2019-04-04 GnuCash IRC logs

00:19:57 *** nimish2711 has quit IRC
00:20:25 *** nimish2711 has joined #gnucash
00:30:17 *** nimish2711 has quit IRC
00:30:46 *** nimish2711 has joined #gnucash
00:35:38 *** nimish2711 has quit IRC
00:36:06 *** nimish2711 has joined #gnucash
00:37:34 *** O01eg has quit IRC
00:45:58 *** nimish2711 has quit IRC
00:46:23 *** O01eg has joined #gnucash
00:46:26 *** nimish2711 has joined #gnucash
00:51:18 *** nimish2711 has quit IRC
00:51:46 *** nimish2711 has joined #gnucash
00:58:33 *** storyjesse has quit IRC
00:59:25 *** storyjesse has joined #gnucash
01:00:49 *** nicoroumeau has joined #gnucash
01:06:38 *** nimish2711 has quit IRC
01:07:07 *** nimish2711 has joined #gnucash
01:16:25 *** nicoroumeau has quit IRC
01:22:06 *** storyjesse has quit IRC
01:31:59 *** nimish2711 has quit IRC
01:32:27 *** nimish2711 has joined #gnucash
01:37:19 *** nimish2711 has quit IRC
01:37:48 *** nimish2711 has joined #gnucash
01:47:58 *** fell_laptop has joined #gnucash
01:47:59 *** ChanServ sets mode: +o fell_laptop
01:49:44 *** fell has quit IRC
01:52:39 *** nimish2711 has quit IRC
01:53:07 *** nimish2711 has joined #gnucash
02:01:59 *** storyjesse has joined #gnucash
02:02:43 *** storyjesse has quit IRC
02:03:14 *** storyjesse has joined #gnucash
02:12:59 *** nimish2711 has quit IRC
02:13:21 *** storyjesse has quit IRC
02:13:27 *** nimish2711 has joined #gnucash
02:17:28 *** warlord has joined #gnucash
02:17:28 *** gncbot sets mode: +o warlord
02:18:19 *** nimish2711 has quit IRC
02:23:51 <warlord> .
02:30:52 *** gjanssens has joined #gnucash
02:30:53 *** ChanServ sets mode: +o gjanssens
02:39:08 *** nimish2711 has joined #gnucash
02:54:00 *** nimish2711 has quit IRC
03:00:03 *** fabior has joined #gnucash
03:09:55 <gjanssens> chris: I'm reviewing the different ways we use in gnucash to find and load report definitions into the system at the file level
03:11:03 <gjanssens> Standard reports are loaded via (resolve-interface ...) here https://github.com/Gnucash/gnucash/blob/maint/gnucash/report/standard-reports/standard-reports.scm#L119
03:12:30 <gjanssens> Business reports on the other hand are injected via plain (use-modules ...) calls here https://github.com/Gnucash/gnucash/blob/maint/gnucash/report/business-reports/business-reports.scm#L113
03:13:39 <gjanssens> And for hand-written reports we tell users to add a (load ...) call in config-user.scm
03:14:00 <gjanssens> So we have 3 different methods to inject reports into the gnucash runtime
03:15:06 <gjanssens> (module-use! with resolve-interface), (use-modules) and (load)
03:16:20 <gjanssens> I'm trying to wrap my head around these and understand the benefits/drawbacks of each method
03:16:48 <gjanssens> For one, I don't think it makes sense to use different mechanisms for reports provided with gnucash
03:17:19 <gjanssens> I mean why on earth would we want different injection mechanisms for standard and business reports ?
03:18:34 <gjanssens> One benefit of the discovery code written for the standard reports is we can easily add reports without having to change the runtime
03:18:52 <gjanssens> If we want to add a business report, we have to update the load code as well.
03:19:38 <gjanssens> On the other hand only explicitly loading what we know we shipped could be considered a safety measure
03:20:53 <gjanssens> warlord: the choice for (use-modules) to inject the business reports, was that yours ? If so do you remember why you chose that over the dynamic code used for standard-reports?
03:22:22 <gjanssens> Continuing on this, I also don't understand why the standard reports discovery code uses (module-use! ...(resolve-interface ...)) rather than plainly running a load command to evaluate the files it found
03:23:34 <gjanssens> It seems to me the resolve-interface way could result in a different file being loaded than was intended (like if a file with the same name and relative path is found somewhere earlier on the guile load path)
03:24:14 <gjanssens> Or was that a deliberate override mechanism ? If so that would be pretty convoluted IMO...
