2018-03-09 GnuCash IRC logs
00:10:05 *** warlord has quit IRC
00:14:15 *** CDB-Away has joined #gnucash
00:18:26 *** marusich has joined #gnucash
00:19:43 *** CDB-Away has quit IRC
00:19:53 *** warlord has joined #gnucash
00:19:58 *** CDB-Away has joined #gnucash
01:38:28 *** Mechtilde has joined #gnucash
01:54:06 *** fell has quit IRC
01:54:44 *** fell has joined #gnucash
02:29:58 *** Mechtilde has quit IRC
02:38:41 *** Mechtilde has joined #gnucash
02:57:56 *** gjanssens has joined #gnucash
02:57:56 *** ChanServ sets mode: +o gjanssens
03:04:56 <gjanssens> .
03:10:28 *** mipesom has quit IRC
03:18:46 *** chris has joined #gnucash
03:24:50 *** mipesom has joined #gnucash
03:40:16 <chris> .
03:50:31 <chris> linechart complete :)
03:52:04 *** Yotson has joined #gnucash
03:53:11 *** fekepp has joined #gnucash
03:58:35 *** jonas has quit IRC
04:12:35 <chris> gjanssens - just wondering if you can remember - html-barchart has logic to display fewer ticks when resized too small - is there any particular reason the change wasn't applied to html-linechart? https://github.com/Gnucash/gnucash/commit/eac798756
04:13:32 <chris> related to https://bugzilla.gnome.org/show_bug.cgi?id=778042
04:14:09 <chris> (this work won't be pushed for a while)
04:32:15 *** fabior has joined #gnucash
04:38:40 *** Mechtilde has quit IRC
04:45:18 *** fabior has quit IRC
05:22:40 *** pilotauto has quit IRC
05:51:52 <gjanssens> chris: that's been a while. Does the line chart have the same issue ? It seems to me the barchart and linechart use a different horizontal axis type. I thought the line chart did handle number of ticks properly by itself.
05:52:06 <gjanssens> On the other hand it may also have been oversight...
05:54:22 <chris> I think it has same issue which can be fixed. Just checked on master 2.7.5. this is minor and can wait until I'm done :)
05:55:00 <chris> oh hang on you're right
05:57:15 <chris> oh with linechart it's different - the number of ticks is initialised at report generation and is not recalculated upon window resize. so I think it'd make sense to merge this getVisualTicks() logic into both.
06:02:18 *** jotrago has quit IRC
06:03:13 *** fabior has joined #gnucash
06:21:08 *** ncv_ has joined #gnucash
06:28:32 *** gncbot sets mode: +o fell
06:44:47 *** fabior has quit IRC
06:46:14 *** fabior has joined #gnucash
07:12:55 *** badger92 has quit IRC
07:16:38 *** badger92 has joined #gnucash
07:29:08 *** badger92 has quit IRC
07:34:34 <warlord> .
07:42:59 *** bob_ has joined #gnucash
07:47:42 *** User_ has joined #gnucash
07:58:08 *** bob_ has quit IRC
08:02:05 *** badger92 has joined #gnucash
08:04:11 *** User_ has quit IRC
08:09:23 *** oozer has joined #gnucash
08:36:44 *** chris has quit IRC
08:51:56 *** Jimraehl1 has joined #gnucash
08:53:09 *** Jimraehl1 has left #gnucash
08:54:39 *** fekepp has quit IRC
08:54:43 *** fekepp has joined #gnucash
09:24:51 *** chris has joined #gnucash
09:46:53 *** darshan has joined #gnucash
10:00:37 *** darshan has quit IRC
10:07:29 *** fekepp has quit IRC
10:29:36 *** darshan has joined #gnucash
10:46:12 *** darshan has quit IRC
11:09:55 <chris> a small change to the upgraded html-charts.scm http://i.imgur.com/Q12iDiE.png allows the yaxis to finally format the numbers as currencies :) http://i.imgur.com/8NxmQWS.png
11:12:11 <chris> my vision of a generic chart report is a generic report that directly translates gnucash's data structures onto jqplot API ;-)
11:42:30 *** ArtGravity has joined #gnucash
12:14:37 *** fabior has joined #gnucash
12:57:03 <gjanssens> chris: is that hardcoded to "$" or will it adapt to the chosen report currency ?
