2020-05-31 GnuCash IRC logs

00:13:12 *** jervin has quit IRC
00:20:21 *** FH_thecat has joined #gnucash
01:12:18 *** omnireq__ has quit IRC
01:12:29 *** omnireq__ has joined #gnucash
01:15:32 *** marusich has quit IRC
01:21:24 *** fell has quit IRC
01:22:44 *** fell has joined #gnucash
01:22:44 *** ChanServ sets mode: +o fell
01:24:03 *** storyjesse has joined #gnucash
01:47:21 *** jervin has joined #gnucash
02:00:49 *** omnireq__ has quit IRC
02:01:00 *** omnireq__ has joined #gnucash
02:08:41 *** FH_thecat has quit IRC
02:20:08 *** suukim has joined #gnucash
02:30:59 *** Mechtilde has joined #gnucash
03:03:19 *** Aussie_matt has joined #gnucash
03:06:29 *** Han has joined #gnucash
03:26:35 *** gjanssens has joined #gnucash
03:26:35 *** ChanServ sets mode: +o gjanssens
03:26:43 <gjanssens> .
03:26:43 <gncbot> gjanssens: Sent 9 hours and 59 minutes ago: <jralls> Moving set_win32_thread_locale() to gnucash-windows-locale() worked as a workaround. I pushed that and the relocation of gi18n.h to C linkage land. That leaves the change to include/c++/cmath.
03:26:44 <gncbot> gjanssens: Sent 9 hours and 57 minutes ago: <jralls> I also pushed updates to gnucash.modules and setup_mingw32.ps1 for the new guile and webkit.
03:34:09 <gjanssens> maint nightly build on windows got stuck again. I have killed it for the master build to proceed.
03:35:49 *** lmat has quit IRC
03:36:27 *** Aussie_matt has quit IRC
03:38:32 *** lmat has joined #gnucash
04:16:43 <chris> gjanssens: would you mind explaining in simple terms why guile & webkit are so troublesome when building?
04:17:55 <chris> (not saying I can be of any help...)
04:23:38 <chris> gjanssens: I've just tested: running gnucash-cli --run-report=XXX on an open book will delete the .LCK file
04:27:19 *** finster has quit IRC
04:32:19 <chris> this is because the book is opened with qof_session_begin(.., .., TRUE, .., ..) -- changing it to FALSE means it won't open a locked book. I don't know how to open a read_only session.
04:33:40 *** User_ has joined #gnucash
04:59:48 *** keiffer has joined #gnucash
05:28:48 *** User_ has quit IRC
05:41:42 *** lmat has quit IRC
05:44:47 *** lmat has joined #gnucash
05:56:26 <chris> re: the ignore_lock parameter in qof_session_begin should be changed to FALSE to be much safer.
05:58:59 <chris> jralls/gjanssens: I'm considering the next-gen query mechanism for reports. One burgeoning idea comes from the cli-reports. Instead of input from report-options and output HTML, it could input JSON, access all book data (accounts/transactions/documents/lots) and output JSON.
05:59:32 <chris> then user can choose a templating language of choice eg mustache or handlebarsjs for presentation
06:01:43 <chris> the main difficulty would be the input JSON file - how to serialize objects?
06:02:19 *** storyjesse has quit IRC
06:02:27 *** storyjesse has joined #gnucash
06:04:11 <chris> gnucash-cli --query=query.json --output-file=data.json
07:03:14 *** lmat has quit IRC
07:04:09 *** lmat has joined #gnucash
07:19:55 *** Han has quit IRC
07:20:24 *** Han has joined #gnucash
07:26:03 *** User_ has joined #gnucash
08:01:18 *** keiffer has quit IRC
08:04:21 *** Han has quit IRC
08:12:27 *** keiffer has joined #gnucash
08:21:07 *** Han has joined #gnucash
08:24:49 *** keiffer has quit IRC
08:25:52 *** Han has quit IRC
08:26:37 *** storyjesse has quit IRC
08:28:13 *** storyjesse has joined #gnucash
08:29:14 <gjanssens> chris: isn't that what the whole c++ migration is aimed at ? That data can be queried directly ? Trying to do that now will probably lead to duplicate work eventually.
08:31:04 <chris> ok.
08:36:58 <chris> what do you think about deselecting ignore_lock ? it'll be safer. but won't be able to access an already opened book.
08:59:05 *** chf has quit IRC
09:01:31 *** chf has joined #gnucash
09:05:55 *** sbluhm has joined #gnucash
09:05:55 *** ChanServ sets mode: +v sbluhm
09:13:33 <chris> gjanssens: I'd vote to disallow opening locked datafiles. Otherwise there's potential dataloss if: (1) datafile is open via GUI (2) datafile is open via CLI (3) datafile is open in another GUI session (4) first session modifies book (5) second session modifies book.
