2019-10-29 GnuCash IRC logs

01:27:26 *** Mechtilde has joined #gnucash
01:39:07 *** JayC has joined #gnucash
01:39:07 *** ChanServ sets mode: +v JayC
02:48:00 *** pohly1 has joined #gnucash
02:53:34 *** Mechtilde has quit IRC
02:54:39 *** sbluhm has joined #gnucash
02:54:39 *** ChanServ sets mode: +v sbluhm
03:04:42 *** mauritslamers has quit IRC
03:06:31 *** sbluhm has quit IRC
03:06:39 *** sbluhm has joined #gnucash
03:06:39 *** ChanServ sets mode: +v sbluhm
03:43:51 *** fell has joined #gnucash
03:43:51 *** ChanServ sets mode: +o fell
03:48:03 *** fabior has joined #gnucash
04:04:23 *** gjanssens has joined #gnucash
04:04:24 *** ChanServ sets mode: +o gjanssens
04:07:34 *** bertbob has quit IRC
04:08:50 *** bertbob has joined #gnucash
04:08:51 *** ChanServ sets mode: +v bertbob
04:27:20 *** fabior has quit IRC
04:36:11 *** storyjesse has quit IRC
05:50:46 *** kriesel has joined #gnucash
06:01:26 *** Mechtilde has joined #gnucash
06:04:17 *** kriesel has quit IRC
06:10:19 *** mauritslamers has joined #gnucash
06:10:19 *** ChanServ sets mode: +v mauritslamers
06:28:50 *** Jimraehl1 has joined #gnucash
06:30:21 *** Jimraehl1 has quit IRC
06:31:39 *** Mechtilde has quit IRC
06:37:06 *** fabior has joined #gnucash
06:43:41 *** fabior has quit IRC
07:22:47 *** oozer has joined #gnucash
07:33:08 *** fabior has joined #gnucash
07:35:07 *** fabior has quit IRC
07:41:39 *** Mechtilde has joined #gnucash
07:43:55 *** fell has quit IRC
07:52:03 <chris> fao gjanssens I'll wait until master to apply patches for srfi-9 as follows: https://pastebin.com/raw/SqP4j4tB
08:00:23 <warlord> .
08:01:36 <chris> anyone near california fires, stay safe :)
08:13:48 *** fell has joined #gnucash
08:13:49 *** ChanServ sets mode: +o fell
08:17:18 <warlord> jralls is in NorCal. I know there are fires up there but don't know how far away.
08:17:30 <warlord> Willdev (who used to hang out here but hasn't been seen in a while) is in SoCal.
08:17:37 <warlord> I don't know anyone else offhand.
09:59:39 *** fabior has joined #gnucash
10:01:51 <gjanssens> chris: ok
10:09:52 *** Mechtilde has quit IRC
10:11:55 *** omnireq has quit IRC
10:17:06 *** fell has quit IRC
10:17:11 *** fell_laptop has joined #gnucash
10:17:11 *** ChanServ sets mode: +o fell_laptop
10:19:50 *** fell_laptop is now known as fell
10:21:24 <fell> In 1985 I was between the fires in the Chianti mountains in Italy - heavy!
10:25:31 *** fabior has quit IRC
10:45:48 *** Mechtilde has joined #gnucash
10:54:24 <fell> gjanssens: all this 'ls: cannot access '/home/frank/git/gnucash-docs/help/pt/figures/*.svg': No such file or directory' (3x per lang and doc) are annoying for 1 svg file.
11:02:27 *** omnireq has joined #gnucash
11:02:27 *** ChanServ sets mode: +v omnireq
11:06:58 *** fell has quit IRC
11:07:06 *** fell has joined #gnucash
11:07:06 *** ChanServ sets mode: +o fell
11:30:20 <fell> chris: from https://code.gnucash.org/builds/win32/build-logs/maint/2019-10/build-maint-2019-10-29-03-21-31.log: * 03:34:13 ERROR <gnc.scm> report.scm error: One of your reports has a report-guid that is a duplicate. Please check the report system, especially your saved reports, for a report with this report-guid: 47f45d7d6d57b68518481c1fc8d4e4ba
11:31:32 *** fabior has joined #gnucash
11:31:42 <fell> That appears several times.
11:31:50 *** kael has joined #gnucash
11:31:50 *** ChanServ sets mode: +v kael
11:41:13 *** mdf has quit IRC
11:42:11 *** guak has joined #gnucash
11:43:41 *** fabior has quit IRC
11:45:04 *** fabior has joined #gnucash
11:47:14 *** calvinct has joined #gnucash
12:00:44 *** fabior has quit IRC
12:05:11 <gjanssens> fell: Ok, I can see that. I have changed the svg detection code to no longer emit that warning.
