2019-07-05 GnuCash IRC logs
00:27:02 *** ShaunKL has quit IRC
00:44:33 *** Mechtilde has joined #gnucash
00:44:33 *** ChanServ sets mode: +v Mechtilde
01:35:23 *** O01eg has quit IRC
01:49:42 *** Mechtilde has quit IRC
02:24:51 *** gjanssens has joined #gnucash
02:24:51 *** ChanServ sets mode: +o gjanssens
02:26:35 *** gnomey has quit IRC
02:32:38 *** gnomey has joined #gnucash
02:59:21 *** fabior has joined #gnucash
03:16:40 *** fabior has quit IRC
03:23:53 *** gour has joined #gnucash
03:23:53 *** ChanServ sets mode: +v gour
03:26:17 *** fabior has joined #gnucash
03:33:29 *** fabior has quit IRC
03:42:30 *** fabior has joined #gnucash
04:04:23 *** Mechtilde has joined #gnucash
04:04:23 *** ChanServ sets mode: +v Mechtilde
05:01:00 *** Aussie_matt has quit IRC
05:16:41 *** Mechtilde has quit IRC
05:27:03 *** mr_sm11th has joined #gnucash
05:27:04 *** ChanServ sets mode: +v mr_sm11th
06:36:58 *** oozer has joined #gnucash
06:37:16 *** Jimraehl1 has joined #gnucash
06:37:54 *** Jimraehl1 has left #gnucash
06:47:27 *** boldstripe has joined #gnucash
06:50:21 *** oozer has quit IRC
06:51:00 *** oozer has joined #gnucash
07:00:43 *** fabior has quit IRC
07:32:39 *** jervin has joined #gnucash
07:53:02 *** Mechtilde has joined #gnucash
07:53:02 *** ChanServ sets mode: +v Mechtilde
08:10:26 <warlord> .
08:14:15 * chris been experimenting with guile threads for computation-heavy price-calculators... no joy
08:16:23 <warlord> Heh
08:16:36 <warlord> Heavy computation will always be faster in C
08:16:41 <warlord> (well written C)
08:16:59 <chris> !@#$
08:17:42 <chris> I think a lot of price-calculator slowdown has nothing to do with loops
08:19:41 <chris> jralls: here's a quick x15 speedup for complicated books price-calculators
08:20:00 <chris> modified gnucash/report/report-system/commodity-utilities.scm
08:20:00 <chris> @@ -801,9 +801,9 @@ construct with gnc:make-gnc-monetary and gnc:monetary->string instead.")
08:20:00 <gncbot> chris: Error: "@" is not a valid command.
08:20:00 <chris> (define (gnc:exchange-by-pricealist-nearest
08:20:00 <chris> pricealist foreign domestic date)
08:20:03 <chris> ;;Used in weighted-average gnc:case-exchange-time-fn only.
08:20:06 <chris> - (gnc:debug "foreign " (gnc:monetary->string foreign))
08:20:09 <chris> - (gnc:debug "domestic " (gnc-commodity-get-printname domestic))
08:20:12 <chris> - (gnc:debug "pricealist " pricealist)
08:20:15 <chris> + ;; (gnc:debug "foreign " (gnc:monetary->string foreign))
08:20:18 <chris> + ;; (gnc:debug "domestic " (gnc-commodity-get-printname domestic))
08:20:21 <chris> + ;; (gnc:debug "pricealist " pricealist)
08:20:24 <chris> (and (record? foreign)
08:20:26 <chris> (gnc:gnc-monetary? foreign)
08:20:29 <chris> date
08:20:33 <chris>
08:20:39 <chris> sorry this deserves spamming the irc log!!!
08:21:55 <chris> removing gnc:debug speeds up a complicated book price calculators from 15s to 1.5s :-o
08:22:06 <chris> w.t.h...
08:24:51 <warlord> Wha?!?!?
08:27:07 <chris> https://pastebin.com/raw/0jTvy26r
08:42:35 <chris> jralls: wish your instrumentation skills on average-balance.scm before and after removing gnc:debug calls :)
09:08:52 <warlord> chris, Could you comment them out one-by-one and see which one is the time hog?
09:14:30 *** jervin has quit IRC
09:18:34 <chris> warlord: all of them
09:19:49 <warlord> So it goes from 1.5s to 15s with ANY single one of those lines?
09:19:59 <chris> ah no it's incremental
09:20:03 <warlord> Or does each line add 5s?
