2017-12-10 GnuCash IRC logs
00:15:58 *** VigilantSky has quit IRC
00:32:12 <CDB-Away> jralls: for reference, the filed report: https://bugzilla.gnome.org/show_bug.cgi?id=791429
01:04:44 *** kael has quit IRC
01:07:02 *** frakturfreak has quit IRC
01:21:49 *** frakturfreak has joined #gnucash
01:36:13 *** fell has joined #gnucash
02:04:51 *** Mechtilde has joined #gnucash
02:27:52 *** yuriks has quit IRC
02:46:11 *** VigilantSky has joined #gnucash
02:59:22 *** yuriks has joined #gnucash
03:31:28 *** hoijui has joined #gnucash
03:44:55 *** fabior has joined #gnucash
04:10:20 *** fell has quit IRC
04:22:08 *** fiddlerwoaroof has quit IRC
04:27:26 *** fiddlerwoaroof has joined #gnucash
04:36:02 *** hoijui has quit IRC
04:48:23 *** VigilantSky has quit IRC
05:01:45 *** VigilantSky has joined #gnucash
05:22:58 *** fabior has quit IRC
05:30:52 *** fell has joined #gnucash
05:39:51 *** fabior has joined #gnucash
05:48:16 *** Mechtilde has quit IRC
06:04:10 *** gjanssens has joined #gnucash
06:04:10 *** ChanServ sets mode: +o gjanssens
06:12:39 *** Mechtilde has joined #gnucash
06:27:47 *** jotrago has quit IRC
06:30:14 *** finster has joined #gnucash
06:40:48 *** Mechtilde has quit IRC
06:42:44 *** User has joined #gnucash
06:52:13 *** fell has quit IRC
06:53:25 *** fell has joined #gnucash
06:59:29 *** User has quit IRC
07:03:34 *** chris has quit IRC
07:03:49 *** chris has joined #gnucash
07:04:14 *** VigilantSky has quit IRC
07:05:57 <gjanssens> .
07:11:35 *** Mechtilde has joined #gnucash
07:15:59 *** fabior has quit IRC
07:17:59 *** jotrago has joined #gnucash
07:39:37 *** jotrago1 has joined #gnucash
07:41:27 *** jotrago has quit IRC
07:41:28 *** jotrago1 is now known as jotrago
07:56:54 *** jotrago has quit IRC
08:27:43 *** VigilantSky has joined #gnucash
08:33:41 *** storyjesse has joined #gnucash
08:35:38 *** jotrago has joined #gnucash
08:40:21 *** jotrago has quit IRC
08:48:30 *** jotrago has joined #gnucash
08:51:46 *** fbruetting has joined #gnucash
08:57:39 *** O01eg has joined #gnucash
08:58:41 *** O01eg has joined #gnucash
09:24:39 *** fbruetting has quit IRC
09:32:35 *** fabior has joined #gnucash
09:35:13 *** jotrago has quit IRC
09:37:21 *** User has joined #gnucash
10:05:44 *** fbruetting has joined #gnucash
10:07:48 *** phebus has quit IRC
10:20:59 *** kael has joined #gnucash
10:30:36 *** VigilantSky has quit IRC
10:37:38 *** fbruetting has quit IRC
11:21:11 *** kael has quit IRC
11:36:06 *** Mechtilde has quit IRC
11:36:24 *** storyjesse has quit IRC
11:38:05 *** fell has quit IRC
11:53:54 <jralls> CDB-Away: Yeah, I saw it.
11:54:32 <jralls> lmat: You know about https://bugzilla.gnome.org/show_bug.cgi?id=787095, right?
11:57:04 <jralls> gjanssens: Well, 784123 has a work-around so it's not completely broken.
12:04:44 *** fabior has quit IRC
12:18:00 <gjanssens> jralls: ok. So users of 2.6.18 can apply the workaround and continue untill 2.8.0 gets released.
12:18:44 <gjanssens> And from your (limited) testing it appears 2.6.18 can even handle DATETIME fields although it's coded to use TIMESTAMPS ?
12:19:08 *** fbruetting has joined #gnucash
12:19:39 <gjanssens> So users could switch between 2.8.0 and 2.6.18 as things are at preset ?
12:20:10 <gjanssens> *present*
12:20:15 <jralls> Other way around, but yes. The reading and writing code doesn't know whether the field is a TIMESTAMP or a DATETIME, only the table creation code does.
