2018-07-07 GnuCash IRC logs
00:11:52 *** pilotauto has quit IRC
00:33:22 *** fell_laptop has joined #gnucash
00:35:14 *** fell has quit IRC
00:36:07 *** boldstripe has quit IRC
00:56:09 *** windboy has quit IRC
01:00:00 *** harshitaneja has joined #gnucash
01:34:49 *** frakturfreak has quit IRC
01:48:00 *** wget has quit IRC
01:49:37 *** frakturfreak has joined #gnucash
01:50:27 *** wget has joined #gnucash
02:01:51 *** Cuare has quit IRC
02:18:16 <chris> @tell gjanssens all fine... I still lean towards blue pen myself; a lilac quill is odd...
02:18:16 <gncbot> chris: The operation succeeded.
02:19:39 <chris> jralls let's clean up #378... all commits except c1b287d and 36ad9e3 are very safe
02:20:01 <chris> I'll push into automated report testing in my private branch and populate the wiki
02:21:18 *** gjanssens has joined #gnucash
02:21:19 *** ChanServ sets mode: +o gjanssens
02:22:35 <gjanssens> .
02:22:35 <gncbot> gjanssens: Sent 9 hours and 3 minutes ago: <warlord> Oh, I couldn't tell that that was a quill pen in #3. I was going to say I like #1, but kind of wish there were the stacks of bills.
02:22:36 <gncbot> gjanssens: Sent 4 hours and 58 minutes ago: <fell> With some shadow the quill might be easier recoqnized. As gag youu could shrink it in the chart.
02:22:37 <gncbot> gjanssens: Sent 4 minutes ago: <chris> all fine... I still lean towards blue pen myself; a lilac quill is odd...
02:22:50 <chris> If I can complete automated pairwise testing (in my private branch) I'm sure more bugs can be shaken out of the tree :) I already found a couple of report crashers/segfaulters
02:23:53 <gjanssens> chris: yes. It reminds me of the fuzz testing that happens on the linux kernel.
02:24:27 <gjanssens> I understand jralls' concern about time it takes to run the tests though
02:25:16 <gjanssens> We may have an independent test command like 'stress-test' or something similar that can be invoked independently of the unit tests (check)
02:25:22 <chris> I'll see how to optimize it. perhaps e.g. the mini-stress-test can be default, and combinatorial-stress-test only run from time to time, e.g. prior to release
02:25:58 <chris> it's mainly for my own protection too, because i'll be the one messing up scheme innards
02:26:37 <chris> (system "./jenny 2 2 2") does the trick :)
02:53:11 *** harshitaneja has quit IRC
02:54:50 *** harshitaneja has joined #gnucash
03:15:51 *** gour has joined #gnucash
04:36:14 *** Mechtilde has joined #gnucash
05:09:38 *** Mechtilde has quit IRC
05:10:01 *** Mechtilde has joined #gnucash
05:13:37 *** Mechtilde has quit IRC
05:17:28 *** Mechtilde has joined #gnucash
05:33:50 *** Mechtilde has quit IRC
06:19:16 *** ncv has joined #gnucash
06:55:55 *** Jimraehl1 has joined #gnucash
06:56:44 *** Jimraehl1 has left #gnucash
07:27:47 *** gour has quit IRC
07:30:19 *** gour has joined #gnucash
08:03:10 *** boldstripe has joined #gnucash
08:16:22 *** finster has joined #gnucash
09:38:13 *** oozer has joined #gnucash
09:40:06 <warlord> .
09:52:17 *** User has joined #gnucash
10:42:43 *** CronnoTR has joined #gnucash
10:55:55 *** bertbob has quit IRC
10:56:22 <CronnoTR> Hi, I have trouble creating income report - I earn in pounds but receive dollars. I'm using the trading accounts when transfering monies. The problem occurs that the commodities are either calculated by weighted average or nearest to the report date which skews the calculations. Is there an option for setting that the commodities are taken at value
10:56:22 <CronnoTR> on the day of transfer?
11:02:12 *** bertbob has joined #gnucash
11:17:43 <gjanssens> warlord, jralls, fell_laptop, chris: a few more iterations on the wiki favicon: https://screenshots.firefox.com/0Cdk2RLros2BR54b/null
11:18:29 <gjanssens> For #4 I chose to use the exact lila colors as on our main website. While the quill is much clearer in larger format, it lacks contrast in the smaller images.
11:19:11 <gjanssens> So for #5 and #6 I chose to only use the darker purple of our site and give it an even darker border.
11:20:18 <gjanssens> #7 nd #8 are exactly like #5 and #6 except the quill is in blue, was chris asked for. I has strong contrast on all sizes, but no link with the main theme colors of the main site.