03:25:16 <warlord> gjanssens, the dynamic code didn't exist at the time.
03:25:19 <gjanssens> Anyway, while I know this code predates your and my involvement with the project I'd love to hear your thoughts on this.
03:25:29 <gjanssens> warlord: ah, interesting
03:25:56 <warlord> At the time everything was (use-module ..)
03:25:57 <gjanssens> So it's actually more of an incomplete rewrite that has gotten us in this mixed up situation
03:26:11 <warlord> There was a standard-reports.scm that did a (use-module...) for all the Std Reports.
03:26:52 <gjanssens> Ok, that at least answers one question
03:27:34 <warlord> But the question was about 3rd party additional reports -- we had a few external developers creating them -- and making it easier to add new reports without having to change the code to load it.
03:27:42 <warlord> I don't know why the biz reports weren't changed, too.
03:29:10 <gjanssens> Ok. I plan to do so now in my current local branch
03:29:51 <gjanssens> All gnucash supplied reports will be loaded with one single method.
03:30:21 <gjanssens> Question is whether it would be with the resolve-interface method or with plain load commands for each report file found
03:31:05 <warlord> I don't know. The main important method is being able to catch a scheme error so GnuCash doesn't crash.
03:31:06 <gjanssens> At this point I'm tempted to use load rather than resolve interface as that at least guarantees it's the discovered file that gets loaded.
03:31:20 <gjanssens> True, we have gfec for that.
03:48:04 *** Mechtilde has joined #gnucash
04:05:55 *** Aussie_matt has quit IRC
04:15:47 *** Mechtilde has quit IRC
04:16:08 *** Mechtilde has joined #gnucash
04:16:19 *** warlord has quit IRC
04:16:27 *** warlord has joined #gnucash
04:16:27 *** gncbot sets mode: +o warlord
04:20:03 <chris> to me load (load) is easiest...
04:20:36 <chris> I can try modify standard-reports.scm to (load) instead of (resolve-interface) which sounds weird
04:20:47 *** Mechtilde has quit IRC
04:20:50 <chris> easiest = shortest
04:24:43 *** warlord has quit IRC
04:24:46 *** warlord has joined #gnucash
04:24:46 *** gncbot sets mode: +o warlord
04:25:09 <warlord> .
04:28:16 *** fabior has quit IRC
04:36:39 *** Mechtilde has joined #gnucash
04:41:19 *** warlord has quit IRC
04:41:28 *** warlord has joined #gnucash
04:41:28 *** gncbot sets mode: +o warlord
04:49:23 <gjanssens> chris: thanks
04:49:51 <gjanssens> In this case I actually took is as a personal excercise so no need for you to take this up :)
04:50:09 <gjanssens> I'm doing more than just using load instead of resolve-interface
04:50:22 <gjanssens> But I wanted to get that bit clear
04:51:09 <warlord> Sorry for popping in and out. Having some network issues here at the conference (clearly).
04:51:17 <gjanssens> I will put it up for review though when it's ready
04:51:20 <warlord> Or could just be my laptop wifi getting more and more flakey..
04:51:20 <gjanssens> warlord: np
04:51:23 <warlord> Not sure which.
04:51:48 <gjanssens> To test your laptop wifi you could experiment with a wifi dongle
04:51:56 <chris> gjanssens you're welcome... scheme is still fun I hope
04:52:05 <gjanssens> Of course you probably don't carry that with you on the conference
04:52:25 <chris> meanwhile I'm breaking out the report<->scm definition in preparation for export to/from .gcm files
04:54:38 <gjanssens> chris: scheme is obviously not my primary language, but it's more interesting now we have someone on board that is cleaning up the biggest warts in there...
04:56:16 *** nimish2711 has joined #gnucash
04:56:38 <gjanssens> could you ever have imagined me asking you for advice now rather than me mentoring you through your first steps in gnucash land :)
04:56:55 <chris> :-o
04:57:23 <warlord> gjanssens, Correct, I do not have a wifi dongle, but you're also correct that I should try that.
05:04:38 *** bertbob has quit IRC
05:07:08 *** bertbob has joined #gnucash
05:07:09 *** ChanServ sets mode: +v bertbob
05:07:31 *** nimish2711 has quit IRC
05:08:00 *** nimish2711 has joined #gnucash
05:14:58 *** nimish2711 has quit IRC
05:17:14 *** Aussie_matt has joined #gnucash
05:17:35 <gjanssens> While we're on the topic - we talk of "standard" reports and "business" reports
05:17:36 *** fabior has joined #gnucash
05:18:53 <gjanssens> A bit of a weird way to name the difference between the two groups.