09:13:51 <chris> unfortunately there is no formal readonly session and I wouldn't want to be responsible for dataloss as above
09:16:51 <warlord> The UI has an "open read-only" option.. So there would seem to be a way?
09:19:59 <chris> warlord: it's not reliable at all: it only sets a read_only bool which is queried by various data modifiers
09:20:09 <chris> even then, inconsistently.
09:20:42 <chris> (e.g. still now or in the past, read_only isn't respected by budget editor)
09:26:54 *** sbluhm has quit IRC
09:40:02 *** FH_thecat has joined #gnucash
09:40:17 <warlord> :(
09:40:41 <chris> Indeed. :( is better than >:(
09:53:18 *** sbluhm has joined #gnucash
09:53:18 *** ChanServ sets mode: +v sbluhm
09:56:18 *** sbluhm has quit IRC
10:19:16 <chris> just checked: the budget editor doesn't respect (read_only) flag
10:28:41 *** sbluhm has joined #gnucash
10:41:35 *** FH_thecat has quit IRC
10:51:33 <gjanssens> chris: that is a bug then
10:51:53 <gjanssens> But I agree the read-only feature was bolted on as an afterthought
10:52:33 <gjanssens> So until we can safely open a book read-only, I'm ok with refusing to do so in gnucash-cli
10:53:08 <gjanssens> We could of course ask the user, but I think the cli tool will be used unattended (in cron jobs).
10:57:28 <gjanssens> jralls: master on Windows did build successfully, but the installer is missing boost program options dlls.
10:58:33 <gjanssens> I'm mostly afk today and tomorrow so I hope you can take it up. Otherwise I could look on Tuesday.
11:08:27 <jralls> chris, re: c++options io, yes, that's done. I don't understand what you mean about "custom scheme reader".
11:08:51 <jralls> chris, thanks for the summary, I'll add that to the release note.
11:09:16 <chris> gjanssens: glad we're on the same page
11:09:38 <chris> jralls: a custom scheme reader would (read) the saved-reports and interpret options one by one... I guess you've already done the work then.
11:11:09 <jralls> fell, Re gnucash-docs do you mean the doc directory in gnucash.git or Help and the T&CG?
11:13:20 *** bertbob has quit IRC
11:14:25 *** sbluhm has quit IRC
11:16:31 <jralls> chris, json is not a query language. There's already an almost universally used one, SQL. Let's use that instead of running off and inventing wheels.
11:18:31 *** bertbob has joined #gnucash
11:18:31 *** ChanServ sets mode: +v bertbob
11:19:37 <jralls> gjanssens, Adding the program-options dll to the installer is a one-liner in gnucash-mingw64.iss, N.P.
11:20:17 <jralls> chris, yes, reading and writing scheme options is already in there.
11:24:32 <fell> Which context, jralls?
11:25:40 <jralls> fell, you said yesterday (my yesterday, anyway) that OpenSuSE had stopped shipping gnucash-docs.
11:26:21 <fell> Their last releaseof the gnucash-docs package was 3.3
11:27:06 <fell> I left a comment about the changed SF download directory
11:27:58 <fell> before we had the docs on SF in a separate directory
11:28:16 <jralls> Ah, got it. They probably have a script that needs tweaking.
11:30:42 <jralls> chris, gjanssens, chris pushed a commit a little while ago replacing a g_warning call with fprintf "because g_warning doesn't actually report error". Is qoflog not included in gnucash-cli?
11:31:25 <jralls> Or does that line happen before logging is started?
11:39:03 <jralls> chris, read-only is enforced in the backends. Even if the budget changes something it won't get written out if the session is read-only.
11:43:11 <jralls> chris, however it is a bug if it's removing the lock file for a read-only open.
12:04:33 *** fell has quit IRC
12:07:08 *** Gerd has quit IRC
12:07:18 *** sbluhm has joined #gnucash
12:07:58 *** fell has joined #gnucash
12:07:58 *** ChanServ sets mode: +o fell
12:08:28 *** Gerd has joined #gnucash
12:13:18 *** sbluhm has quit IRC
12:15:52 <chris> jralls: qoflog may be included but g_warning wasn't outputting anything. this small commit can be reverted easily. Moreover the g_warning/fprintf for qof_session_get_error_message wasn't outputting anything either because qofsessions's push_error was null for LOCKED_FILE
12:16:22 <jralls> chris, I see the problems. The logic is a bit too tangled to fix right away, but I'll work on that next week.