13:00:09 *** fabior has quit IRC
13:01:53 *** oozer has quit IRC
13:02:56 *** oozer has joined #gnucash
13:03:53 *** ncv_ has quit IRC
13:07:05 *** harshitaneja has joined #gnucash
13:09:23 *** harshitaneja has joined #gnucash
14:06:59 *** frakturfreak has joined #gnucash
14:09:01 *** harshitaneja has quit IRC
14:18:28 <jralls> chris: I think it would make more sense for the graph report to take a query result rather than connecting directly to gnucash's data structures.
14:31:23 <jralls> ARGH! The stupid build script on unstable failed last night but completed without issues when I ran it by hand just now.
14:41:20 <fell> Jralls, it's lonesome. ;-)
14:45:28 <jralls> fell: Why? we're both here, along with all of the lurkers! ;-)
14:46:03 <fell> The script in the build VM
14:46:45 <jralls> Oh, the *script* is lonesome. So I should set up another scheduled task to keep it company?
14:47:18 * jralls wonders what such a task would do...
14:47:36 *** Mechtilde has joined #gnucash
14:50:24 <fell> BTW: How about I backport "Remove all references to the now-defunct Yahoo! quote retrieval" without what I complained and update the doc?
14:50:59 <jralls> fell: Hmm, the failure in master is interesting. It's in gnucash-docs: warning: failed to load external entity "http://docbook.sourceforge.net/release/xsl/current/htmlhelp/htmlhelp.xsl" .
14:51:54 <fell> I never saw that ref
14:54:53 <fell> It occurs only in chm.make on master
14:56:50 <jralls> Of course. It's used to convert the regular html output to chm for windows.
14:56:52 <fell> In Maint we have "...current/htmlhelp/htmlhelp.xsl"
14:57:36 *** Mechtilde has quit IRC
14:58:41 <fell> oops
15:00:29 <fell> No, there is no chm.make
15:01:42 <fell> you created it on 15.09.2015
15:03:54 <jralls> Right. I wrote it to support building the windows help files in Mingw64. Maybe, considering SourceForge's recent problems, deciding to point to the source over the net instead of keeping a local copy like maint does wasn't so smart.
15:05:53 *** Mechtilde has joined #gnucash
15:14:46 *** Mechtilde has quit IRC
15:17:32 *** bob_ has joined #gnucash
15:23:01 <fell> JRalls, If you fix chm, Andi is written "Köhler", sems to have wrong encoding.
15:23:55 <jralls> Maybe I should just go with Koehler and skip trying to get the encoding right.
15:25:47 <fell> I thought, we have in between everithing in UTF-8?
15:26:28 *** lfs has joined #gnucash
15:27:30 <jralls> Emacs tells me that it's a DOS file, but not what charset it is. FWIW Emacs shows the character correctly as ö.
15:32:50 *** lfs has quit IRC
15:32:53 *** lfs has joined #gnucash
15:35:19 *** lfs has quit IRC
15:42:36 <fell> Eclipse tells me inhereted fromt the container: UTF-8 and shows "�"
15:43:52 *** josephcocoa has joined #gnucash
15:51:39 <jralls> If I leave it as-is it displays in the unix shell (e.g. with less or head) as K<F6>hler. If I change it to utf8 then it displays as K�hler. Neither is very good.
15:52:47 <jralls> github knows what to do with it in Latin1: https://github.com/Gnucash/gnucash-docs/blob/master/chm.make
15:57:51 <jralls> I'm seriously leaning towards Koehler.
16:12:50 <fell> we spend too much time on 1-2 bytes
16:13:22 <warlord> Why is it so hard to display ö?