05:19:02 <gjanssens> "business" is clear, but "standard" ?
05:19:33 <gjanssens> Would it make sense to call those something else, like "personal" or "general"
05:19:59 <chris> tbh personally i'd prefer break out standard into general/assets/income/budget just like the menus
05:20:16 <warlord> I think the original idea was "standard" vs "custom"
05:20:25 <warlord> They have *always* been considered standard reports.
05:20:39 <gjanssens> Or is it perhaps time to drop the distinction completely as we have done in most of the rest of gnucash and just put all reports together in one directory
05:20:46 <warlord> I named mine "business reports" because the original concept was to have a separate "gnucash-business" package that would "plug in"
05:21:10 <gjanssens> Yeah, I got it "standard" is there for historical reasons
05:21:26 <gjanssens> And business was added later
05:21:37 <gjanssens> as a plug in
05:22:20 <gjanssens> So as there's no longer a separate business plugin we can freely reconsider the report organisation in the sources/install location
05:22:51 <warlord> Agreed
05:23:06 <gjanssens> chris: would you actually also store the report scm files in such a hierarchy ?
05:24:08 <gjanssens> Wouldn't that mean extra work if we decide to move the report to another sub-menu ? (Like, in both changing the menu-path and having to move it to a different directory)
05:25:37 <gjanssens> IMO the list of report files is not *that* long - they could all live in one directory
05:25:42 <chris> I wouldn't think it really matters; git tracks moves well
05:25:59 <chris> It may ease devlopment of helper files eg. all business/budget functions within the folder
05:26:15 <gjanssens> Oh, it's not git I'm worried about. It's requiring changes to CMakeLists.txt that shouldn't be forgotten
05:26:23 <warlord> I think the reports themselves declare what menu they live in, so having the SCM files together doesn't affect that.
05:26:36 <warlord> BIAB -- lunchtime.
05:26:36 <chris> ^correct
05:26:42 <gjanssens> warlord: absolutely, that's my point
05:27:27 <gjanssens> However chris' suggestion was to put the files in subdirectories similar to the way they are in the menus.
05:27:59 <gjanssens> And in that case if you choose to move a report from one menu to another, you'd also want to move it to another subdirectory to keep consistency
05:28:35 <gjanssens> Which in turn means updating CMakeLists.txt files.
05:28:46 <chris> I don't mind at all either way... for now they've lumped budget-* together which helps visually which is why I figured folders
05:31:01 <gjanssens> And having said all that I *was* considering subdirectories, but based on different criteria...
05:31:17 <chris> (btw how to conver gchar to SCM string?)
05:32:04 <gjanssens> chris: a gchar as in a single character or a string of type gchar* ?
05:32:35 <chris> string
05:32:47 <gjanssens> gchar is an "alias" for char, so that could mean a char* type string
05:32:59 <gjanssens> scm_from_c_string I believe should work
05:33:19 <gjanssens> (That's a c function of course)
05:33:36 <gjanssens> As for the directory structure I have in mind
05:33:57 <gjanssens> Besides the "standard" and "business" reports we also still have locale-specific and utility reports
05:35:07 <gjanssens> The utility section is mainly sample reports
05:35:29 <gjanssens> Locale-specific is under-developed but a good concept to expand on in the future IMO
05:35:43 <gjanssens> Add enhancement request https://bugs.gnucash.org/show_bug.cgi?id=773198 to this mix
05:36:26 <chris> aha scm_from_utf8_string and friends
05:36:26 <gjanssens> And I would think a should add an "sample" report directory to store anything that's proof-of-concept or demo like
05:36:43 <gjanssens> Ah sorry yes those are the ones I meant
05:37:10 * chris doesn't like locale-specific... it hides them all... I'd prefer to move them to be visible in all locales
05:37:47 <gjanssens> So I'd end up with these directories: standard (comprising both current standard and business reports), locale-specific, sample and one "reports" directory under GNC_USERCONFIG_DIR for hand-written reports
05:38:02 <gjanssens> chris: that's the part of locale-specific I meant is under-developed
05:38:33 <gjanssens> They should indeed be all visible, or at least it should be possible to make them visible on user request
05:40:38 <gjanssens> However they have the distinguishing trait they are only meaningful in specific "locale" contexts
05:41:11 <gjanssens> And for the "sample" reports they are different in that they shouldn't be visible by default at all.