11:21:11 <gjanssens> #4, #5 and #7 have a slightly larger quill than I had in #2 and even than in #3
11:21:54 <gjanssens> #6 and #8 implement fell_laptop's suggestion to reduce the quill size to the diagram. I have also removed the other diagram details on these to to reduce clutter.
11:22:30 <gjanssens> I only kept the haze on the quill
11:23:01 *** CronnoTR has quit IRC
11:23:38 <gjanssens> warlord: as for not seeing it's a quill pen, I had also found more detailed svg's but they were too detailed to work in a small icon. So I went for a more abstract one.
11:24:32 <gjanssens> Personally I'm most happy with #5 or #6. Both have strong points and I find it hard to choose between those two
11:25:07 *** finster has left #gnucash
11:25:38 <gjanssens> Oh, and I didn't add shadows as these would make the the smaller icons more incomprehensible: too much detail to handle on such small surface.
12:07:40 *** CronnoTR has joined #gnucash
12:17:10 * jralls likes #5 & #7, with #7 slightly favored because it has better contrast on the smallest image. The quills on #4, #6, & #8 are pretty much invisible at the smallest scale... OTOH I'm looking at it on a 2560x1440 monitor so the smallest renders at 1mm^2.
12:18:51 <jralls> chris: Do you have any comment on https://github.com/Gnucash/gnucash/pull/379?
12:21:19 <jralls> gjanssens: Bob has answered your comments on https://github.com/Gnucash/gnucash/pull/374...
12:22:31 <gjanssens> jralls: 2560x1440 is that HiDPI already ?
12:22:59 <gjanssens> I'll see if I can still aswer Bob
12:23:03 *** rasmus has joined #gnucash
12:23:33 <gjanssens> Starting Tuesday I'll be mostly afk until the end of the month
12:23:41 <gjanssens> Hollidays in Italy
12:24:13 <rasmus> nice
12:25:52 *** Agfarmer18 has joined #gnucash
12:26:21 <jralls> gjanssens: I guess it depends on what meaning of "HiDPI" you use. Apple's is > 192dpi and since it's a 68cm monitor it's a bit short of that. I think at that size it needs to be "4K" to qualify as "Retina".
12:26:31 <jralls> Where in Italy?
12:27:13 <gjanssens> First a few days in the mountains around bibbiena and afterwards more to the coast, direction Rimini
12:28:41 <rasmus> Hi everyone! I'm having trouble retrieving currency quotes via gnucash gui. I'm running ubuntu 18.04, perl Finance::Quote version 1.47 is installed. The gnc-fq-check script produces output. but so far i havent been able to get any stock quote via gnc-fq-dump.
12:29:56 <jralls> rasmus: Have you done everything in https://wiki.gnucash.org/wiki/FAQ#Q:_Why_doesn.27t_online_quoting_work.3F?
12:30:50 <jralls> rasmus: Note particularly that gnc-fq-dump can't read your GnuCash preferences so you need to set the environment variable.
12:30:58 <rasmus> i got an alphavantage API Key. But the tab Online Quotes in the gnucash edit -> preferences does not exist.
12:31:48 <jralls> rasmus: Then you're using GC2.6.x. You need the environment variable for both GnuCash and gnc-fq-dump.
12:32:16 <rasmus> thanks jralls, i should have checked that out. i think this will solve my issue
12:39:39 *** CronnoTR has quit IRC
12:39:44 <gjanssens> GC2.6.19 or older. I have added the preference panel for Alpha Vantage in gnucash 2.6.20
12:43:36 <jralls> Oh, OK. IIRC Ubuntu 18.04 has 2.6.19.
12:46:58 <rasmus> Thanks for the hint, it worked. Thanks to all people involved in the gnucash project. It introduced me to serious bookkeeping. I already learned a lot and i will learn a lot more in the future.
12:51:51 *** Cuare has joined #gnucash
12:51:56 <jralls> gjanssens: How well do you understand Gtk input modules?
12:52:38 <gjanssens> jralls: I'm a total beginner :(
12:53:07 <gjanssens> But I had it on my plan to study it. The reference webpages are already open in my browers
12:53:13 <gjanssens> browser*
12:53:37 <gjanssens> It's something I want to understand better.
12:53:53 <gjanssens> This is probably concerning the register issues ?
12:56:44 *** rasmus has left #gnucash
12:57:27 <jralls> Yeah. I'm sure https://bugs.gnucash.org/show_bug.cgi?id=796734 is because imquartz is eating the selection. I suspect that https://bugs.gnucash.org/show_bug.cgi?id=796723 might too.
12:58:55 <jralls> Then there's https://bugs.gnucash.org/show_bug.cgi?id=796733.
13:00:27 <jralls> Back in 2010 https://github.com/Gnucash/gnucash/commit/5c5b031c5d49038a10b4e49e14aa880f28b36761 added a bunch of IM handling code to gnucash-sheet, and I wonder how much of it is really necessary.