12:05:38 <gjanssens> You'll have to rerun autogen.sh and configure to get the new code into your makefiles
12:06:40 <gjanssens> The cmake code didn't have this issue ;)
12:17:42 *** Mechtilde has quit IRC
12:24:36 <jralls> warlord: Alex Aycinena lives in San Mateo and I think Tommy Trussell is in Santa Clara. The serious NorCal fire is up in Sonoma, about 75 miles away. There have been a few smaller ones closer but they've all been quickly dealt with so far.
12:28:34 <jralls> gjanssens: https://groups.google.com/forum/#!topic/wx-dev/Z-FteLjJblk is about using git notes instead of commit messages for changelog comments. AFAIK we haven't used git notes for anything. Maybe worth thinking about?
12:31:01 *** fabior has joined #gnucash
12:36:26 <gjanssens> jralls: interesting topic. However it confuses me. Isn't the changelog simply a summary of changes (hence the commit messages) ? Are you instead thinking of storing release note snippets in git notes ?
12:38:03 <jralls> I think wx uses their changelog as more like our NEWS file, so yes, I was thinking of release note snippets. There might be other uses for git-notes as well.
12:39:27 <jralls> There's also Ollie's mention of git-merge-changelog. We have a custom git-log snippet for doing that, but maybe the purpose-built program would work better. Like Vadim I hadn't noticed its existence before.
12:42:56 <fell> Thanks, gjanssens!
12:44:50 <gjanssens> fell: you're welcome
12:47:51 <gjanssens> jralls: remind me, how do we generate our changelog again ?
12:49:03 <jralls> https://github.com/Gnucash/gnucash/blob/maint/util/git-release-notes.pl
12:49:04 <gjanssens> Oh nevermind, I see the cmake rules
12:49:51 <jralls> Oh, sorry, wrong script. Yes, the cmake rules for changelog.
12:50:53 *** kael has quit IRC
12:54:15 <gjanssens> The git-merge-changelog man page is extremely sparse...
12:54:45 <gjanssens> It doesn't give any details on which log formats it supports (Olly hints this may be important)
12:55:52 <gjanssens> IIURC this tool would be useful when maintaining a changelog on two branches (like maint and master) and one wants to merge one into the other.
12:56:11 <gjanssens> It does not extract changelog data from the git repo itself.
12:56:46 <gjanssens> So it would only be one component of generating changelogs, right ?
12:57:17 <gjanssens> We'd still have to extract the log entries somehow from our commits.
12:57:59 <jralls> Hmm, it seems to be separate from git as well. At least it doesn't turn up in https://git-scm.com/docs.
12:58:19 <gjanssens> It's a separate tool indeed. I had to install it from the Fedora repo
12:59:16 <gjanssens> But back to the original topic - I do like the idea of storing supplementary release note snippets in git notes.
13:00:00 <gjanssens> That also allows us to tweak them if something went wrong (silly typo's or a commit being reverted for not being ready yet,...)
13:00:17 <jralls> Me too, especially the part about being able to change them later without changing the actual commit.
13:01:46 <gjanssens> Indeed
13:01:59 *** ArtGravity has joined #gnucash
13:02:00 *** ChanServ sets mode: +v ArtGravity
13:09:19 *** calvinct has quit IRC
13:25:01 *** kael has joined #gnucash
13:25:01 *** ChanServ sets mode: +v kael
13:26:59 *** fabior has quit IRC
13:28:49 <jralls> gjanssens, fell: Time to continue the what gets saved where discussion?
13:29:11 <gjanssens> Ok, I have a few cycles
13:29:32 <fell> OK
13:30:26 <jralls> We left off with "if we can save things in new tables in SQL we can do the same with new elements in XML".
13:30:34 *** tzhuang has quit IRC
13:31:22 <jralls> The only fly in that ointment is that the current XML backend will refuse to load a file with unknown elements while the SQL backend will simply ignore extra tables.
13:32:36 <gjanssens> True. That's a general issue with the xml file format -- or at least with how our backend is designed.
13:33:10 <fell> Hm, in theory, yes. But I see SQL grants and views as a different thing and so woukd keep them in separate places in XMLL.
13:33:23 <gjanssens> This is not directly related to whether or not to store report configurations in xml though.
13:33:41 *** tzhuang has joined #gnucash
13:33:50 <gjanssens> fell: I was not considering sql views
13:34:01 <gjanssens> I don't even know if sqlite supports this
13:34:04 <jralls> It is in the sense that the alternative is to hugely increase the amount of KVP. That's not pretty.
