2017-09-01 GnuCash IRC logs

00:35:13 *** storyjesse has joined #gnucash
00:37:31 *** Mechtilde has joined #gnucash
00:55:52 *** tuxd00d has joined #gnucash
01:19:55 *** Mechtilde has quit IRC
02:11:20 *** jotrago has quit IRC
02:14:09 *** storyjesse has quit IRC
02:21:03 *** gjanssens has joined #gnucash
02:21:03 *** ChanServ sets mode: +o gjanssens
02:21:19 <gjanssens> .
02:25:06 *** karelk has quit IRC
02:28:15 <tuxd00d> Are we working towards a full multiuser environment? I’ve been out the development loop for a while.
02:43:56 *** shoonya has joined #gnucash
02:51:20 *** hoijui has joined #gnucash
02:56:25 *** fabior has joined #gnucash
03:45:16 *** jotrago has joined #gnucash
03:55:03 *** jotrago has quit IRC
03:56:26 *** jotrago has joined #gnucash
04:15:20 *** jotrago1 has joined #gnucash
04:16:08 *** jotrago has quit IRC
04:16:08 *** jotrago1 is now known as jotrago
04:21:45 *** jotrago1 has joined #gnucash
04:22:08 *** jotrago has quit IRC
04:22:09 *** jotrago1 is now known as jotrago
04:31:16 *** mrklintscher18 has joined #gnucash
04:42:13 *** Artefact2 has joined #gnucash
04:43:05 *** jotrago1 has joined #gnucash
04:43:27 *** jotrago has quit IRC
04:43:28 *** jotrago1 is now known as jotrago
04:45:13 <gjanssens> tuxd00d: yes, though it will take us many years still at the current pace
04:45:40 <gjanssens> We're converting the project to c++ as a prerequisite
04:46:03 <gjanssens> During that refactoring the multi-user end goal is kept in mind all the time
04:46:11 <tuxd00d> Awesome
04:46:45 <tuxd00d> That’s a big task.
04:47:12 <gjanssens> It is. Help is always welcome :)
04:47:56 <gjanssens> The second goal is to move away from glib in the core libraries to improve portability
04:48:31 <gjanssens> So wherever possible we use C++11 features supplemented by the boost library
04:48:45 <tuxd00d> I wish I could devote some time to it right now, as I really like the project. But my c++ is rusty. If it were C…
04:49:50 <gjanssens> That's ok. I can't resist to try each time again though ;)
04:50:35 <gjanssens> My C++ was rusty as well by the way. Before we chose to port to it it had been since 2000 that I had written my last line of C++
04:50:53 <gjanssens> So I had some studying to do to get up to speed with modern C++ standards
04:51:10 <gjanssens> I'm still learning each time I work on the code
04:52:14 <tuxd00d> I may be able to through in my two cents now and then. But I’m spread really thin at the moment.
04:52:17 <gjanssens> But maybe you can help in testing. We're close to releasing our first 2.7 development snapshot
04:52:29 <tuxd00d> It’s was the mid 90’s for me.
04:53:19 <tuxd00d> If it’s on linux or OSX, I can help test. I don’t do Windows (I value my time).
04:53:33 <tuxd00d> ;)
04:53:44 <gjanssens> We're currently sidetracked by external dependency issues. Early May wee learned webkitgtk1 was about to be deleted from most distributions
04:54:17 <gjanssens> So we had to replace it urgently. With that we had to switch from gtk2 to gtk3 as well
04:54:21 <tuxd00d> That’s a headache.
04:54:47 <gjanssens> Those were pretty huge changes to be done so close to alpha/beta.
04:55:08 <tuxd00d> Absolutely
04:55:17 <gjanssens> Testing on linux and OS X will be highly appreciated :)
04:56:18 <gjanssens> On those two platforms things are pretty much ready for release. On Windows unfortunately we're still trying to get it to build and run with the new dependencies...
04:56:40 <gjanssens> (Ready barring bugs still to be discovered of course)
04:59:48 <tuxd00d> Why are we using bugzilla instead of githubs bug tracker… just curious.
05:02:25 <tuxd00d> How do you pronounce Geert?
05:35:08 <gjanssens> tuxd00d: we are using bugzilla because we have been doing that long before github came to life
05:35:32 <tuxd00d> Makes sense
05:35:38 <gjanssens> We may have to change this soon because the gnome project is likely moving away from it
05:36:46 <gjanssens> jralls did a first evaluation on the gnucash-devel mailing list end of July
05:37:27 <gjanssens> And github's bugtracker was missing some features we are relying upon.
05:38:00 <gjanssens> Nothing has been decided yet though. We're focussing on getting 2.7 ready
05:39:02 <gjanssens> As for my name that's difficult for non-native speakers :(
05:39:25 <tuxd00d> Over at Asterisk we are using JIRA
05:39:58 <gjanssens> 'G' in Flemish is a sound somewhere in between 'k' and 'h'. I haven't found an English word yet that uses that sound.
