2018-08-30 GnuCash IRC logs
00:42:55 *** fell has quit IRC
00:50:40 *** puck has quit IRC
00:53:15 *** puck has joined #gnucash
01:20:14 *** gour has joined #gnucash
01:41:59 *** Mechtilde_ has joined #gnucash
02:01:57 *** Mechtilde_ has quit IRC
02:13:00 *** gjanssens has joined #gnucash
02:13:00 *** ChanServ sets mode: +o gjanssens
02:13:18 *** Cuare has joined #gnucash
02:17:21 *** harshitaneja has joined #gnucash
02:28:21 *** harshitaneja has quit IRC
02:35:38 *** harshitaneja has joined #gnucash
02:52:26 *** warlord has quit IRC
03:08:58 *** boldstripe has joined #gnucash
03:29:28 *** warlord has joined #gnucash
03:43:29 *** Mechtilde_ has joined #gnucash
03:56:40 *** bertbob has quit IRC
03:58:30 *** bertbob has joined #gnucash
04:02:29 *** ncv has joined #gnucash
04:04:04 *** bertbob has quit IRC
04:05:07 *** gour has quit IRC
04:05:26 *** bertbob has joined #gnucash
04:08:46 *** pilotauto has quit IRC
04:09:21 *** gour has joined #gnucash
04:26:26 *** jotrago has quit IRC
04:49:24 *** Aussie_matt has joined #gnucash
04:50:00 *** luwum[m] has quit IRC
04:50:00 *** DiogoGomes[m] has quit IRC
04:50:02 *** mmkodali[m] has quit IRC
04:51:20 *** DiogoGomes[m] has joined #gnucash
05:06:19 *** boldstripe has quit IRC
05:21:42 *** User has joined #gnucash
05:22:25 *** fabior has joined #gnucash
05:25:05 *** Mechtilde_ has quit IRC
05:29:48 *** fabior has quit IRC
05:33:43 *** fabior has joined #gnucash
05:36:42 *** fabior has quit IRC
05:37:23 *** fabior has joined #gnucash
05:40:24 *** fabior has quit IRC
05:41:05 *** fabior has joined #gnucash
05:50:53 *** User has quit IRC
06:12:02 *** warlord has quit IRC
06:43:42 *** mmkodali[m] has joined #gnucash
06:43:46 *** luwum[m] has joined #gnucash
06:55:18 *** fekepp1 has joined #gnucash
06:55:30 *** Aussie_matt has quit IRC
06:55:46 *** fekepp has quit IRC
06:55:46 *** fekepp1 is now known as fekepp
06:56:21 *** harshitaneja has quit IRC
07:01:29 *** ncv has quit IRC
07:02:36 *** ncv has joined #gnucash
07:03:25 *** fabior has quit IRC
07:04:25 *** harshitaneja has joined #gnucash
07:18:29 *** jotrago has joined #gnucash
07:19:49 *** oozer has joined #gnucash
07:21:59 *** Mechtilde_ has joined #gnucash
07:36:38 *** ncv has quit IRC
07:39:59 *** ncv has joined #gnucash
07:41:38 *** warlord has joined #gnucash
07:50:18 *** Jimraehl1 has joined #gnucash
07:52:44 *** Jimraehl1 has left #gnucash
07:56:00 *** jotrago has quit IRC
08:02:26 *** fell has joined #gnucash
08:14:57 *** harshitaneja has quit IRC
08:15:33 *** harshitaneja has joined #gnucash
08:15:51 *** oozer has quit IRC
08:57:03 *** storyjesse has quit IRC
09:15:16 *** Mechtilde_ has quit IRC
09:22:09 *** warlord has quit IRC
09:22:33 *** Mechtilde_ has joined #gnucash
09:25:12 *** fabior has joined #gnucash
09:31:50 *** fabior has quit IRC
09:32:41 *** Mechtilde_ has quit IRC
09:48:22 *** fabior has joined #gnucash
10:08:23 *** fabior has quit IRC
10:12:54 *** jotrago has joined #gnucash
10:15:40 *** oozer has joined #gnucash
10:19:41 *** storyjesse has joined #gnucash
10:24:49 *** fabior has joined #gnucash
10:25:50 *** ncv has quit IRC
10:26:54 *** ncv has joined #gnucash
10:28:18 *** windboy has joined #gnucash
10:28:24 *** fabior has quit IRC
10:32:15 *** windboy has quit IRC
10:36:02 *** windboy has joined #gnucash
10:54:05 *** fabior has joined #gnucash
11:49:14 *** warlord has joined #gnucash
11:51:50 *** gncbot sets mode: +o fell
11:59:48 <CDB-Man> <CDB-Away> hmm, when I tried to run the finance quote script to update prices, the perl screen hangs permanently -- I let it run for 10 hours while I was at work and no response. contract this to before where it would at least update 1/2 of my prices. did something change?