13:34:14 <fell> In some way that is, what reports are.
13:34:30 <gjanssens> fell: no, a report is more than just a view
13:35:00 <gjanssens> A view only holds a query, a report also holds user configurable representation data
13:35:02 <jralls> SQL views are supported in SQLite3, but views are just automagically maintained joins. They're not really relevant here.
13:35:27 *** kael has quit IRC
13:35:45 <gjanssens> jralls: as for storing in xml, I think I suggested much earlier to define a new xml tag for this.
13:36:15 <gjanssens> We can introduce it in the 3.x series without any means to actually set one via the UI or other apis
13:36:17 <jralls> And while reports get their data by running a query they're not really SQL views either. The register is like a SQL view that joins a transaction with its splits.
13:36:25 <gjanssens> And only have 4.x really write to this new tag
13:36:44 <fell> The whole gnc-2 schema should be overhault to get rid of most slots.
13:37:17 <gjanssens> Absolutely, though that's too big a refactoring for 4.x
13:37:55 <jralls> If we're going to get 4.x out in the next 4-6 months it sure is.
13:38:44 <gjanssens> Hmm, re 3.x not writing to the new tag, that's always tricky...
13:39:15 <gjanssens> Once 4.x writes to this new tag, 3.8+ should be able to at least read it and ideally write it as well.
13:39:33 <gjanssens> A bit similar to the budget fixes chris has been working on.
13:40:07 <gjanssens> The other option is to have 3.x only read from the new tag and write changes to the saved-reports file again
13:40:20 <jralls> If 3.8 can write then it would break 3.7.
13:40:24 <gjanssens> Reopening the file in 4.x should then rerun the import.
13:41:04 <gjanssens> 3.8 should only write *if* there are already these tags used.
13:41:05 <fell> iMHO saved repots should be per user, not per file.
13:41:23 <gjanssens> That means, if the file has already been saved at least once with 4.x
13:41:43 <jralls> fell: Yeah, we know. But in everyone else's HO it should be per file.
13:42:15 <gjanssens> fell: no. We had this discussion before. Saved reports can't be per user, as some data in the report options is book specific. One example: account lists
13:43:05 <gjanssens> We can consider import/export functions for the reports, with the clear warning some options may not be valid in other books.
13:43:24 <fell> Per user and file, precisely
13:43:31 <jralls> gjanssens, can we guard it with a feature flag or will sixtp-dom-parser already have exploded because of the unknown element?
13:43:57 <gjanssens> jralls: that's a good question
13:44:46 * gjanssens will run a quick experiment...
13:47:00 <fell> My imagination, which is now possible: I want to share my file with it's external accounting views with my CPA, but noot my internal accounting views.
13:47:46 <jralls> fell: What does "accounting views" mean in this context?
13:49:10 <fell> He is not interested in my https://en.wikipedia.org/wiki/Cost_accounting
13:50:18 <fell> external means reports generated for the bank, the tax office, ..
13:51:18 <gjanssens> fell: if he's not interested, he isn't required to look at them
13:51:53 <gjanssens> So far I don't think I intend to also save which reports and tabs are open
13:52:08 <fell> But it would be hard for him to find the right reports between the 101 others.
13:52:12 <gjanssens> Only how the saved report should look when opened
13:52:40 <gjanssens> That's like saying it would be hard to find the right expense account within 300 others
13:53:01 <gjanssens> You will have to tell your accountant the names of the relevant reports
13:53:51 <gjanssens> And turning your argument around, if you want your accountant to get your external accounting views you still need to provide them to your accountant in some way. Currently that means either via pdf or html export
13:54:18 <jralls> fell: Use a naming convention, e.g. "cost_foo" for cost-accounting reports, or "accountant_bar" for ones the accountant will want to use.
13:54:44 <gjanssens> I don't think you can expect users to go in and manually manipulate a copy of saved reports to send to the accountant with instructions where to install it to get the actual reports.
13:54:53 *** User_ has joined #gnucash
13:55:50 <fell> Currently I can copy/create the required reports under a 2. user.
13:55:52 <gjanssens> jralls: unfortunately our backend code currently bails on a bad tag before the feature codes are interpreted :(
13:56:17 <jralls> :( indeed.
13:56:49 <gjanssens> I wonder if we can change that.
13:58:07 <jralls> Going forward, of course we can as long as we use a stream parser. We just have to put the feature flags up front right after the xmlns decls.
13:58:37 <jralls> And always use a non-validating SAX parser.
