2017-05-06 GnuCash IRC logs
01:02:49 *** fell_ has joined #gnucash
01:05:05 *** fell has quit IRC
01:11:11 *** Mechtilde has joined #gnucash
01:13:17 *** pilotauto has joined #gnucash
01:44:23 *** chf has quit IRC
01:46:00 *** chf has joined #gnucash
02:15:34 *** gjanssens has joined #gnucash
02:15:34 *** ChanServ sets mode: +o gjanssens
02:19:02 *** Mechtilde has quit IRC
02:20:39 *** Mechtilde has joined #gnucash
02:55:03 *** Mechtilde has quit IRC
03:00:40 *** Mechtilde has joined #gnucash
03:08:02 *** Mechtilde has quit IRC
03:08:03 *** mrklintscher has quit IRC
03:09:46 *** mrklintscher has joined #gnucash
03:23:06 *** fabior has joined #gnucash
03:33:15 *** mrklintscher has quit IRC
03:34:27 *** mrklintscher has joined #gnucash
03:36:03 *** sixwheeledbeast has quit IRC
03:40:03 *** sixwheeledbeast has joined #gnucash
03:44:01 *** jotrago has joined #gnucash
03:55:26 *** federvie1 has quit IRC
03:55:37 *** chf has quit IRC
03:57:45 *** federvieh has joined #gnucash
03:58:24 <gjanssens> .
03:58:24 <gncbot> gjanssens: Sent 14 hours and 38 minutes ago: <lmat> how is your timespec work going? Completely stalled. I ran into a few walls, and could not see my way out. In fact, at this point, I would probably start over and make smaller incremental changes.
03:58:25 <gncbot> gjanssens: Sent 14 hours and 7 minutes ago: <jralls> lmat codesmythe Ninja takes -j just like make does.
03:58:26 <gncbot> gjanssens: Sent 13 hours and 54 minutes ago: <jralls> We build our own Guile on autotools because Guile-1.8 and Guile-2.0 have different autoconf macros and we couldn't get them to do the right thing. That's probably not a problem for CMake.
03:59:50 <gjanssens> jralls: thanks for refreshing my memory. OTOH building our own guile executable would probably also have avoided the mix of 2.0 cflags with 2.2 executable.
04:00:32 *** chf has joined #gnucash
04:01:20 <gjanssens> lmat: ok. Do you intend to do so as well (I mean start over) ?
04:05:56 *** federvieh has quit IRC
04:14:22 *** federvieh has joined #gnucash
04:17:07 *** sixwheeledbeast has quit IRC
04:17:59 *** sixwheeledbeast has joined #gnucash
04:26:17 * gjanssens finished reading the remainder of yesterday's communication
04:26:35 <gjanssens> lmat: as jralls said: feel free to share your work early and often!
04:37:12 *** gjanssens has quit IRC
04:51:40 *** gjanssens has joined #gnucash
04:51:40 *** ChanServ sets mode: +o gjanssens
05:21:30 *** gjanssens has quit IRC
05:35:08 *** gjanssens has joined #gnucash
05:35:08 *** ChanServ sets mode: +o gjanssens
05:55:16 *** mrklintscher has quit IRC
05:56:45 *** mrklintscher has joined #gnucash
06:05:17 *** pilotauto has quit IRC
06:31:13 *** mrklintscher has quit IRC
06:32:31 *** mrklintscher has joined #gnucash
06:55:20 *** federvieh has quit IRC
06:55:20 *** Artefact2 has quit IRC
06:55:20 *** GabrieleV has quit IRC
06:55:20 *** xmaka has quit IRC
06:55:20 *** Unhammer has quit IRC
06:55:20 *** weasel has quit IRC
06:55:20 *** mrklintscher has quit IRC
06:55:20 *** sixwheeledbeast has quit IRC
06:55:20 *** jotrago has quit IRC
06:55:20 *** fabior has quit IRC
06:55:20 *** jchonig has quit IRC