12:30:08 *** ncv has quit IRC
12:30:57 *** ncv has joined #gnucash
12:32:36 *** storyjesse has quit IRC
12:34:41 *** Mechtilde_ has joined #gnucash
12:58:04 *** ncv has quit IRC
13:01:05 *** ncv has joined #gnucash
13:05:24 *** gour_ has joined #gnucash
13:07:36 *** gour has quit IRC
13:07:49 <jralls> CDB-Man: Not in GnuCash.
13:13:35 *** harshitaneja has quit IRC
13:15:23 *** harshitaneja has joined #gnucash
13:20:20 *** fabior has quit IRC
13:43:54 *** Mechtilde_ has quit IRC
13:47:22 *** bertbob has quit IRC
13:47:38 *** ncv has quit IRC
13:48:50 *** ncv has joined #gnucash
13:52:08 *** Mechtilde_ has joined #gnucash
14:00:33 *** bertbob has joined #gnucash
14:03:15 *** fell has quit IRC
14:05:55 <jralls> chris: Go ahead and refactor commodity-utilities if you like. I've just pushed a fix for the one glaring error turned up by the unit tests.
14:06:23 *** fell has joined #gnucash
14:07:34 <jralls> chris: Do be careful, though: https://bugs.gnucash.org/show_bug.cgi?id=775368 shows the pitfalls of not thoroughly understanding what's going on. I hope the unit tests help clarify how things fit together.
14:11:22 <jralls> chris: One other thing, most of the functions in commodity-utilities are exported even though they're used only internally, so you'll have to preserve at least the outer bits when refactoring.
14:12:28 <jralls> gjanssens: How is the IM work going?
14:17:42 *** ncv has quit IRC
14:22:05 <gjanssens> jralls: not very well
14:22:26 <gjanssens> I have spent 3 days this weekend on it and I'm still not much further...
14:22:43 <gjanssens> I have had to set it aside for a couple of days due to other obligations
14:23:20 <jralls> OK. What have you tried and what problems did you encounter?
14:23:37 <gjanssens> My first effort is focused on making the gnucash-item-edit the primary (preferably single) widget for keyboard handling.
14:24:19 <gjanssens> But model controller and view are spread out over 3 directories in the register so I keep on hitting unexpected interferences.
14:24:57 <gjanssens> For contrast, in the current code the gnucash-sheet is the primary widget for keyboard handling
14:25:27 <gjanssens> But that's in essence a GtkLayout which has no built-in keyboard handling intelligence at all
14:25:41 <gjanssens> So it is basically reimplementing everything from scratch
14:26:06 <gjanssens> The gnucash-item-edit on the other hand is at its core a GtkEntry
14:26:09 *** boldstripe has joined #gnucash
14:26:28 <gjanssens> That widget is one of Gtk's primary keyboard accepting widgets, so it's much more tuned for it.
14:27:06 <gjanssens> I'm relatively close in the pure gtk aspect of this.
14:27:41 <gjanssens> It's the way the register's mvc code is set up that keeps me running in circles.
14:28:00 <jralls> What happens if you just rip all of the event handling out of gnucash-sheet and make it a pure layout container?
14:29:19 <gjanssens> For example it implements callbacks to insert in several parts of the gtk widget callbacks to handle stuff like the calculator function in number fields or the speed keys in the date field.
14:30:15 <gjanssens> My current idea is to redo the widget hierarchy to make specialized widgets for each of the cell types the register supports and move this specialized key interception code into each of them.
14:30:28 <gjanssens> Will need some more time for that though :)
14:30:58 <gjanssens> And for gnucash-sheet as a pure layout container, we can't.
14:31:07 <jralls> Because?
14:31:19 <gjanssens> It needs to handle at least mouse-clicks to move the cursor around
14:31:36 <gjanssens> The item-edit can't intercept those
14:32:29 <gjanssens> Equally when a read-only line is active there is no item-edit at all, but we still want to be able to move the cursor around.