12:16:25 <chris> jralls: ditto lock file is deleted for ignore_lock.
12:16:43 <jralls> Yeah, that's what my last was about.
12:16:53 <chris> ok gtg now late here.
12:17:04 <jralls> OK. I'll look at the logging thing.
12:17:12 * chris willbe happy when the 2 small commits are reverted and fixed properly.
12:17:32 <jralls> You mean the ones from this morning?
12:18:06 <chris> yeah my ones in the past 4 hours.
12:18:26 <chris> b0effbc36 is good to stay
12:18:35 <jralls> OK, *my* this morning. ;-)
12:21:27 *** User_ has quit IRC
12:33:49 *** TownsendHardware has joined #gnucash
12:38:35 *** sbluhm has joined #gnucash
12:38:35 *** ChanServ sets mode: +v sbluhm
12:39:44 <chris> jralls: another small issue: 658241d3b adds fprintf because the qof_session_get_error_message(session) for "locked-file" returns "". qofsessions's push_error are pushing blanks.
12:39:55 <chris> with that, off now!
12:40:48 *** omnireq__ has quit IRC
12:41:01 *** omnireq__ has joined #gnucash
12:46:05 <jralls> chris, I put logging back (with PERR instead of g_warning, but that eventually does essentially the same thing). Works fine. I think you were expecting it to write to stderr but you didn't tell it --logto=stderr.
12:46:45 <jralls> gjanssens, maybe the default in gnucash-cli should be to log to stderr instead of gnucash.trace?
12:47:34 *** sbluhm has quit IRC
12:57:15 *** sbluhm has joined #gnucash
12:57:15 *** ChanServ sets mode: +v sbluhm
13:03:18 <fell> jralls, you did not remove the obsolete patches after the guile update?
13:03:57 <fell> -on-windows
13:04:50 <jralls> fell, no, because jhbuild pulls them from the repository so deleting them would break anyone who's running a local and unupdated gnucash.modules.
13:07:32 <fell> Ah, is there a schedule for the remove?
13:08:13 <jralls> That would imply a schedule for people to update their local gnucash-on-windows. Do you know how to enforce one?
13:09:33 <jralls> There's also the matter that if they're deleted from the repo they'll exist only in my local clone of guile. Since the guile folks weren't interested in them we'll continue to need them until we get rid of guile.
13:10:44 *** sbluhm has quit IRC
13:15:57 <fell> They would still exist in the git history, but if you think, we will need them again ...
13:16:44 <jralls> I *know* we will need them again.
13:17:58 <fell> Something else: you maint branch was not up to date, when you merged it into master.
13:20:53 <fell> but perhaps you resolved the conflict, I was unable to resolve. So I will try to merge again..
13:21:58 <fell> yep
13:33:19 *** tonysoar has joined #gnucash
13:35:59 *** suukim has quit IRC
13:39:35 *** sbluhm has joined #gnucash
13:39:35 *** ChanServ sets mode: +v sbluhm
13:42:07 <fell> done
13:53:58 <gjanssens> jralls: yes, I think gnucash-cli should default log to stderr
13:54:36 <jralls> Can you think of a quick way to do that?
14:00:59 *** jervin has joined #gnucash
14:01:39 <gjanssens> For example test if the logto command line parameter is empty and set it to stderr in that case before the call to qof_log_init ?
14:01:51 <gjanssens> I'm not sure the code is structured to allow that easily.
14:03:12 <gjanssens> I think you can add this test and correction in Gnucash::GnucashCli::parse_command_line_options
14:03:20 <gjanssens> s/correction/override/
14:04:41 <jralls> That would affect both cli and gui. I think it wants to go in GnucashCli::configure_program_options.
14:05:44 <gjanssens> GnucashCli::parse_command_line won't affect gui
14:06:24 <jralls> Sorry, misread the class.
14:09:47 *** jervin has quit IRC
14:13:40 <jralls> That works, but requires making m_log_to_filename protected. I guess that's not a concern since there are several other protected GnucashCoreApp member variables.
14:26:09 <jralls> Committed & pushed. Thanks!
14:26:47 *** frakturfreak has joined #gnucash
14:26:47 *** ChanServ sets mode: +v frakturfreak
14:33:11 *** tonysoar has quit IRC
14:33:35 *** jervin has joined #gnucash
14:43:38 *** jervin1 has joined #gnucash
14:44:05 *** jervin has quit IRC
14:44:05 *** jervin1 is now known as jervin
14:48:31 *** jervin has quit IRC
14:48:38 *** jervin has joined #gnucash
14:50:47 *** jervin has quit IRC
15:02:09 *** jervin has joined #gnucash
15:20:50 *** sbluhm has quit IRC
15:25:41 <gjanssens> jralls: thanks
15:26:23 <jralls> fell, NP, though at this point I've lost track of what for...