09:23:50 <chris> no third one's the worst
09:25:48 <chris> 1.5s -> 2.3s or 3.8s with either of first two, 1.5s -> 13s with third
09:26:18 <chris> because average-balance will add all monetaries for *every date* hence most vulnerable to exchanging-fn
09:27:56 <warlord> And now you know! :)
09:29:45 *** Mechtilde has quit IRC
09:30:19 <chris> is the (gnc:make-commodity-collector) from your time? i think it's rather ugly... prefer purely functional approach (apply + (map gnc:gnc-monetary-amount list-of-mon))
09:33:37 <warlord> It predates me
09:53:27 *** oozer has quit IRC
10:00:44 *** gour has quit IRC
10:01:59 *** gour has joined #gnucash
10:23:54 *** jervin has joined #gnucash
10:36:33 <warlord> gjanssens, flatpak maint build died July 3?
10:41:33 <gjanssens> Oh, didn't notice...
10:41:39 <gjanssens> I'll see if I can fix it.
10:41:53 <warlord> Thanks
10:54:36 *** oozer has joined #gnucash
10:59:46 *** jervin has joined #gnucash
11:08:36 *** boldstripe has quit IRC
11:09:30 *** boldstripe has joined #gnucash
11:39:57 *** Robert847 has joined #gnucash
11:39:57 *** ChanServ sets mode: +v Robert847
11:45:49 <Robert847> Hello. I would like to ask about the View > Refresh function as it works in release 3.6. In release 2.6.19 in Ubuntu 18.04, if I create a new sub-account by entering a new name in the account field of an exisiting transaction then look in the 'with sub-accounts' view of the parent account, that transaction is missing. The View > Refresh menu does not pick up the new sub-account.
11:47:05 <Robert847> Is this worth creating a bug report or has it already been addressed?
11:47:24 <jralls> chris, the problem isn't so much the gnc:debug as evaluating what's passed to it. If you can figure out a way to prevent evaluation unless the log domain 'gnc.scm' is set to 'debug' then we can get the speed up without losing the ability to get tracing information when a user has a problem.
11:49:27 <jralls> Robert847: Have you committed the transaction?
11:50:42 <Robert847> Yes, and saved the file and I am in the 'with sub-accounts' window when I click View > Refresh
11:51:01 <chris> jralls: what's the C code to check log-domain set to 'debug' ?
11:51:38 <Robert847> Oh, I also have opened the new child account and verified that the transaction appears there
11:55:23 <jralls> chris: quof_log_check(QofLogModule log_domain, QofLogLevel log_level);
11:56:22 <jralls> oops, s/quof/qof/
11:59:41 *** fabior has joined #gnucash
12:00:27 <chris> qof-log-check isn't wrapped :(
12:00:49 <jralls> So wrap it.
12:01:38 <jralls> Or duplicate it in Scheme.
12:02:04 <jralls> Scratch that, it needs access to the log-level hash.
12:02:25 <jralls> Only wrapping will work.
12:05:53 <sunbird> jralls: hey, thanks for the response. yes, the matcher page is blank. i've never had this issue with any account before. transactions were definitely not imported before.
12:06:47 *** jonp` has quit IRC
12:07:25 <chris> sunbird: check acct-currency = ofx-currency
12:07:32 <jralls> sunbird: Then check the FITID field in the OFX file and compare to previous OFX files from that bank. They're probably reusing numbers. We've seen instances of that in the past. It's not permitted by the OFX spec.
12:15:39 *** jervin has quit IRC
12:16:40 *** fabior has quit IRC
12:17:44 <jralls> Robert847: Still the case in 3.6, so file a bug (after checking that there isn't one already, of course).
12:19:11 <Robert847> Thank you. Will do.
12:23:07 <chris> jralls: i can't do it sorry
12:23:25 <chris> %include "qoflog.h" at least
12:26:29 <chris> off to bed now
12:34:12 *** chris has quit IRC
12:34:17 *** chris has joined #gnucash
12:34:17 *** ChanServ sets mode: +v chris
12:34:20 *** jervin has joined #gnucash
12:57:09 <sunbird> chris jralls: thx. i'll check that when i get home.
13:02:46 <Robert847> sunbird: If I recall, GnuCash remembers that a transactiion is imported even if later deleted, so it is still impossible to re-import it.