14:32:58 <jralls> Gtk will send the event to the right widget as long as we don't get in its way. If the widget handles the event, done. Otherwise Gtk passes it to the widget's parent container and so on up to the toplevel.
14:33:04 <gjanssens> (examples of a read-only line are invoice/bill transactions in the AR/AP accounts)
14:33:44 <jralls> So the widgets in a read-only line need to know that they're read only and not handle the event.
14:34:05 <gjanssens> There are no widgets in a read-only line other than the sheet
14:34:23 *** frakturfreak has joined #gnucash
14:34:29 <gjanssens> It's one big drawing board. The only real widget is the gnucash-item-edit
14:34:58 <gjanssens> Which is superimposed on the simple drawing to create the illusion of an all active register.
14:35:10 <jralls> Then the sheet isn't a pure container. Anything that isn't in a gnucash-item-edit should be in a GtkLabel.
14:35:41 <gjanssens> Correct, the sheet is not a pure container.
14:35:47 <gjanssens> But it's very efficient.
14:35:59 <gjanssens> Working with gtklabels will add quite a bit of overhead IMO
14:36:13 <jralls> Shouldn't, they're really lightweight.
14:36:16 <gjanssens> (I think a GtkTreeView is doing something similar in fact)
14:36:44 <gjanssens> I'm off for supper now...
14:36:50 <jralls> OK.
14:58:45 *** gncbot sets mode: +o fell
15:12:11 *** boldstripe has quit IRC
15:22:48 <gjanssens> jralls: back
15:23:47 <gjanssens> The challenge is currently how drastic a change am I willing to make in the middle of the 3.x series.
15:23:49 <jralls> Well, I guess reviving Reg2 so that we actually use GtkTreeView is one way to get there.
15:24:06 <gjanssens> :) I was about to mention that.
15:24:33 <gjanssens> But it's pretty drastic a change :(
15:24:44 *** harshitaneja has quit IRC
15:25:26 *** harshitaneja has joined #gnucash
15:25:36 <gjanssens> We've had quite a few poor user experiences to fix after 3.0 release. I'd like to not introduce a new one so I'm being conservative.
15:25:53 <gjanssens> For 4.x I'm very much geared towards GtkTreeView
15:25:55 <jralls> It is. So is rewriting gnucash-sheet. Even if we decide that either path is OK for 3.x it's not likely to get done in time for 3.3 and there are some pretty bad UX issues that we need to fix.
15:26:33 <jralls> And yes, we sure don't need any new ones.
15:27:19 <gjanssens> As I was very much focussed on this register clean-up, can you remind me of the major ones we should get fixed by 3.3 ?
15:27:48 <gjanssens> Are there still IM issues after your last fixes ?
15:29:56 <jralls> https://bugs.gnucash.org/show_bug.cgi?id=796734 is the biggie, though I'm not sure that the underlying problem is in GnuCash or imquartz.
15:30:28 <jralls> https://bugs.gnucash.org/show_bug.cgi?id=795253 is the other.
15:32:56 <jralls> Though that one is "needinfo" waiting confirmation that my fix from 5 June did the trick. I'm inclined to take silence == assent after almost 3 months.
15:41:46 <gjanssens> Can you reproduce 796734 yourself ?
15:42:34 <gjanssens> It does work as expected on Fedora indeed (although it gets messy when I enable ibus as GTK_IM_MODULE)
15:52:18 <gjanssens> jralls: unfortunately I still haven't recovered my windows test box, so currently there's not much I can add for 795253
15:52:27 *** RASSRQ has joined #gnucash
15:53:37 <jralls> Yes, I can reproduce 796734, and I think 796253 is fixed.
15:53:37 <gjanssens> And can you reproduce 796734 in any of your other VMs? (Fedora/debian,...)
15:54:13 <jralls> No, it's a mac-only problem which is why I suspect that the problem might be in imquartz instead of GnuCash.
15:54:37 <gjanssens> Well, that's why I asked. It's a reasonable clue to assume so at least.
15:56:02 <jralls> There's something about the preedit handling that deletes the selection. It actually happens in that input method handling block in gnucash-sheet.c which is why I hoped that you could just make it go away.
15:56:49 <jralls> But since it looks like we're stuck with it form now I'll have to dig deeper and figure out what is actually deleting it.
16:00:23 <gjanssens> I haven't given up completely yet, but I had a few non-gnucash related matters that took priority.