05:41:51 <gjanssens> I would load them only if gnucash is run with the additional command line parameter "--extra"
05:42:28 <gjanssens> The reasoning is (partly following bug 773198) these reports are only useful as examples for people interested in writing their own reports
05:43:00 <gjanssens> For other's they're just distracting and confusing
05:43:22 * chris agrees
05:43:52 <chris> they should be loaded for travis though
05:44:13 <chris> or, shouldn't they be loaded but not added into menu?
05:44:22 <chris> because saved-reports may try calling them
05:44:34 *** fabior has quit IRC
05:45:11 <gjanssens> They probably should be loaded, and that's also how I planned to do it.
05:45:36 <gjanssens> The only change is whether or not to show the extra "Sample" submenu
05:46:56 <gjanssens> I didn't express this clearly indeed
05:48:19 *** warlord has quit IRC
05:48:53 *** warlord has joined #gnucash
05:48:53 *** gncbot sets mode: +o warlord
05:49:17 <chris> that sounds fine
06:01:14 *** warlord has quit IRC
06:25:02 *** warlord has joined #gnucash
06:25:02 *** gncbot sets mode: +o warlord
06:26:59 <warlord> speaking of locale-specific things... I think it would be nice to allow the user to select different locale-based hierarchies for New Hierarchy Assistant regardless of current locale. I.e., it would be nice to have a drop-down of available locales to switch the available choices.
06:28:15 <gjanssens> warlord: yes, absolutely
06:37:48 <warlord> :)
06:40:23 *** Jimraehl1 has joined #gnucash
06:41:21 *** Jimraehl1 has left #gnucash
06:56:19 *** warlord has quit IRC
06:58:32 *** warlord has joined #gnucash
06:58:32 *** gncbot sets mode: +o warlord
07:01:44 *** User_ has joined #gnucash
07:13:27 *** User_ has quit IRC
07:14:25 *** oozer has joined #gnucash
07:14:49 *** warlord has quit IRC
07:16:28 <chris> I've done tabbed-reports save/load as json \o/ \o/
07:17:33 *** storyjesse has joined #gnucash
07:17:45 *** warlord has joined #gnucash
07:17:45 *** gncbot sets mode: +o warlord
07:19:30 <chris> https://pastebin.com/raw/iz0E57Hx
07:25:48 <warlord> chris, looks great!
07:27:03 *** storyjesse has quit IRC
07:33:52 <gjanssens> chris: \o/
07:34:18 <chris> I've overwritten the read code in the commit, so, someone else will need to fix it please ;)
07:34:28 <chris> so, no backward compatibility here
07:40:24 *** warlord has quit IRC
07:40:36 *** warlord has joined #gnucash
07:40:36 *** gncbot sets mode: +o warlord
07:42:28 <chris> I'll upgrade guile-json in maint to clean up the PR
07:43:26 *** nimish2711 has joined #gnucash
07:58:48 *** storyjesse has joined #gnucash
08:04:21 <chris> !@#$ style-sheet also saves as live scheme
08:08:31 *** boldstripe has joined #gnucash
08:15:08 *** nimish2711 has quit IRC
08:23:42 <gjanssens> yeah, plenty of surprises like that everywhere :(
08:26:19 *** warlord has quit IRC
08:29:32 *** warlord has joined #gnucash
08:29:32 *** gncbot sets mode: +o warlord
08:31:45 <warlord> chris, yeah, sorry about that. (not my fault, btw)
08:46:43 *** Aussie_matt has quit IRC
09:13:59 *** storyjesse has quit IRC
09:17:39 *** fabior has joined #gnucash
09:25:37 *** fabior has quit IRC
09:31:49 *** storyjesse has joined #gnucash
09:48:14 *** fabior has joined #gnucash
09:55:05 *** ArtGravity has joined #gnucash
09:55:05 *** ChanServ sets mode: +v ArtGravity
09:57:41 *** Mechtilde has quit IRC
10:03:24 *** Mechtilde has joined #gnucash
10:06:46 <chris> gjanssens it would be super nice if the Window menu could reflect the open tabs... I typically operate with 20+ tabs
10:07:21 <warlord> chris, it does for me.
10:07:50 <gjanssens> warlord: do you open each tab in a separate window ?