06:55:20 *** warlord has quit IRC
06:55:20 *** Primus has quit IRC
06:55:20 *** Derperperd has quit IRC
06:55:20 *** lmat has quit IRC
06:55:20 *** ian has quit IRC
06:55:20 *** Cyp_ has quit IRC
06:55:20 *** cyphase has quit IRC
06:55:20 *** jralls has quit IRC
06:55:20 *** CDB-Away_ has quit IRC
06:55:20 *** raeburn1 has quit IRC
06:55:26 *** mrklintscher has joined #gnucash
06:55:26 *** sixwheeledbeast has joined #gnucash
06:55:26 *** federvieh has joined #gnucash
06:55:26 *** jotrago has joined #gnucash
06:55:26 *** fabior has joined #gnucash
06:55:26 *** jchonig has joined #gnucash
06:55:26 *** warlord has joined #gnucash
06:55:26 *** Primus has joined #gnucash
06:55:26 *** Derperperd has joined #gnucash
06:55:26 *** Artefact2 has joined #gnucash
06:55:26 *** lmat has joined #gnucash
06:55:26 *** ian has joined #gnucash
06:55:26 *** Cyp_ has joined #gnucash
06:55:26 *** GabrieleV has joined #gnucash
06:55:26 *** cyphase has joined #gnucash
06:55:26 *** xmaka has joined #gnucash
06:55:26 *** jralls has joined #gnucash
06:55:26 *** CDB-Away_ has joined #gnucash
06:55:26 *** Unhammer has joined #gnucash
06:55:26 *** raeburn1 has joined #gnucash
06:55:26 *** weasel has joined #gnucash
06:55:26 *** irc.eagle.y.se sets mode: +oo warlord jralls
07:01:58 *** mrklintscher has quit IRC
07:03:12 *** mrklintscher has joined #gnucash
07:19:43 *** fabior has quit IRC
07:22:23 *** chf has quit IRC
07:30:40 *** chf has joined #gnucash
07:45:40 *** rickoehn has joined #gnucash
07:58:53 *** Jimraehl1 has joined #gnucash
08:00:32 *** Jimraehl1 has left #gnucash
08:23:04 *** User has joined #gnucash
09:03:06 *** mikee_ has joined #gnucash
09:03:35 *** mikee has quit IRC
09:03:42 *** mikee_ is now known as mikee
09:03:42 *** ChanServ sets mode: +o mikee
09:23:04 *** User has quit IRC
09:29:31 *** User has joined #gnucash
09:38:32 *** fabior has joined #gnucash
10:15:09 *** codesmythe has joined #gnucash
10:33:00 *** mrklintscher has quit IRC
10:34:16 *** mrklintscher has joined #gnucash
10:43:19 *** mlncn has joined #gnucash
10:54:31 <gjanssens> jralls: with the GncDateTime string constructors done, I'm now looking into the GncNumeric string constructor to make it more useful for the csv importer (and other importers in the future)
10:57:08 *** User has quit IRC
11:00:37 <gjanssens> Hmm, I'll get back to this later. Something else came up...
11:10:40 *** mrklintscher has quit IRC
11:12:00 *** mrklintscher has joined #gnucash
11:31:39 *** codesmythe has quit IRC
11:35:23 *** mrklintscher has quit IRC
11:36:43 *** mrklintscher has joined #gnucash
12:22:18 *** mrklintscher has quit IRC
12:23:28 *** mrklintscher has joined #gnucash
12:49:58 *** rickoehn has quit IRC
12:56:05 *** rickoehn has joined #gnucash
13:56:13 *** Mechtilde has joined #gnucash
13:56:52 <jralls> gjanssens: OK.
13:57:33 <jralls> gjanssens, warlord: I'll be away next week, I'll check email at least once a day but won't be on IRC.
13:58:07 <gjanssens> jralls: me too :)
13:58:22 <gjanssens> I'll have less time to check my mail even...
13:58:32 <gjanssens> I'll be away mon-fri
14:00:08 <jralls> I'm leaving oh-dark-thirty Monday morning and will be back Sunday afternoon.