13:00:59 <gjanssens> jralls: I think most of it can go away if we reorganize the widget hierarchy a bit
13:01:21 <jralls> OK, what do you propose?
13:01:22 <gjanssens> Back in 2010 we were using gnome-canvas which did all by itself
13:01:50 <gjanssens> Now we have a GtkLayout for overall display and text entry happens in an ordinary GtkEntry widget
13:02:15 <gjanssens> However the keyboard handling was not fully transferred to the GtkEntry
13:02:22 <gjanssens> I think that's what should happen
13:02:35 <gjanssens> The GtkEntry should grab the focus rather than the GtkLayout
13:03:16 <gjanssens> Any keys related to moving around should be passed up to the GtkLayout, but other than that the GtkEntry should do most of the work
13:03:55 <gjanssens> I hope/expect we can ride on the built-in IM support in GtkEntry to solve many of our issues.
13:04:09 <jralls> Autofill has to get triggered somehow, so should it be a gnc widget derived from GtkEntry with the autofill functions?
13:05:54 <gjanssens> IIRC a GtkEntry has built-in support for autofill (via a GtkCompletion)
13:06:13 <gjanssens> So I assume Gtk already solved this issue for us.
13:06:25 <gjanssens> We only need to write our custom GtkCompletion callback function
13:07:38 <gjanssens> We do this in other parts of gnucash as well
13:08:20 <gjanssens> This callback is of course different depending on the column we're in.
13:08:58 <gjanssens> Only the date column may need some special attention as we pop down our own custom date widget.
13:09:04 <jralls> And the account column is a GtkCombo...
13:09:17 <gjanssens> I don't even think so
13:09:39 <gjanssens> It's still the GtkEntry with a custom arrow button that handles the pop-down
13:09:40 <jralls> No? How does it display its pick list?
13:10:07 <gjanssens> The register predates these advanced widgets I think
13:10:21 <gjanssens> And time was too short to refactor all that before release.
13:10:36 <gjanssens> But it has been nagging ever since :(
13:11:37 <jralls> Register-gnome was written for gtk-2.0. I don't know how much i18n support was even in Gtk back then.
13:12:00 <jralls> s/written/re-written/
13:12:30 <gjanssens> It was re-written using gnome-canvas
13:12:53 <gjanssens> So the question would be how much i18n support was in gnome-canvas.
13:13:06 <gjanssens> Probably not much as it got abandoned quite some time ago
13:14:01 <gjanssens> For your information just to be able to do these i18n hacks, register-gnome maintained an invisible GtkEntry and passed keys it intercepted from the canvas back and forth to it
13:14:49 <gjanssens> That's why there's still these odd things like locking and unlocking text entry on that widget. That's part of this old hack that didn't get cleaned out properly.
13:15:42 <gjanssens> It did all this to mimic more or less a GtkEntry's behaviour, but everything, including the active cell was hand-painted on the canvas
13:16:00 <jralls> Yuck.
13:16:23 <jralls> No wonder Bob's effort to convert it to GtkTreeView didn't work out.
13:16:38 <gjanssens> Indeed. It took me 3 attempts over multiple years to understand it enough to rewrite it to GtkLayout
13:17:20 <gjanssens> And while doing so still introducing bugs because I didn't grasp all the subtleties :(
13:17:20 <jralls> Damn good thing you had that mostly ready last summer!
13:17:33 <gjanssens> Yeah
13:18:18 <gjanssens> I think another attempt at using GtkTreeView may still be worth it though.
13:18:31 <jralls> So how to approach the bugs? We can't completely rewrite the register for 3.x, and maybe not even for 4.x if we're going to release that in 2021.
13:18:42 <gjanssens> There's still way too many wheels being reinvented in that code.
13:19:02 <gjanssens> When is 3,3 planned ?
13:19:51 <jralls> September 30.
13:19:59 <gjanssens> (Grmbl. Changed my keyboard today to have a comma on the numeric keypad as that's the normal decimal separator around here, but I keep mistyping release numbers because of muscle memory)
13:20:19 <gjanssens> Ok.
13:20:28 <jralls> That's OK, I understand either way. ;-)
13:21:22 <gjanssens> I could work on this in August and September - that is do the keyboard focus switch from layout to entry and see what happens
13:23:06 <gjanssens> If that doesn't help at least we're in normal Gtk territory to start debugging.
13:23:34 <gjanssens> Unless you want to dive into the GUI register code this month already...
13:26:30 <jralls> Hmm. Well, I'd like to nail those bugs, but it looks like I have a lot of study ahead to catch up to you. I'll keep at it but I probably won't be ready to start coding before you get back from Italy so I should probably work on other issues as well.
13:27:07 <gjanssens> Ok
13:27:14 <jralls> E.g. I *still* haven't begun the Timespec removal I'd planned for 3.2.