10:08:11 <warlord> I have registers open in separate windows, reports open as tabs
10:08:37 <warlord> The titlebar shows the current tab in that window
10:08:47 <gjanssens> Interesting, because my Window menu only lists individual windows, not all tabs in all windows
10:08:48 <warlord> And the window menu option reflects the titlebars
10:09:35 <gjanssens> chris: you *can* right-click the tab bar itself to get a list of open tabs in that window
10:09:36 <warlord> Oh, you want to see a NON-CURRENT tab in the window list?
10:09:51 <chris> ah right-click, forgot that
10:10:03 <warlord> The list does show all open tabs (where by "open" I mean "the currently visible tab in a window")
10:10:29 <gjanssens> Oh, right. Different interpretation of the request indeed
10:11:09 <chris> especially Ctrl-PgUp/PgDn doesn't consistently switch tabs, but this is minor
10:13:13 <gjanssens> BTW I have been experimenting with (load) instead of (resolve-interface) for the better part of the day now
10:13:38 <gjanssens> Guile certainly is picky in this area...
10:14:15 <gjanssens> I can use (load), but it insists on recompiling all the reports at gnucash start-up and ignores the pre-compiled files :(
10:14:39 <gjanssens> I tried load-from-path instead but it didn't make any difference
10:15:42 <gjanssens> (load-compiled) does work. The drawback there is it would force us to rerun the build each time when making report changes
10:16:14 <gjanssens> This is less nice when doing interactive hacking on reports (that is directly tweak the installed files for testing)
10:17:06 <gjanssens> The last thing I am about to check now is whether passing the .scm file extension has anything to do with this.
10:19:37 <warlord> Interesting...
10:20:41 <gjanssens> And yes. Guile is particularly picky about the file extension...
10:21:09 <warlord> Which means if you leave off the .scm extension...????
10:21:25 <gjanssens> Running (load-from-path gnucash/report/reports/standard/transaction.scm) will trigger an autocompile
10:21:36 <gjanssens> Running (load-from-path gnucash/report/reports/standard/transaction) won't
10:21:49 <gjanssens> (with the proper quoting of course)
10:21:59 <warlord> What if the .scm is newer?
10:22:50 <warlord> (will it recompile when it needs to?)
10:23:12 <chris> (load "./code.scm") seems to recompile when needed
10:23:26 <gjanssens> warlord: it will recompile
10:24:13 <gjanssens> chris: using the extension .scm will cause guile to ignore our own precompiled files an all cases and only check in the local file cache
10:24:45 <gjanssens> I don't know of (load "./code") will automatically try to load code.scm
10:24:58 <chris> it doesn't seem to
10:25:04 <gjanssens> I know load-from-path will try to add the extension if missing
10:25:23 <gjanssens> chris: ?
10:25:40 <chris> (load "./code") complains of missing filename
10:26:19 <gjanssens> Ok. That's what I suspected. But I can confirm load-from-path works with a missing extension
10:26:29 <warlord> chris, but (load-from-path "./code") works?
10:26:46 <chris> no because my current-dir is not in %load-path
10:27:02 <gjanssens> That would of course require code.scm to be somewhere in a %load-path directory
10:28:31 <gjanssens> Also note that switching to (load-from-path) from (resolve-interface) now causes 6 duplicate report errors.
10:28:43 * gjanssens still has to investigate why exactly
10:31:02 *** jervin has joined #gnucash
10:31:03 *** doubledev has joined #gnucash
10:31:04 <warlord> Hmmmm
10:31:33 *** doubledev has left #gnucash
10:38:28 <chris> (load-from-path "./rnrs") works
10:38:51 *** fabior has quit IRC
10:39:08 <gjanssens> Looks like calling (use-modules (some report)) from another report is the culprit of the duplicate report warnings
10:42:31 <warlord> hmm
10:44:56 <gjanssens> Yep. Removing (use-modules (gnucash report reports standard category-barchart)) ; for guids of called reports
10:45:27 <gjanssens> from net-charts.scm (after moving the guids to a common location) fixes 4 out of 7 of the duplicate report errors
10:45:56 <chris> and income-gst-statement.scm
10:46:23 <chris> ^ I may have to export from transaction.scm and allow income-gst to import them
10:47:52 <gjanssens> chris: yes that's another offender
10:47:56 *** guak has joined #gnucash
10:48:04 <gjanssens> What exactly does income-gst require from transaction ?
10:48:43 <chris> erm, the renderer and options-generator, i.e. everything meaty!