16:00:44 <gjanssens> However it's be safer if you could pinpoint the cause with the code as is indeed :(
16:01:47 <gjanssens> Anyway, time to call it a day here.
16:01:57 <jralls> I'll take that on next week. This weekend is the local highland games and I'll be AFK from this evening til Monday.
16:02:03 <jralls> Good night!
16:02:11 <gjanssens> Heh, have fun!
16:02:12 <gjanssens> Bye
16:03:15 *** gjanssens has quit IRC
16:32:14 *** RASSRQ has quit IRC
16:56:25 *** RASSRQ has joined #gnucash
17:13:03 *** harshitaneja has quit IRC
17:14:49 *** boldstripe has joined #gnucash
17:16:11 *** chris has quit IRC
17:21:24 *** fell has quit IRC
17:32:31 *** jonas has joined #gnucash
17:41:05 <RASSRQ> Credit-Notes have a seriousl bug that Geert said he has fixed and will be in ver 3.3. I really need this so have installed the maint binary Windows installer from 2018-08-23 from https://code.gnucash.org/builds/win32/maint/. The installer seems successful, but when I run it the Tip-Of-The-Day windows comes up and the GnuCash splash screen starts. I dismiss the Tip windows, and the splash screen works for a bit and then closes and Gn
17:41:07 <RASSRQ> uCash is not running, and no log file. I tried the installer from 2018-08-29 with exactly the same result. Is there something special that needs to be done to get a maint version to run successfully?
17:42:29 <jralls> RASSRQ: I haven't tested the Windows nightlies lately. Hang on while I give it a try.
17:44:26 *** boldstripe has quit IRC
17:52:58 <jralls> RASSRQ: This morning's nightly starts up for me with no issues, so maybe something in your data. Try running it from a shell (CMD or Powershell is fine) with --nofile and see if it starts up. If it does, try --debug so it writes more output to the tracefile.
17:53:33 <RASSRQ> JRALLS: Will do. THANK YOU.
18:04:48 *** gour_ has quit IRC
18:06:37 *** frakturfreak has quit IRC
18:08:39 <RASSRQ> JRALLS: with --nofile it starts. Opening my file crashes GnuCash (which opens with stable 3.2). I used --debug so where do I find the .log file? There is no .log file in the GnuCash program dir, nor in the dir where my .gnucash file is. Where would I find a log? Also, FYI, using stable 3.2 a Customer Report shows an anomoly -- 2 invoices in the customer report, showing different $-values, open the same invoice. So I can beli
18:08:41 <RASSRQ> eve the database is munged, perhaps because of the invoice/credit-note bugs Geert was working on. If a run a Repair in stable 3.2, the problem doesn't go away, so if the munged DB is why the Maint version crashes, how could I repair the DB? In short, what can I do to move the ball forward on this? THX.
18:09:51 <jralls> RASSRQ: $HOME/AppData/Local/Temp/gnucash.trace.XXXXXX.log.
18:10:59 *** chf has quit IRC
18:11:02 <jralls> There will be a lot of them, so sort by date... probably easiest to do in Windows Explorer unless you're a CMD or Powershell whiz.
18:11:24 *** chf has joined #gnucash
18:12:10 <jralls> RASSRQ: Is there a bug associated with the credit note thing gjanssens was working on?
18:30:41 *** RASSRQ has quit IRC
18:31:47 *** kus has joined #gnucash
18:32:20 *** RASSRQ has joined #gnucash
18:36:24 *** pilotauto has joined #gnucash
18:37:03 <RASSRQ> JRALLS: The very end of the log file has this:
18:37:08 <RASSRQ> * 18:01:14 INFO <qof.session> [QofSessionImpl::load] new book=0ae61af0
18:37:08 <RASSRQ> * 18:01:14 CRIT <gnc.backend.dbi> [error_handler()] DBI error: -5: An invalid name was passed to libdbi
18:37:08 <RASSRQ> * 18:01:14 CRIT <gnc.backend.dbi> [error_handler()] DBI error: -5: An invalid name was passed to libdbi
18:37:08 <RASSRQ> * 18:01:14 CRIT <gnc.backend.dbi> [error_handler()] DBI error: -5: An invalid name was passed to libdbi
18:37:08 <RASSRQ> * 18:01:14 CRIT <gnc.backend.dbi> [error_handler()] DBI error: -5: An invalid name was passed to libdbi
18:37:08 <RASSRQ> * 18:01:14 CRIT <gnc.backend.dbi> [error_handler()] DBI error: -5: An invalid name was passed to libdbi
18:37:09 <RASSRQ> * 18:01:14 CRIT <gnc.backend.dbi> [error_handler()] DBI error: -5: An invalid name was passed to libdbi
18:37:45 <jralls> RASSRQ: Oh, you're using SQL... which backend?