13:03:51 *** chris has quit IRC
13:04:25 *** chris has joined #gnucash
13:04:25 *** ChanServ sets mode: +v chris
13:09:16 *** boldstripe has quit IRC
13:10:12 *** boldstripe has joined #gnucash
13:12:16 *** jervin has quit IRC
13:24:18 *** calvinct has joined #gnucash
13:30:24 *** Robert847 has left #gnucash
13:35:04 *** calvinct has quit IRC
13:38:02 *** Mechtilde has joined #gnucash
13:38:02 *** ChanServ sets mode: +v Mechtilde
13:42:36 *** Mechtilde has quit IRC
13:44:17 *** Mechtilde has joined #gnucash
13:44:17 *** ChanServ sets mode: +v Mechtilde
14:06:34 *** frakturfreak has joined #gnucash
14:18:17 *** gour1 has joined #gnucash
14:19:59 *** gour has quit IRC
14:41:26 *** boldstripe has quit IRC
14:49:15 *** boldstripe has joined #gnucash
15:04:36 *** frakturfreak has quit IRC
15:06:32 *** frakturfreak has joined #gnucash
15:06:33 *** ChanServ sets mode: +v frakturfreak
15:23:34 *** frakturfreak has quit IRC
15:26:12 *** frakturfreak has joined #gnucash
15:26:13 *** ChanServ sets mode: +v frakturfreak
15:32:09 *** boldstripe has quit IRC
15:45:25 *** User_ has joined #gnucash
15:55:30 *** boldstripe has joined #gnucash
15:57:02 *** User_ has quit IRC
16:19:03 *** boldstripe has quit IRC
16:19:08 <gjanssens> warlord: I see there are a few other flatpak maint builds that failed mysteriously
16:19:35 <gjanssens> You can find them by looking for left-over start logs here https://code.gnucash.org/builds/flatpak/build-logs/
16:20:03 <gjanssens> I don't know why that happens as there's no log info for me to look at
16:20:11 <gjanssens> And later builds work again.
16:20:47 <gjanssens> Do you find local log files for the failed builds ?
16:21:12 <gjanssens> Perhaps they have more info but the build script failed to upload them ?
16:21:46 <gjanssens> I also find the master flatpak builds are currently failing due to some chart components not being found.
16:21:57 <gjanssens> To look at when I have a bit of spare time.
16:22:35 <warlord> gjanssens, I just get an email that says the build failed.
16:22:45 <warlord> I got ~3 of them over the past week or so.
16:25:06 *** guak has joined #gnucash
16:25:20 <gjanssens> There are for on the link I gave, one from January (probably while we were testing), two from June and one from July
16:26:09 <gjanssens> Perhaps that's just due to your limited network bandwidth, some download timing out or something like that ?
16:26:28 *** marlon has joined #gnucash
16:32:58 *** jonp has joined #gnucash
16:50:49 <warlord> could be. Or a DNS outage, or something.
16:52:49 *** boldstripe has joined #gnucash
16:53:11 *** KevinDB has quit IRC
16:54:35 *** KevinDB has joined #gnucash
16:54:35 *** ChanServ sets mode: +v KevinDB
16:57:39 *** gjanssens has quit IRC
17:50:29 *** boldstripe has quit IRC
18:02:27 *** boldstripe has joined #gnucash
18:11:28 *** ArtGravity has joined #gnucash
18:11:28 *** ChanServ sets mode: +v ArtGravity
18:27:50 *** gour1 has quit IRC
18:47:52 *** frakturfreak has quit IRC
18:55:23 *** boldstripe has quit IRC
18:55:29 *** jervin has joined #gnucash
18:57:41 *** jervin has quit IRC
19:02:00 *** jervin has joined #gnucash
19:17:39 *** jervin has quit IRC
19:35:10 *** bertbob has quit IRC
19:35:29 *** jervin has joined #gnucash
19:37:49 *** bertbob has joined #gnucash
19:37:49 *** ChanServ sets mode: +v bertbob
19:40:09 *** guak has quit IRC
19:44:54 *** mr_sm11th has quit IRC
21:04:44 *** marlon has quit IRC
21:49:06 *** oozer has quit IRC
22:24:46 *** fell has quit IRC
22:47:59 *** chris has quit IRC
23:12:29 *** fell has joined #gnucash
23:12:29 *** ChanServ sets mode: +o fell
23:19:00 *** chris has joined #gnucash
23:19:00 *** ChanServ sets mode: +v chris
23:19:10 <chris> https://github.com/christopherlam/gnucash/commits/wrapqoflog please see commit 'wrap qoflog' -- errors when exporting QOF-LOG-DEBUG or G-LOG-LEVEL-DEBUG -- what's wrong?
23:35:34 *** chris has quit IRC
23:36:18 *** chris has joined #gnucash
23:36:18 *** ChanServ sets mode: +v chris
23:36:22 <chris> jralls if you wish to experiment with multicore price-calculators I've done a branch at https://github.com/christopherlam/gnucash/tree/maint-parallelism