10:49:37 <gjanssens> wow :)
10:49:54 <gjanssens> Ok, so how would you change this ?
10:50:21 <chris> never thought about that. the messy option is to move them into transaction.scm
10:50:48 <chris> I'll have to experiment and see what can work
10:51:09 <gjanssens> Ok
10:53:07 <gjanssens> Perhaps (resolve-interface) instead of (use-module) can help, if used specifically from within income-gst
10:53:48 <gjanssens> What should be avoided is that (define-report) gets called more than once
10:54:30 *** jervin has quit IRC
10:54:35 <chris> :(
10:55:13 <gjanssens> Well, more than once per module
10:58:07 *** warlord has quit IRC
11:03:33 <chris> https://pastebin.com/raw/mS6veek1 this runs but still complains of guid duplicate
11:03:52 <chris> (in guile #; will neutralize the next s-exp)
11:04:36 *** fabior has joined #gnucash
11:05:38 * chris confused
11:11:28 *** fabior has quit IRC
11:13:19 <gjanssens> So resolve-interface still triggers execution of the (define-report) call at the end of the module :(
11:14:50 <gjanssens> I have also tried the @ form as described in use-modules here https://www.gnu.org/software/guile/manual/html_node/Using-Guile-Modules.html#index-use_002dmodules
11:15:10 <gjanssens> but even that one still triggers the execution of the (define-report) call
11:15:45 <gjanssens> In this case of course it wouldn't hurt as it's the exact same module being reloaded
11:18:00 <gjanssens> You can move the renderer and options generator into a separate file of course. That would also fix this particular issue.
11:21:45 *** storyjesse has quit IRC
11:22:27 *** nimish2711 has joined #gnucash
11:27:12 *** boldstripe_ has joined #gnucash
11:27:27 *** boldstripe has quit IRC
11:27:28 *** boldstripe_ is now known as boldstripe
11:28:58 <chris> I think moving into a report-utilities.scm type of file is the right answer here :(
11:35:55 * chris wonders why existing standard-reports.scm doesn't trigger duplicate guids
11:36:56 <chris> general-ledger.scm also suffers
11:43:22 <chris> but ultimately it would make sense that trep-renderer at least gets upgraded, it's now a query engine
11:49:02 <gjanssens> chris: yes. It looks like my change is revealing some modularity inconsistencies
11:50:43 <gjanssens> Note I can get work around the duplicate guid errors by forcing a certain file load order
11:51:20 <gjanssens> If I force the code to load transaction before income-gst using load-from-path the errors go away
11:51:41 <gjanssens> It looks as if "load-from-path" forces a module (re)load
11:52:20 <gjanssens> Whereas this won't happen in case of calling use-modules. The latter will only cause a load if the module wasn't loaded yet
11:52:39 <gjanssens> That's probably why resolve-interface was used rather than load-from-path...
11:55:09 <gjanssens> I don't know if there's a way to test whether a file defines a module and only load the file if that module isn't known yet in guile...
11:57:21 <chris> module-ref ?
11:57:31 <gjanssens> How would that work ?
11:57:40 <chris> (module-ref module 'gnc:define-report) or something?
11:57:48 * chris flying blind here
11:58:45 <gjanssens> Hmm, I don't see how that would work.
11:59:23 <gjanssens> It tests on an already loaded module, but we haven't found a way yet to extract the details we look for from the unloaded file
12:00:56 <gjanssens> Perhaps I'm overcomplicating things here and I should just stick with the (resolve-interface) code that was present
12:01:12 *** jervin has joined #gnucash
12:01:35 <gjanssens> I had hoped to get to one single load function for both user created reports and gnucash provided ones, but that may be stretching it
12:02:14 <gjanssens> Instead I'll simply force loading user generated reports first and the proceed to resolve-interface to load the gnucash provided reports from various directories
12:03:17 <gjanssens> This would equally allow the user to replace gnucash provided reports with a hand-written one -
12:04:09 <gjanssens> If the user puts a modified version of transaction.scm in GNC_CONFIG_HOME/reports without altering the guid and define-module line
12:04:40 <gjanssens> that report would be loaded before the resolve-interface would reach gnucash' own transaction.scm
12:05:15 <gjanssens> As the module is already defined (by user) it won't try to load the one from gnucash
12:06:39 <chris> ok sounds like an interesting development; caveat emptor as usual
12:06:43 <gjanssens> And the big question now is whether that's really a good idea...