14:04:47 *** mlncn has quit IRC
14:05:03 *** frakturfreak has joined #gnucash
14:05:52 *** mlncn has joined #gnucash
14:11:10 *** mlncn has quit IRC
14:20:42 *** jotrago1 has joined #gnucash
14:21:14 *** jotrago has quit IRC
14:21:15 *** jotrago1 is now known as jotrago
14:24:28 *** jotrago has quit IRC
14:26:33 *** jotrago has joined #gnucash
14:55:56 *** mrklintscher has quit IRC
14:57:10 *** mrklintscher has joined #gnucash
15:07:00 *** Mechtilde has quit IRC
15:09:33 *** User has joined #gnucash
15:12:27 *** mrklintscher has quit IRC
15:13:54 *** mrklintscher has joined #gnucash
15:14:12 *** fabior has quit IRC
15:15:58 <gjanssens> jralls: what I meant to bring up regarding the GncNumeric string constructor
15:16:32 <gjanssens> I'm working on an implementation that does handle thousands separators and locale dependent decimal points
15:16:54 <gjanssens> It's mostly working, barring one regex that's not doing what I expect.
15:17:09 <gjanssens> I'll figure it out but not tonight.
15:17:48 <gjanssens> I did want to show you what I've got so far for your feedback.
15:18:31 <gjanssens> I've pushed to my github repo here: https://github.com/gjanssens/gnucash/tree/gnc_numeric_str_constr
15:19:05 <gjanssens> The string constructor tests all work except for the ones using a "," as decimal point.
15:20:07 <gjanssens> What I did was add an overloaded function that also accepts a locale as parameter and use that one to extract the ts and dp from
15:20:28 <gjanssens> The simple function just calls the overloaded one with the system locale.
15:21:28 <gjanssens> The current issue with the , dp also breaks the input stream operator which is to be expected as it simply calls on the string constructor.
15:22:45 <gjanssens> While working on it I also noted the output operator also extracts the "," separator from the locale to pretty print a number.
15:23:43 <gjanssens> Oh no. I was going to say something stupid. It uses the locale from the stream which is perfect.
15:24:54 <gjanssens> I intend to add the same to the input stream operator (that is get the locale from the stream object) and pass it to the string constructor. That way both will always have consitent results.
15:24:56 *** youssef has joined #gnucash
15:25:53 <gjanssens> Last thing I intend to do still is to improve the input stream operator to be a bit smarter about rationals with whitespace around the /
15:26:31 <gjanssens> I intend to do a first read operation
15:26:55 <gjanssens> If the read string already has a / the string will be passed to the string constructor at once
15:27:31 <gjanssens> If not, the function should look ahead in the stream until it finds the next non-whitespace character.
15:28:06 *** youssef has quit IRC
15:28:59 <gjanssens> If that happens to be a /, a second read operation is initiated after that / and the combined result of the first read the slash and a second read are passed to the string constructor
15:29:40 <gjanssens> Leaves me only with the case of a space as thousands separator.
15:30:28 <gjanssens> The curious thing is if you imbue a locale that has such a ts and stream into a numeric variable (an int, double,...) the stream operator will correctly handle this.
15:30:45 <gjanssens> It's only because we stream into a string this isn't working.
15:31:16 <gjanssens> So I'm still mulling how to handle this exactly.
15:31:35 <gjanssens> But I have a feeling I'm close to solving this as well :)
15:32:50 <gjanssens> Oh, and if you don't have time to look at it before you leave, that's obviously no problem. I won't have much time either myself to work on it until I get back.
15:33:54 <gjanssens> Last remark: you'll find the string constructor decomposed in a couple of smaller functions.
15:34:12 <gjanssens> I felt that made for less regexes.
15:35:02 <gjanssens> The main constructor splits the string based on a / (accepting optional surrounding whitespace)
15:35:39 <gjanssens> The numerator is mandatory. If missing the function throws. The denominator is optional.
15:36:03 <gjanssens> Both parts are then treated similarly to get them converted to real numbers
15:37:21 *** mrklintscher has quit IRC
15:37:42 <gjanssens> Both are passed on to parse_num_str to be split in an integral and optional decimal part.