12:21:01 <gjanssens> Then I propose to be conservative after all and not make additional changes in 2.6
12:22:27 <jralls> Both fields use the same ISO date string, it's just a more restricted date range in TIMESTAMP. The other difference is that DATETIME takes the string as-is while TIMESTAMP assumes that it's in local time and converts it to UTC.
12:23:29 <gjanssens> Hmm, so would that mean we could see timezone jumps when alternating between 2.6 and 2.8 ?
12:23:53 <gjanssens> And worst case of-by-one-day issues ?
12:24:20 <gjanssens> (argh) *off*-by-one-day
12:26:05 *** sbluhm has quit IRC
12:27:22 <jralls> It could certainly produce weird results in New Zealand, yes: It would change 10:59:00 to 22:59:00 the previous day.
12:27:54 <jralls> So that's something I need to look into.
12:29:30 *** Mechtilde has joined #gnucash
12:29:54 *** kael has joined #gnucash
12:32:12 *** sbluhm has joined #gnucash
13:01:05 <jralls> Meanwhile I'm working on https://bugzilla.gnome.org/show_bug.cgi?id=791422. The problem is setting GNC_DBD_DIR incorrectly in etc/gnucash/environment, but setting it in the actual environment is overridden by the incorrect value in etc/gnucash/environment. Isn't that supposed to be the other way around, i.e. the environment settings should override settings in the environment file?
13:05:35 <gjanssens> jralls: no, the environment file adds to the existing environment
13:08:59 <gjanssens> jralls: as of gnucash 2.7 you can override the the build computed values in etc/gnucash/environment with values set in etc/gnucash/environment.local
13:09:03 <markvandenborre> gjanssens: I just googled for "gnucash argenta"
13:09:05 <gjanssens> That file is never overwritten by gnucash
13:09:13 <markvandenborre> and the donations page comes up :-)
13:10:12 <gjanssens> markvandenborre: and I suppose you were looking for something else... :)
13:10:52 <jralls> gjanssens: But both override what the user specifies in the shell. I think that's counterintuitive for most experienced Unix users. It also breaks things like desktop files.
13:11:23 <markvandenborre> gjanssens: correct
13:11:42 <markvandenborre> it's probably trivial, so I'll start searching first before bothering you with it
13:11:47 <markvandenborre> but it was quite funny to notice
13:11:53 <gjanssens> jralls: IIRC GNC_DBD_DIR is an optional parameter that should only be added to the environment file if it's not system-default, which means when not overridden at build time
13:12:30 <gjanssens> Perhaps this part got lost in the conversion to cmake
13:14:49 <gjanssens> Other than that, it may be useful to add syntax to prevent overwriting existing environment variables.
13:14:53 <jralls> The logic is screwed up in cmake, it sets it twice, the second time unconditionally and wrong if the (also broken) logic to find the directory fails.
13:15:30 <gjanssens> So that should be fixed first.
13:15:39 <jralls> Right.
13:15:49 <gjanssens> Then we can think of lower-priority syntax.
13:17:19 <gjanssens> For path-type variables this doesn't matter much as the environment file can simply expand the existing paths and add to it. It's rather for variables that take only one value like the GNC_DBD_DIR one.
13:17:34 <gjanssens> markvandenborre: indeed
13:18:29 <jralls> But it's generally unnecessary as long as the dbi drivers are in a "dbd" subdirectory of the one where libdbi is, which is almost always the case.
13:19:29 <gjanssens> jralls: as for the "non-intuitive" part of the issue, I believe I have been considering this in the past and I chose to implement it as I did because of security considerations.
13:20:06 <gjanssens> My thought was the locations where gnucash looks for shared libraries shouldn't be easily alterable after installation.
13:20:58 <gjanssens> For a user-local installation the user does as he/she wishes (and can override either at build time or at runtime via environment.local)
13:21:21 <gjanssens> For a system wide installation the same override mechanisms apply.
13:22:01 <gjanssens> I wouldn't want a malicious script or whatever have the user load a manipulated dbd driver that tracks aqbanking interactions for example
13:22:21 <jralls> Careful. Where GnuCash looks for shared libraries is determined before the program starts running. Changing LD_LIBRARY_PATH in etc/gnucash/environment will have no effect at all. You mean to say where GnuCash looks for loadable modules.
13:22:45 <gjanssens> Indeed, so the whole point is a bit moot I suppose.
13:23:57 *** gjanssens is now known as gjanssens_afk
13:24:11 <markvandenborre> gjanssens: crash when trying to import the csv on the latest debian stable packaged version of gnucash :-(
13:24:17 * markvandenborre tries again
13:27:28 <jralls> Not moot, just different from what you said. Dbd drivers are loadable modules.