12:07:11 <gjanssens> It's certainly very convenient for someone who understands what's going on
12:07:33 <gjanssens> But it can equally be very confusing for someone that doesn't or barely does
12:10:58 <gjanssens> Especially since we wouldn't even get a warning any more
12:11:21 <gjanssens> The vetting would happen deep inside guile logic, not in our own code
12:12:00 <chris> it may be easier to keep the current mechanism and move all business/locale report into the same loading mechanism
12:13:21 <gjanssens> :)
12:13:52 <gjanssens> Yeah, I was hoping to tackle several improvements in one go (though in separate commits)
12:14:14 <gjanssens> Moving those reports will still continue
12:15:05 <gjanssens> The other part was about simplifying (from a user's perspective) the inclusion of user made reports
12:15:13 <chris> reading guile's source code espeically boot-9.scm is rather enlightening
12:15:49 <chris> yeah about loading reports we can add a dirsweep over GNC_CONFIG_HOME/reports and load up everything
12:16:03 <gjanssens> Exactly
12:16:22 <gjanssens> The only detail there is what to do with duplicate report guids ?
12:16:59 <gjanssens> Right now I believe it's best to stick with the current policy: print an error and skip the dupe
12:17:06 *** jervin has quit IRC
12:17:07 <chris> well I can amend report.scm to generate a new one, and perhaps rename it as well?
12:17:39 <gjanssens> We can document your suggestion of first undefining the report if people insist on overwriting a standard one
12:17:48 <chris> the report will be usable (can run) but unusable (can't save!)
12:18:09 <gjanssens> They can add such a line right before the define-report call in their own report
12:18:32 <gjanssens> I prefer the user should take an explicit action to override gnucash provided reports
12:18:52 <gjanssens> If we document how, it should not be that hard if a user needs this
12:19:02 <chris> ok
12:20:03 <chris> well it's bed time here
12:20:05 <gjanssens> At least if it works. Did you try it ?
12:20:29 <chris> no but report.scm defines it well, and it's sound
12:20:36 <chris> i'll try next few days
12:20:44 <gjanssens> Ok
12:20:48 <gjanssens> Good night
12:21:03 *** nicoroumeau has joined #gnucash
12:29:12 *** nicoroumeau_ has joined #gnucash
12:29:22 *** nicoroumeau has quit IRC
12:29:52 *** nicoroumeau_ is now known as nicoroumeau
12:32:11 *** fabior has joined #gnucash
12:35:48 *** fabior has quit IRC
12:59:44 *** bizaff has quit IRC
13:01:29 *** fabior has joined #gnucash
13:04:12 *** fabior has quit IRC
13:08:19 *** Robert847 has joined #gnucash
13:10:15 *** ChanServ sets mode: +v Robert847
13:11:16 *** boldstripe has quit IRC
13:12:16 *** boldstripe has joined #gnucash
13:14:28 <Robert847> Hi, I am still using release 2.6.17 in Ubuntu 18.04, not quite ready to try 3.5. However, I get very frustrated by the tendency of the register to not observe where I am clicking my mouse button when trying to accept an edit in a transaction that has more than one line in the current account.
13:14:57 <Robert847> I am curious if this is improved in release 3.5
13:16:33 <Robert847> I might click in the next transaction up the list but the highlight ends up in a different transaction altogether afer the edit is committed.
13:19:10 <Robert847> I think that this was supposed to be addressed in the Register2 code.
13:25:47 *** Mechtilde has quit IRC
13:52:07 *** fabior has joined #gnucash
13:52:49 *** Mechtilde has joined #gnucash
13:53:40 <gjanssens> Robert847: the register2 code is still not in use in 3.5, so it's most likely the situation hasn't improved yet. I'm sorry.
13:56:40 *** Mechtilde has quit IRC
13:57:10 *** Mechtilde has joined #gnucash
14:01:58 *** oozer has quit IRC
14:05:09 *** calvinct has joined #gnucash
14:22:48 *** Robert847 has left #gnucash
14:24:20 *** ArtGravity has quit IRC
14:25:33 *** Mechtilde has quit IRC
14:44:57 *** nimish2711 has quit IRC
14:48:36 *** frakturfreak has joined #gnucash
14:48:36 *** ChanServ sets mode: +v frakturfreak
14:58:10 *** ChanServ sets mode: +v eric_
15:07:36 *** oozer has joined #gnucash
15:09:48 *** nimish2711 has joined #gnucash
15:11:58 *** boldstripe has quit IRC
15:13:25 *** boldstripe has joined #gnucash
15:50:31 *** nimish2711 has quit IRC
16:08:18 *** calvinct has quit IRC
16:50:21 *** fabior has quit IRC
17:09:26 *** frakturfreak has quit IRC
17:11:41 *** mikee has joined #gnucash
17:13:08 *** boldstripe has quit IRC
17:13:15 *** mikee has quit IRC
17:13:30 *** mikee has joined #gnucash
17:14:04 *** lmat has quit IRC
17:14:35 *** boldstripe has joined #gnucash
17:17:46 *** gjanssens has quit IRC
17:31:25 *** gncbot sets mode: +v mikee
17:31:37 <mikee> .