15:26:33 <gjanssens> In retrospect doing this in GnucashCli::configure_program_options would conceptually be even better.
15:27:07 <gjanssens> There are ways to set default options for program_options, but I don't know if that can be done still after the option has been defined.
15:27:31 <gjanssens> If what you added works that's good enough for now.
15:30:01 *** gjanssens has quit IRC
15:45:38 *** mr_pants has joined #gnucash
15:47:14 <mr_pants> Hello everyone, GNUCash version 3.10 - When I open it up, it presents me the past due invoices. What is the column "CN?: mean? It is a check box column and can't figure out why some are checked and others aren't.
15:47:15 <mr_pants> Thank you
15:47:46 <jralls> Dunno. warlord?
15:54:32 *** guak has joined #gnucash
16:05:36 *** storyjesse has quit IRC
16:06:54 *** bertbob has quit IRC
16:09:02 <CDB-Man> chris: saw the fixes over the past couple days. I've been super busy this week, do hoping to test them out this Tuesday
16:09:23 <fell> mr_pants: "Credit note"?
16:09:43 *** bertbob has joined #gnucash
16:09:43 *** ChanServ sets mode: +v bertbob
16:11:42 <fell> This abbreviation is the column heading for the condition "Is this invoice a Credit Note?"
16:12:47 * fell is wondering if we can attach tooltips to header columns.
16:17:11 *** jervin1 has joined #gnucash
16:17:37 *** jervin has quit IRC
16:17:37 *** jervin1 is now known as jervin
16:21:42 *** Mechtilde has quit IRC
16:21:46 <jralls> fell: Not exactly, but we should be able to attach them to the labels.
16:22:22 <jralls> Or to every checkbox. Write a bug for BobIT.
16:37:07 <fell> https://bugs.gnucash.org/show_bug.cgi?id=797774
18:24:23 *** guak has quit IRC
18:26:37 *** jervin has quit IRC
19:00:50 *** Aussie_matt has joined #gnucash
19:13:49 <fell> I did the SuSE update myself: https://build.opensuse.org/package/show/home:felle:branches:GNOME:Apps/gnucash-docs
19:20:48 *** omnireq__ has quit IRC
19:21:50 *** omnireq has joined #gnucash
19:21:50 *** ChanServ sets mode: +v omnireq
19:46:54 *** Aussie_matt has quit IRC
19:48:01 *** Aussie_matt has joined #gnucash
19:50:21 *** mr_pants has quit IRC
19:59:37 *** frakturfreak has quit IRC
20:02:59 <warlord> jralls, was the "dunno" question about "CN?"
20:03:10 <jralls> Yes.
20:03:18 <jralls> Was fell right?
20:04:51 <jralls> gtg
20:05:02 *** jralls is now known as jralls_afk
20:05:23 <warlord> jralls_afk, yes
20:05:39 <warlord> (at least I am 90% sure; gjanssens would be the ideal person to answer)
20:07:48 *** omnireq has quit IRC
20:08:02 *** omnireq has joined #gnucash
20:08:02 *** ChanServ sets mode: +v omnireq
21:08:48 *** omnireq has quit IRC
21:10:07 *** omnireq has joined #gnucash
21:10:07 *** ChanServ sets mode: +v omnireq
21:56:06 <chris> jralls: PINFO() for report_session_percentage isn't quite right. gnucash --run-report="Transaction Report" large-datafile.gnucash - it should show a nice progress report.
21:56:38 <chris> both PINFO are designed to be displayed at all times
22:21:13 *** CDB-Man_ has joined #gnucash
22:21:13 *** ChanServ sets mode: +v CDB-Man_
22:23:38 *** CDB-Man has quit IRC
22:32:18 *** omnireq has quit IRC
22:32:39 *** omnireq has joined #gnucash
22:32:39 *** ChanServ sets mode: +v omnireq
22:48:05 *** Gerd1 has joined #gnucash
22:48:41 *** Gerd has quit IRC
22:48:41 *** Gerd1 is now known as Gerd
23:14:48 *** omnireq has quit IRC
23:17:12 *** omnireq has joined #gnucash
23:17:12 *** ChanServ sets mode: +v omnireq
23:43:56 *** waeking has quit IRC
23:45:52 *** waeking has joined #gnucash
23:45:52 *** ChanServ sets mode: +v waeking