16:15:51 <fell> Because every program makes other assumtions for a plain text file.
17:12:01 *** pilotauto has joined #gnucash
17:12:32 *** robin has joined #gnucash
17:46:06 *** gjanssens has quit IRC
18:04:28 *** frakturfreak has quit IRC
18:10:35 *** bob_ has quit IRC
18:25:53 <chris> gjanssens - as usual you're right, the $%'d came from jqplot's own examples. but isn't the alist->json utility such a breath of fresh air?
18:26:07 <chris> no more escaping strings, it's inbuilt.
18:26:33 <jralls> chris do an at-tell. gjanssens is off to bed.
18:26:54 <chris> @tell gjanssens - as usual you're right, the $%'d came from jqplot's own examples. but isn't the alist->json utility such a breath of fresh air? no more escaping strings which is built-in
18:26:54 <gncbot> chris: The operation succeeded.
18:27:08 <chris> and also magic+emacs is awesome :)
18:27:53 *** bertbob has quit IRC
18:31:07 <chris> my observations with hacking charts so far
18:31:46 <jralls> chris: I think your work is better described as un-hacking!
18:33:52 <chris> i think *all* charts in menus use xaxis = date and yaxis = currency, so, it'd be good to optimize to those... this means barchart could be converted to send dates to jqplot instead of numbers, and jqplot can use them directly
18:34:12 <chris> also jqplot also looks like it's been hacked together over the years with sometimes inconsistent api ... i may explore some alternatives as an experiment and will report back if something looks much better and consistent
18:42:01 *** bertbob has joined #gnucash
18:42:18 <jralls> chris: OK. You might pause and contemplate the overall reporting architecture to see if a redesign could be done that would allow better ad-hoc reporting without users having to learn to code.
18:50:25 <chris> i was think along those lines -- either continue to add filters (eg add business-type filters) and output formats to the TR (e.g. add invoice items as well, and add json/csv/ods export) which means the TR will become "visual-SQL", or alternatively open up a web server to allow Excel to query a Gnucash session, which is really SQL
18:51:34 <jralls> Huh? A web server to allow Excel to query a Gnucash session? That doesn't make sense.
18:52:09 <jralls> And how is a GnuCash session really SQL?
18:53:47 <chris> well that's the idea - a "report" written in scheme which opens up a webserver port, accepts SQL-type queries and outputs data
18:54:30 <jralls> Sounds like a security hole you could drive an aircraft carrier through.
18:54:37 <chris> yup
18:55:23 <chris> unless the "report" is offered read-only API only.
18:56:37 <chris> that's all I came up with so far
18:57:08 <jralls> And has a a very carefully written parser. But even operating as intended it's a way to leak one's financial data if someone else can connect to the computer.
18:58:30 <jralls> But if one wants SQL access one can already get that, no report required. Just use a SQL backend and query away.
18:58:53 <chris> yups, it's massive duplication of effort
19:08:36 <jralls> So I was going more towards separating the report architecture into query->intermediate format->export-or-presentation.
19:11:59 <jralls> The query engine already exists (it sucks, but that's a separate issue). The intermediate format is the problem, the query engine returns data structures that can't be exported as-is, so we go to custom code for almost every report that post-processes the data structures and emits html or javascript.
19:18:09 <chris> well, TR which does little processing other than subtotals could export ods/json
19:20:25 <chris> but i haven't gone through others thoroughly yet to figure out the very common themes
19:20:59 <jralls> Right, that's why I was suggesting a pause-and-reflect.
20:06:09 *** ArtGravity has quit IRC
20:22:03 *** luc14n0 has left #gnucash
20:44:53 *** oozer has quit IRC
21:32:42 *** luc14n0 has joined #gnucash
22:41:20 <fell> jralls: Asia and Brazil disappeared with Yahoo, the rest still exists.
22:50:10 *** storyjesse has joined #gnucash
23:19:25 *** mipesom has quit IRC
23:36:50 *** mipesom has joined #gnucash