17:31:46 *** nicoroumeau has quit IRC
17:31:47 <mikee> @op
17:31:47 *** gncbot sets mode: +o mikee
17:32:32 <mikee> Hello all.
17:33:35 <mikee> This voice thing is new.
17:35:18 *** nicoroumeau has joined #gnucash
17:41:58 *** Aussie_matt has joined #gnucash
17:43:08 *** fell_laptop is now known as fell
17:44:10 <fell> Yes, mikee, after several spam bot attacks ...
17:44:24 *** markolino has joined #gnucash
17:44:25 *** ChanServ sets mode: +v markolino
17:44:39 <mikee> Understandable. It's been a while.
17:45:21 <fell> Nice, to see you here again.
17:45:38 *** nicoroumeau has joined #gnucash
17:45:59 <fell> BTW: will Wales stay in the EU or Briatnny?
17:46:07 <mikee> I've just pushed some code to GitHub that talks to the HMRC API for Making Tax Difficult.
17:46:47 <mikee> Ha!. Personally I'd link with Brittany.
17:47:00 <mikee> I like Europe.
17:47:23 <mikee> Anyways: https://github.com/EvansMike/hmrc_self_submit
17:48:40 <mikee> So do ops have to manually approve speakers?
17:50:38 <fell> No, it is auotmatic done after they register/identify against Nickserv on the join.
17:51:30 <fell> Typo: Diffcult
17:52:28 <mikee> @Fell OK. Thanks. I've forgotten how this all works :)
17:52:28 <gncbot> mikee: Error: "Fell" is not a valid command.
17:52:41 <mikee> .Fell OK. Thanks. I've forgotten how this all works :)
17:52:52 <mikee> .Fell OK. Thanks. I've forgotten how this all works :)
17:53:00 <mikee> See!
17:54:53 <fell> I hope, https://wiki.gnucash.org/wiki/IRC will explain everything important.
17:56:13 <mikee> Wales mostly voted to leave. And yet Wales probably received a lot more EU funding than much of England. Nuts.
17:56:33 <mikee> .me reads wiki
17:57:16 <mikee> again.
17:59:10 <fell> I wrote that mostly to have parts at hand, which I need only each couple of years.
18:00:54 <mikee> I do that with my blog. It's mostly just a reminder fo me. Nobody else read it. Apparently
18:26:24 *** dizzydeane has joined #gnucash
18:39:05 *** dizzydeane has quit IRC
19:02:53 *** BF has quit IRC
19:04:40 *** dizzydeane has joined #gnucash
19:21:18 *** fell has quit IRC
19:25:04 *** fell has joined #gnucash
19:25:04 *** ChanServ sets mode: +o fell
19:59:30 *** boldstripe has quit IRC
20:00:25 *** boldstripe has joined #gnucash
20:22:13 *** fell has quit IRC
20:24:36 *** fell has joined #gnucash
20:24:36 *** ChanServ sets mode: +o fell
20:47:20 *** nimish2711 has joined #gnucash
20:54:47 *** markolino has quit IRC
21:53:10 *** markolino has joined #gnucash
21:53:10 *** ChanServ sets mode: +v markolino
21:55:27 *** oozer has quit IRC
21:56:01 *** oozer has joined #gnucash
22:00:10 *** boldstripe has quit IRC
22:01:36 *** boldstripe has joined #gnucash
22:04:36 *** markolino_ has joined #gnucash
22:05:03 *** guak has quit IRC
22:06:59 *** markolino has quit IRC
22:07:37 *** oozer has quit IRC
22:08:01 *** markolino_ has quit IRC
23:05:45 *** tonysoar has joined #gnucash
23:28:26 *** Agfarmer18 has quit IRC
23:31:06 *** tonysoar has quit IRC