18:37:54 <RASSRQ> sqlite3
18:38:35 <jralls> OK, let me try loading a SQL test file...
18:40:13 <jralls> OK, I can replicate that.
18:43:13 <RASSRQ> JRALLS: To answer your earlier question: In version 3.2 the Credit-Note got 'inverted', e.g., a CREDIT of $279 would actually CHARGE $279. Several people had that problem. Also a problem that unposting and then reposting an invoice ended up 2 separate invoices, and that's how I got the weird thing where the Invoice links in a Customer Report both pointed to the same Invoice. Worse, the Customer report showed Invoice-1 at $0, Invo
18:43:13 <RASSRQ> ice-2 showed it as $X, but clicking either link opened the invoice whose total was really $Y. SO, I'm hoping Geerts' recent changes fixes all that.
18:44:44 <RASSRQ> Sounds promising that you can replicate the problem with the Maint build. Can you advise me what is next or when I should try again? Again, THANKS!
18:45:06 <jralls> RASSRQ: OK, that's not really what I was asking about. Doesn't matter, this is clearly a separate problem altogether. You might go back to 3.2, save-as your file as XML, and then try again with the nightly just to see if gjanssens's fix worked for you.
18:45:33 <RASSRQ> Great idea. Will do.
19:02:58 *** Cuare has quit IRC
19:05:01 *** windboy has quit IRC
19:09:38 *** fell has joined #gnucash
19:19:28 <RASSRQ> JRALLS: I'm working with an XML file now. Geert's change fixed NEW Credit-Notes having the correct value, but left old credit notes that show a wrong value in the Customer Report but, when opened, have no transaction and cannot be UnPosted. I.e., I can't find a way to correct the errors caused by the 3.2 bug. I'd like to report this to him with some screen shots. Can you please tell me how I can get that to him? Thx.
19:20:19 <jralls> RASSRQ: The best way would be to update the bug report if there is one or to file a new one if there isn't.
19:23:41 <jralls> RASSRQ: The only bug that comes up when searching for "credit note" is https://bugs.gnucash.org/show_bug.cgi?id=796766. That doesn't look like the original problem.
19:25:01 <jralls> RASSRQ: So you should go ahead and file a new one.
19:43:45 <jralls> RASSRQ: Meanwhile I've found the cause of the DBI crash.
19:58:58 *** CharlesIC has joined #gnucash
19:58:59 <CharlesIC> hi
19:59:07 <CharlesIC> what's the best way to use gnucash between more than one computer/device?
19:59:19 <CharlesIC> i think storing its files on google drive might be dangerous
20:00:14 <RASSRQ> JRALLS: Thx for addressing the DBI crash so quickly. I just filed bug 796836. I hope the attachment to the report has enough snapshots and description to be clear.
20:00:20 *** chris has joined #gnucash
20:02:44 *** pilotauto_ has joined #gnucash
20:03:46 *** pilotauto has quit IRC
20:05:46 <jralls> RASSRQ: And fixed. You should be able to open your SQLite3 file with tomorrow's nightly.
20:23:10 *** CharlesIC has quit IRC
20:27:05 <RASSRQ> JRALLS: Great. Many thx.
20:57:38 <chris> jralls - an experiment - in latest commodity-utilities.scm, line 452, try surround ((caadr known-pair) 'total #f) with (pk)... i.e. line 452 is now (pk ((caadr known-pair) 'total #f))
20:57:58 <chris> (pk x) is a debugging tool - it'll dump x while running program
20:58:24 <chris> afterwards run test-commodity-utils.scm - it will show there are a few div/0 occurring
21:04:46 <chris> well there's 1 div/0
21:11:03 *** luc14n0 has quit IRC
21:34:47 *** luc14n0 has joined #gnucash
21:35:26 *** bertbob has quit IRC
21:37:34 *** bertbob has joined #gnucash
22:03:57 *** RASSRQ has quit IRC
22:23:12 *** oozer has quit IRC