to store preferences really matters in that case, I don't know
12:14:54 *** wol has joined #gnucash
12:15:15 <gjanssens> GnuCash is not really strong in that area currently, largely due to our heavy reliance on gtk/glib no doubt
12:32:46 *** bhardwajs1 has quit IRC
12:37:27 *** MechtiIde has joined #gnucash
12:38:30 *** jralls has quit IRC
12:39:03 *** jralls has joined #gnucash
12:39:04 *** gncbot sets mode: +o jralls
12:48:42 <lmat> gjanssens: Thanks for that!
12:48:45 *** ErKa has quit IRC
12:49:55 <lmat> gjanssens: You did forget to talk about version control.
12:52:08 <lmat> gjanssens: and common tasks like searching for the word "start" among all the settings.
12:52:31 <lmat> gjanssens: Another question is: what settings are configureable anyway? Is it only UI settings, or backend settings, too?
12:57:42 <gjanssens> lmat: version control is an interesting point.
12:58:18 <gjanssens> I don't manage windows pc's so I have no idea how a sysadmin would go about managing settings in version control on Windows
12:59:13 <gjanssens> lmat: re searching: you can do so in regedit, you can do so with the dconf tool on dconf based systems, so I don't see an issue there
12:59:29 <jralls> gjanssens: No need to write strfmon.c. Here's the FreeBSD/Darwin version: http://www.opensource.apple.com/source/Libc/Libc-320/stdlib/FreeBSD/strfmon.c. Just stick it in lib/ and adjust the build to use the system version if it's there and the "borrowed" one if it isn't.
13:00:11 <gjanssens> lmat the last question is more interesting
13:00:22 <gjanssens> And hard to answer in one sentence
13:00:47 <gjanssens> It touches on the layered structure of our settings
13:01:06 <gjanssens> Some are user level, others are file level
13:02:10 <gjanssens> jralls: cool, thanks
13:02:37 <lmat> gjanssens: Okay, so it's straightforward to have several sets of settings and change them out? (comparable to renaming settings files? ^_^)
13:03:02 <gjanssens> I was already reworking the whole assistant-loan code to work around the missing strfmon.c
13:03:21 <gjanssens> Borrowing it is probably easier :)
13:03:25 <jralls> lmat: If we want to get into iOS we have to use defaults. There's no way to edit an INI file on an iOS system.
13:05:32 *** bhardwajs has joined #gnucash
13:06:55 <lmat> jralls: ouch
13:10:24 <jralls> Rather, the only way to adjust settings from outside of GnuCash is to use defaults, which integrates with a panel in the Settings app. But if we don't do it that way we're likely not to get accepted by the i
13:10:38 <jralls> darn. iTunes store.
13:13:34 *** gjanssens is now known as gjanssens_
13:18:58 <lmat> nice
13:19:20 <lmat> gnucash should be able to modify its own configuration, right...is that allowed?
13:20:12 <warlord> it is, but not "expected" on iOS
13:29:58 *** ErKa has joined #gnucash
13:37:43 *** bhardwajs has quit IRC
13:41:43 *** agaric has quit IRC
13:51:10 *** agaric has joined #gnucash
13:52:39 *** bhardwajs has joined #gnucash
14:03:53 *** bhardwajs1 has joined #gnucash
14:11:43 *** bhardwajs has quit IRC
14:27:18 *** wol has quit IRC
14:40:05 *** wol has joined #gnucash
14:40:17 *** wol has quit IRC
14:51:23 *** mikee_ has joined #gnucash
14:59:02 *** mikee has quit IRC
14:59:02 *** mikee_ is now known as mikee
15:08:11 *** bhardwajs1 has quit IRC
15:09:38 *** bhardwajs has joined #gnucash
15:17:39 *** bhardwajs1 has joined #gnucash
15:17:53 *** wol has joined #gnucash
15:24:10 *** bhardwajs has quit IRC
15:29:45 *** wol has quit IRC
15:29:47 *** wol has joined #gnucash
15:39:25 *** gjanssens_ has quit IRC
15:44:22 *** linas__ has joined #gnucash
15:44:23 *** linas_ has quit IRC
16:00:08 *** bhardwajs has joined #gnucash
16:05:19 *** bhardwajs1 has quit IRC
16:19:58 *** MechtiIde has quit IRC
16:24:38 *** bhardwajs1 has joined #gnucash
16:31:12 *** bhardwajs has quit IRC
16:31:28 *** bhardwajs1 has quit IRC
16:34:02 *** agaric has quit IRC
16:59:21 *** bhardwajs has joined #gnucash
17:07:01 *** ErKa has quit IRC
17:13:45 *** bhardwajs has quit IRC
17:19:25 *** bhardwajs has joined #gnucash
17:24:52 *** wol has quit IRC
17:31:10 *** grantbow has joined #gnucash
17:35:02 <lmat> Do the unit tests ensure that allocated values are released?