13:59:03 <gjanssens> They are part of the book's slots which appear to be written before anything else
13:59:26 *** User_ has quit IRC
14:00:56 <gjanssens> That's all possible but probably not something to introduce at this stage...
14:01:13 <gjanssens> Which means currently our only reasonable option is slots
14:01:54 <gjanssens> fell: you can also copy your book to a second user and remove the reports you don't want your accountant to see
14:02:25 <gjanssens> More even, you no longer need a second user. You simply make a copy of the book.
14:04:38 <fell> That can be done once per year, but there are other use cases:
14:05:27 <fell> Imagine one guy does only sales (client reports) and another the rest.
14:06:17 <fell> Each day stripping the not required parts?
14:06:39 <gjanssens> fell: we're talking about gnucash! not a big company/role-separated accounting application
14:07:17 <fell> Yep, but currently this is possible.
14:07:29 <gjanssens> The reports all relate to the same data, we don't have privilege separation so trying to split all this per user is overkill IMO
14:07:36 <jralls> fell: Possible and supported are very different.
14:08:06 <gjanssens> We hardly support sharing books for that matter...
14:09:48 <gjanssens> To properly support per-user reports we should first introduce a user concept in gnucash
14:09:53 <jralls> In any case, naming conventions work as well or better than having the saved reports outside of the book as soon as two users need the same saved report.
14:11:03 <gjanssens> File system trick are not a good replacement for this.
14:11:10 <gjanssens> tricks*
14:11:48 <gjanssens> fell: and the use cases you are defending here, are you using them yourself ?
14:12:12 *** storyjesse has joined #gnucash
14:12:46 <fell> I remember we discussed them sometimes on gnucash-de.
14:12:51 *** Mechtilde has joined #gnucash
14:13:18 <gjanssens> Ok
14:13:34 <fell> I don't know, if it is really used.
14:14:30 <jralls> Yeah, and someone (Tommy Trussell?) claims to have run a Fortune-500 company's books on GnuCash. Having been involved in corporate audits I find the claim utterly unbelievable.
14:14:57 * gjanssens too
14:15:55 <jralls> GnuCash's complete lack of internal controls--that "user" concept foremost--makes it unsuitable for more than one or two people having their fingers in the books.
14:16:02 <gjanssens> So I go back to the idea that a gnucash book is essentially single-user.
14:16:42 *** TownsendHardware has joined #gnucash
14:16:53 *** storyjesse has quit IRC
14:16:57 <jralls> Or two at most: The proprietor and his accountant... though apparently finding an accountant interested in using GnuCash is difficult.
14:17:01 <gjanssens> That single user may have multiple devices to work on this book (a PC + a laptop is not uncommon, or a book stored on some file sharing system such as dropbox)
14:17:10 <jralls> fell may be very lucky in that regard.
14:17:34 <gjanssens> This user would no doubt want to have all his reports available on each device
14:17:46 <gjanssens> Having to remember to copy the saved-reports file is a chore
14:18:03 <gjanssens> This is my counter use case, and I suspect it to be more common actually
14:18:06 <fell> Anyway, what I support is splitting "saved-reports" in <book>-reports.
14:18:14 <jralls> Not to mention copying GNC_DATA_DIR when getting a new computer.
14:19:51 <gjanssens> fell: can you detail that a bit ?
14:20:03 * gjanssens has about 10 mins left...
14:20:28 <fell> There is currently one saved reports for all books.
14:21:36 <fell> Imagine I have the company in DE and my residence in BE or vice versa.
14:22:35 <fell> So I have different accounts in my personal and my private book.
14:24:33 <fell> And all get mixed in the saved-reports.
14:25:48 <fell> personal and business book (2 lines ^)
14:28:20 <jralls> fell: So you want to have a saved report section in each GNC_DATA_DIR/books/*.gcm?
14:33:03 *** calvinct has joined #gnucash
14:34:30 <fell> *.gcm could store, which saved reports to use.
14:34:59 <fell> I am not shure, if it should become a section in gcm.
14:35:00 <gjanssens> It seems to me your primary concern is filtering
14:35:15 <gjanssens> Let's go back to your accountant use case
14:35:25 <gjanssens> Imagine all reports are stored in the book
14:35:44 <gjanssens> What's missing is a decent filter option to only show reports relevant to the accountant
14:36:02 <gjanssens> The most basic approach would be to add a filter based on report name
14:36:39 <gjanssens> So using jralls' suggestion of naming reports for the accountant "Accountant_*", if we allow the accountant to enter "Accountant_" in a filter field
14:36:47 <gjanssens> that would reduce the list of reports to show
14:37:09 <gjanssens> I am fine with making this *filter* per user/per book
14:37:41 <gjanssens> In that case it doesn't matter there are more reports in the book, if the accountant saves the filter, he will only see those he cares about
14:38:17 <gjanssens> That would equally apply to the use case of one person does only sales another only purchases,...