13:27:27 <gjanssens> Right
13:28:05 <jralls> Too many crashers. Those seem mostly under control (touch wood).
13:28:49 <gjanssens> That's a bit my problem as well. I had planned this register work a while back too. I even asked Bob to back off until I got it done.
13:30:06 <gjanssens> But I need some undivided time to focus on it and that didn't happen lately. So instead I did a few bugs that could get solved in shorter time spans.
13:30:25 <gjanssens> hasn't happened lately*
13:30:38 <gjanssens> (I think)
13:34:25 *** bertbob has quit IRC
13:36:11 <jralls> Yeah. Only chris has had time to focus on the big picture lately. :-(
13:36:25 *** bertbob has joined #gnucash
13:36:25 <chris> :-o
13:36:41 * gjanssens waves at chris :)
13:36:56 <chris> jralls #379 is currently harmless. IMNSHO it is rather silly - (set! x 2) (test-equal x 2).
13:37:52 <chris> I'd expect options.scm to at least test the functionality - (define x (make-date-option)) (option-set-data! x (cons 'relative 'nothing)) --> error 'nothing is not a valid option
13:37:57 <jralls> OK. I'll look it over, or feel free to comment on it yourself. I was mostly concerned about it being at cross purposes with 378.
13:38:03 *** Agfarmer18 has quit IRC
13:38:25 <chris> Will do when time comes, don't wish to scare the beginner :)
13:38:45 <chris> 378 and 379 are independent
13:39:07 <jralls> Speaking of 378, do you still want me to cherry-pick out of it or are you going to split it up into bug-fix and stress-test?
13:39:51 <chris> go ahead cherry-pick... I'll rebase and refine the stress-test, and seek opinion from devel this week
13:40:23 <jralls> For stress-test, I like the idea of a separate target for it. We could set up a Travis job to run it periodically.
13:40:54 *** gour has quit IRC
13:41:09 <chris> I expect the combinatorial stress test to dislodge numerous lurking bugs - perhaps scheme can query the environment
13:41:19 <jralls> OK. I'll work on that this afternoon. I need to go for my bike ride now before it gets too hot. We're expecting 30C today.
13:41:40 <jralls> Back in a couple of hours.
13:41:43 * chris needs to Zzz soon, finished a late emergency shift ...
13:41:45 *** jralls is now known as jralls_afk
13:42:13 *** gour has joined #gnucash
13:42:43 *** mdforbis_ has quit IRC
13:43:21 <chris> ^ back above, if testing (make-date-option) it shouldn't error on (cons 'relative 'nothing)... it should error on (cons 'absolute "string") because 'absolute can accept time64 and timepair; whereas 'relative accepts 'symbol (IIRC)
13:45:15 * chris waves back, my preference any #5-#8 with #6 on top
13:48:53 *** mdf has joined #gnucash
13:53:28 *** chris is now known as chris-zzz
14:02:45 *** fabior has joined #gnucash
14:19:08 *** User has quit IRC
14:37:11 *** fell_laptop is now known as fell
14:37:24 *** gncbot sets mode: +o fell
14:39:28 *** kus has joined #gnucash
14:46:28 *** Robert8471 has joined #gnucash
15:52:42 *** jralls_afk is now known as jralls
16:13:08 *** Robert8471 has left #gnucash
16:39:15 <warlord> gjanssens, jralls -- I'm fine with any of #5-#8
16:53:17 *** fabior has quit IRC
17:19:50 *** boldstripe has quit IRC
17:26:36 *** ncv has quit IRC
17:26:45 *** ncv has joined #gnucash
17:36:37 *** gjanssens has quit IRC
17:55:06 *** Cuare has quit IRC
18:08:50 *** tonysoar has joined #gnucash
18:20:26 *** gour_ has joined #gnucash
18:22:45 *** gour has quit IRC
18:30:41 *** ncv has quit IRC
18:34:50 *** ncv has joined #gnucash
18:36:34 *** ncv has joined #gnucash
18:59:37 *** tonysoar has quit IRC
19:09:43 *** kus has quit IRC
19:57:28 *** gour_ has quit IRC
20:12:18 *** tonysoar has joined #gnucash
20:43:42 *** oozer has quit IRC
21:47:43 *** mipesom has quit IRC
21:47:48 *** tonysoar has quit IRC
21:56:37 *** windboy has joined #gnucash
22:03:47 *** windboy has joined #gnucash
22:05:01 *** mipesom has joined #gnucash
22:11:58 *** windboy has quit IRC
22:16:31 *** windboy has joined #gnucash
22:18:49 *** windboy has quit IRC
22:25:44 *** windboy has joined #gnucash
22:43:21 *** ncv has quit IRC
23:28:01 *** fell has quit IRC
23:29:19 *** pilotauto has joined #gnucash