17:35:21 <lmat> I guess I could check that easily enough ^_^ But do you know of any checks like that that exist?
17:40:23 *** agaric has joined #gnucash
17:46:32 <jralls> lmat: I'm confident that utest-(Account|Transaction|Split) and the tests in libqof/qof do. Not so much about any of the rest.
17:46:49 <lmat> jralls: okay. Why those? Is it written specially? (nothing automatic?)
17:47:16 *** lwells has joined #gnucash
17:47:38 <jralls> lmat: Those because either I wrote them or Muslim Choclov did under my direction in GSoC 2011.
17:48:36 <lwells> I am trying to import transactions from the Android app, but it is not pulling in the expense properly. I have doing the ofx export from Android
17:48:47 <jralls> And by 'ensure' I mean that they make sure that they release any resources that they acquire, not that they check for proper release of resources that might be acquired int the code under test. The latter is next to impossible for a unit test to check.
17:49:08 <jralls> That's why there are leak checkers like Valgrind.
17:49:25 *** bhardwajs has quit IRC
17:49:47 <lmat> jralls: They make sure by careful coding? Or by some other method?
17:51:28 <lwells> Funny this is, the qif format does work, pulls in everything propely
17:51:32 <lmat> jralls: What would be satisfactory for me would be for kvp_value_new_frame (for instance) to keep track of everything it hands out and kvp_value_delete to keep track of everything coming in. At a certain point, the code can say "is anything outstanding?"
17:51:41 <lmat> jralls: I'm asking if there's anything like that.
17:52:16 <lmat> jralls: Alternatively (I know this is possible with c++, not so sure with glib, stdlib, etc.), one can override operator new and operator delete to conduct this logging...but that's somewhat tricky.
17:53:30 <lwells> I imported the accounts into the android app from the gnucash desktop, so that should match up properly right?
17:56:27 <jralls> lmat: Read up on "RAII". Once you're in C++ the constructors and destructors should take care of resource allocation and deallocation. For checking GObjects, look at test_object_checked_destroy() in src/test-core/unttest-support.c
17:56:41 <lmat> jralls: thanks
17:59:35 <jralls> For stuff allocated on the heap outside of constructors, allocate with std::make_shared or std::make_unique as much as possible. If you have to hand off a pointer for C to use, try to isolate it so that when the consumer is rewritten in C++ it can be done in a more leak-resistant way.
18:08:27 <lmat> jralls: later!
18:08:28 *** lmat has quit IRC
18:09:48 *** bhardwajs has joined #gnucash
18:54:15 *** bhardwajs1 has joined #gnucash
18:59:25 *** himaxx has joined #gnucash
19:02:13 *** himaxx has quit IRC
19:03:26 *** bhardwajs has quit IRC
19:48:59 *** jralls has quit IRC
19:50:01 *** jralls has joined #gnucash
19:50:01 *** gncbot sets mode: +o jralls
19:59:52 *** bhardwajs1 has quit IRC
20:07:20 *** bhardwajs has joined #gnucash
21:03:00 *** GabrieleV_ has joined #gnucash
21:03:07 *** GabrieleV has quit IRC
21:03:07 *** GabrieleV_ is now known as GabrieleV
21:29:20 *** ErKa has joined #gnucash
21:53:56 *** ErKa has quit IRC
21:58:01 *** GabrieleV_ has joined #gnucash
21:58:09 *** GabrieleV has quit IRC
21:58:09 *** GabrieleV_ is now known as GabrieleV
22:05:45 *** agaric has quit IRC
22:33:02 *** GabrieleV_ has joined #gnucash
22:33:04 *** GabrieleV has quit IRC
22:33:05 *** GabrieleV_ is now known as GabrieleV
22:36:15 *** bhardwajs has quit IRC
22:58:31 *** fell__ has joined #gnucash
22:58:33 *** gncbot sets mode: +o fell__
23:07:25 *** fell_ has quit IRC
23:32:09 *** vt-kuga has joined #gnucash
23:42:23 *** vt-kuga has quit IRC