14:39:15 <gjanssens> If you can have each user define his own filters, what they see would be similar, but the convenience much higher
14:39:24 <fell> I wonder, if we should better move the reports into the gcm.
14:39:44 <fell> That seems the more appropriate place for me.
14:39:47 <gjanssens> No more need to shuffle around external files. Just the book is sufficient
14:40:04 <gjanssens> The filter belongs in gcm, the saved reports don't
14:40:46 <fell> Ok, I think that is a possible way.
14:41:40 <gjanssens> Glad we found some common ground :)
14:42:04 * gjanssens has to leave now. Guests have arrived...
14:42:12 <gjanssens> Perhaps TTYL this evening
14:42:13 <fell> Hve fun!
14:42:16 *** gjanssens is now known as gjanssens_afk
14:45:37 <jralls> fell: Putting saved reports in the gcm along with the window layouts seems a particularly *bad* idea: Consider the most common case, a user with a desktop (big screen) and a laptop (little screen). She may well use different window sizes and arrangements on the two. If she has two monitors on her desktop the window might even be placed outside the visible screen area on the laptop.
14:46:46 <fell> right
14:48:01 <jralls> gjanssens's idea of filters would be better since they can be more easily re-typed on each machine that shares the book.
15:00:09 *** calvinct has quit IRC
15:05:12 *** sbluhm has quit IRC
15:13:11 *** bertbob has quit IRC
15:15:32 *** bertbob has joined #gnucash
15:15:32 *** ChanServ sets mode: +v bertbob
15:18:46 *** fell_laptop has joined #gnucash
15:18:46 *** fell has quit IRC
15:18:47 *** ChanServ sets mode: +o fell_laptop
15:21:09 *** bertbob has quit IRC
15:22:34 *** sbluhm has joined #gnucash
15:23:23 *** bertbob has joined #gnucash
15:23:24 *** ChanServ sets mode: +v bertbob
15:41:49 *** frakturfreak has joined #gnucash
15:42:17 *** sbluhm has quit IRC
15:43:42 *** sbluhm has joined #gnucash
15:43:42 *** ChanServ sets mode: +v sbluhm
15:44:30 *** bertbob has quit IRC
15:54:30 *** bertbob has joined #gnucash
15:54:31 *** ChanServ sets mode: +v bertbob
16:01:51 *** mauritslamers has quit IRC
16:16:46 *** fell_laptop is now known as fell
16:20:27 *** g5pw has quit IRC
16:28:41 *** Mechtilde has quit IRC
17:00:12 *** sbluhm has quit IRC
17:11:40 *** gjanssens_afk has quit IRC
17:28:07 *** mauritslamers has joined #gnucash
17:28:07 *** ChanServ sets mode: +v mauritslamers
17:29:26 *** JayC has quit IRC
17:36:46 *** guak has quit IRC
17:42:56 *** guak has joined #gnucash
17:43:26 *** jervin has joined #gnucash
17:43:32 *** calvinct has joined #gnucash
17:49:28 *** calvinct has quit IRC
17:57:16 *** frakturfreak has quit IRC
18:01:53 *** omnireq has quit IRC
18:06:52 *** pohly1 has quit IRC
18:11:38 *** jeaye has quit IRC
18:54:06 *** fell has quit IRC
18:54:15 *** fell has joined #gnucash
18:54:15 *** ChanServ sets mode: +o fell
19:12:35 *** CDB-Man has joined #gnucash
19:12:36 *** ChanServ sets mode: +v CDB-Man
19:12:46 *** omnireq has joined #gnucash
19:12:58 *** CDB-Man_ has quit IRC
20:33:18 *** ArtGravity has quit IRC
20:36:56 *** gggg_ has quit IRC
20:55:19 *** fell has quit IRC
20:58:18 *** guak has quit IRC
21:17:59 *** omnireq_ has joined #gnucash
21:19:10 *** omnireq has quit IRC
21:24:44 *** jervin has quit IRC
21:37:06 *** oozer has quit IRC
22:17:31 *** o01eg_ has joined #gnucash
22:18:46 *** o01eg has quit IRC
23:14:26 *** TommyT has joined #gnucash
23:14:26 *** ChanServ sets mode: +v TommyT
23:45:35 *** gggg has joined #gnucash
23:53:57 *** TommyT has quit IRC
23:57:15 *** omnireq_ has quit IRC
23:58:17 *** omnireq_ has joined #gnucash