13:29:25 <jralls> gjanssens_afk: But aside from that if a bad actor can insert a credential-stealing libdbdmysql.so he can also modify etc/gnucash/environment to load it.
13:30:50 <markvandenborre> gjanssens_afk: working now
13:30:55 <markvandenborre> thx for the great work!
13:31:46 <markvandenborre> is there a way create presets for importing csv files of a specific type?
13:31:59 <markvandenborre> case in point: this argenta import
13:32:15 <markvandenborre> it's not perfect yet, but it is very useful already
13:32:43 <markvandenborre> or is there a way to share those import settings with other gnucash users?
13:33:52 <markvandenborre> afk for a moment, but please do let me know if there is; would be a great way to make my tiny contribution, as I have accounts in a few different places...
13:45:27 *** User has quit IRC
14:10:54 *** kael has quit IRC
14:26:37 *** weasel has quit IRC
14:33:47 *** gjanssens_afk is now known as gjanssens
14:34:49 <gjanssens> jralls: a bad actor could have the user download such a libdbdmysql.so into user directories. He doesn't need elevated privileges for that
14:35:19 *** weasel has joined #gnucash
14:35:27 <gjanssens> But I was thinking if the user can override LD_LIBRARY_PATH and LTDL_LIBRARY_PATH, the bad actor could just leverage that
14:36:38 <gjanssens> Heh, except etc/gnucash/environment prepends those paths, making it less obvious again to override
14:36:50 <gjanssens> Anyway, glad you got it working
14:37:46 <gjanssens> markvandenborre: gnucash 2.8 will have csv import presets which you can create yourself
14:38:05 <gjanssens> Sharing them will be a bit more difficult
14:38:52 <gjanssens> I know aqbanking also has csv import facilities and it has quite a range configurations for different banks.
14:39:01 <jralls> True. libdbi doesn't use LTDL so it's just LD_LIBRARY_PATH, possibly overridden by GNC_DBD_DIR.
14:39:38 <gjanssens> The bad thing is gnucash largely ignores those aqbanking import features so they're currently of not much use
14:39:59 <gjanssens> My plan was to pick this up in the next iteration of the import conversions to c++
14:40:10 <gjanssens> But those won't make 2.8 unfortunately.
14:40:41 *** gjanssens is now known as gjanssens_afk
15:02:43 *** Mechtilde has quit IRC
15:33:07 *** gjanssens_afk is now known as gjanssens
15:55:18 *** Ya_ALLAH_Ya_Muhmd has joined #gnucash
15:55:18 *** Ya_ALLAH_Ya_Muhmd has left #gnucash
16:19:30 *** fbruetting has quit IRC
16:21:09 *** ChanServ sets mode: +qo warlord warlord
16:21:11 *** warlord sets mode: +o gncbot
16:21:13 <warlord> .
16:26:52 *** ArtGravity has joined #gnucash
16:31:34 *** Ya_ALLAH_Ya_Muhmd has joined #gnucash
16:31:34 *** Ya_ALLAH_Ya_Muhmd has left #gnucash
16:34:11 *** jralls sets mode: +b *!Ya_ALLAH_Y@ip29.ip-87-98-166.eu
16:39:19 *** ArtGravity has quit IRC
16:40:09 *** ArtGravity has joined #gnucash
16:45:00 *** pilotauto has joined #gnucash
16:48:06 <markvandenborre> gjanssens: thx
16:56:21 *** VigilantSky has joined #gnucash
17:10:28 *** VigilantSky1 has joined #gnucash
17:11:55 *** VigilantSky has quit IRC
17:11:56 *** VigilantSky1 is now known as VigilantSky
17:19:19 *** ArtGravity has quit IRC
17:33:04 *** ArtGravity has joined #gnucash
17:45:03 *** fell has joined #gnucash
18:00:01 *** gjanssens has quit IRC
18:07:15 *** bertbob has quit IRC
18:32:18 *** bertbob has joined #gnucash
20:29:40 *** fell has quit IRC
20:53:01 <lmat> jralls: I had not followed that XML file crash bug, but I was aware of that bug in kvp on master. My pull request removes all kvp parsing, thus fixing that issue.
21:00:37 <lmat> jralls: The suggested patch is correct, too.
21:22:22 *** ArtGravity has quit IRC
23:23:04 *** frakturfreak has quit IRC
23:45:38 *** jotrago has joined #gnucash