15:38:11 <gjanssens> (And while writing this, I see I have to to the is_hex test in this function rather than in the next one)
15:38:57 *** mrklintscher has joined #gnucash
15:39:01 <gjanssens> Each of these parts if found are further processed in part_int_str, which does the actual conversion via a stream operator (to handle thousands separators according to locale).
15:39:26 <gjanssens> The results are then combined when returning to the respective calling functions.
15:42:52 <gjanssens> All the way back up to the string constructor function itself.
15:47:40 *** kael has joined #gnucash
15:48:54 *** codesmythe has joined #gnucash
16:04:08 *** jonas has quit IRC
16:05:44 *** gjanssens has quit IRC
16:09:11 *** jonas has joined #gnucash
16:16:16 <codesmythe> jralls: I'm unable to reproduce the pthread problem on Win10.
16:22:57 <codesmythe> jralls: Nevermind ... find an error in my setup. Still trying.
16:42:30 *** codesmythe has left #gnucash
16:49:27 *** codesmythe has joined #gnucash
16:54:06 <jralls> codesmythe: Did you get it to reproduce?
16:54:10 *** kael has quit IRC
16:54:13 *** kael has joined #gnucash
16:57:25 *** mrklintscher has quit IRC
16:58:41 *** mrklintscher has joined #gnucash
17:20:17 *** codesmythe has quit IRC
17:22:32 *** karelk has quit IRC
17:22:33 *** mrklintscher has quit IRC
17:24:43 *** mrklintscher has joined #gnucash
17:42:27 *** codesmythe has joined #gnucash
17:43:47 *** mlncn has joined #gnucash
17:52:23 <codesmythe> jralls: Yep, just did.
17:57:43 *** mrklintscher has quit IRC
17:59:08 *** mrklintscher has joined #gnucash
18:22:26 *** mrklintscher has quit IRC
18:23:50 *** mrklintscher has joined #gnucash
18:36:14 *** mlncn has quit IRC
18:47:56 *** kael has quit IRC
19:02:28 *** karelk has joined #gnucash
19:36:20 *** User__ has joined #gnucash
19:36:34 *** User has quit IRC
19:43:04 <codesmythe> jralls: Error is caused by inlining of libguile function 'scm_double_cell' when optimization is enabled. Using -fno-inline in CMAKE_C(XX)?_FLAGS_RELEASE does get past the error. I don't know if that is too blunt an instrument or if we just want to link in -lpthreads. Also, there's an uninitialized variable error in src/engine/gnc-pricedb.c:2276. With that fixed and -fno-line, GnuCash builds with build type Release.
20:10:53 *** codesmythe has quit IRC
20:13:18 *** federvie1 has joined #gnucash
20:15:17 *** federvieh has quit IRC
20:37:34 *** CDB-Away has joined #gnucash
20:38:27 *** CDB-Man_ has joined #gnucash
20:39:47 *** CDB-Man has quit IRC
20:40:13 *** CDB-Away_ has quit IRC
20:40:43 *** codesmythe has joined #gnucash
20:45:09 *** rickoehn has quit IRC
20:51:14 *** frakturfreak has quit IRC
20:53:59 <jralls> codemsythe: Ah, very good. But why does not defining CMAKE_BUILD_TYPE at all not inline the function (and presumably others)?
20:55:32 <jralls> codesmythe: And yes, inlining is good, so if the tradeoff is that we have to add -lpthread to all of the swigged libs then we should do that.
21:02:12 *** User__ has quit IRC
21:07:00 *** fell_ has quit IRC
21:28:02 *** mlncn has joined #gnucash
22:00:13 *** User has joined #gnucash
22:05:52 *** mlncn has quit IRC
22:30:54 *** User has quit IRC
23:03:43 *** mrklintscher has quit IRC
23:04:56 *** mrklintscher has joined #gnucash
23:33:33 *** mrklintscher has quit IRC
23:35:01 *** mrklintscher has joined #gnucash