05:40:21 <gjanssens> 'ee' is 'a' in 'cale'
05:40:24 <tuxd00d> What about Spanish?
05:40:46 <tuxd00d> Or French?
05:40:46 <gjanssens> 'r' is again something never pronounced in English
05:41:13 <gjanssens> the 'G' is very typical to Flemish. The French have trouble with my name as well.
05:41:47 <gjanssens> It's close to the Spanish 'J' in 'Jesus', but not exactly the same :)
05:42:21 <gjanssens> Even Dutch people have trouble with it and we're supposed to have the same formal language :\
05:43:20 <tuxd00d> In my dutch translator it sounds like “key ert”
05:43:43 <gjanssens> Jira was mentioned as well, but not not yet evaluated. It may be an option
05:44:07 <gjanssens> Your Dutch translator ?
05:44:36 <tuxd00d> I don’t know what Jira’s open source project pricing is.. but it seems very customizable… but I’ve never been on the server side.
05:44:50 <tuxd00d> https://translate.google.com/#nl/es/geert
05:45:15 *** fekepp has joined #gnucash
05:45:45 <gjanssens> Heh, that's how the Dutch pronounce it indeed :)
05:46:08 <gjanssens> The 'r' is different in flemish. Closer to the french 'r'
05:46:48 <tuxd00d> It’s almost 3 a.m. … need to get some sleep. It was nice talking with you. I’ll see what I can do to help out with the project.
05:47:07 <gjanssens> That's wonderful. Sleep well !
05:55:46 *** fekepp has quit IRC
06:40:20 *** User has joined #gnucash
06:45:25 *** hoijui has quit IRC
06:53:35 *** Jimraehl1 has joined #gnucash
06:54:39 *** fabior has quit IRC
06:54:45 *** Jimraehl1 has left #gnucash
07:02:54 *** fekepp has joined #gnucash
07:18:41 *** fekepp has quit IRC
07:31:28 *** pilotauto has quit IRC
07:57:02 *** fekepp has joined #gnucash
07:59:10 *** jotrago has quit IRC
08:01:20 *** Chris has joined #gnucash
08:01:48 *** fekepp has quit IRC
08:07:00 *** hoijui has joined #gnucash
08:22:59 *** shoonya has quit IRC
08:27:45 *** hoijui has quit IRC
08:36:37 *** mrklintscher18 has quit IRC
08:40:48 *** mrklintscher18 has joined #gnucash
08:53:10 *** Chris has quit IRC
09:21:33 *** jotrago has joined #gnucash
09:43:47 *** fabior has joined #gnucash
10:23:31 *** kael has joined #gnucash
10:46:29 *** bertbob has quit IRC
10:57:48 *** rickoehn has quit IRC
10:57:53 *** rickoehn has joined #gnucash
11:00:15 *** bertbob has joined #gnucash
11:11:38 *** karelk has joined #gnucash
11:16:13 *** fekepp has joined #gnucash
11:30:51 *** kael has quit IRC
11:46:40 *** Mechtilde has joined #gnucash
12:01:26 <gjanssens> jralls: Is there somewhere I can pull your latest work on the Windows build ? Just in case I find some time this WE to look at it...
12:05:28 <jralls> gjanssens: I got gnc-filepath-utils fixed and rebased my dodge-boost-filesystem branch and the stupid build broke with a problem compiling gettext.scm: libgncmod-app-utils won't link "The specified module could not be found".
12:06:13 <jralls> If I don't get that fixed by the time I have to leave (about 2.5 hours) I'll just push what I've got onto master and you can have a go at it.
12:06:47 *** codesmythe has joined #gnucash
12:07:04 <gjanssens> Ok. Perhaps codesmythe's PR may fix it (if it's a build sequence issue it would)
12:07:19 <gjanssens> Speaking of the devil... :)
12:07:26 <codesmythe> Ha
12:08:04 <jralls> It doesn't seem to be a sequence issue. The error is from the M$ dynamic linker... the file is there.
12:08:17 <gjanssens> codesmythe: while I think your PR is a welcome improvement I wonder if it's the best approach
12:08:53 <jralls> So I'm about to point Dependency Walker at it.
12:09:44 <gjanssens> The way it's done means a dev that creates a new gncmodule has to remember to add it to the list in a file that has nothing to do with the new gncmodule. And the same when removing a module (which will be the more likely use case in the future)
12:10:13 <gjanssens> In the latter case I presume one would get cmake errors due to missing dependencies though
12:10:42 <gjanssens> Is there a way to assemble this similar to how the dist list is assembled ?
12:11:09 <gjanssens> Perhaps I'm being too pedantic though... I don't know. This was my first reflection when I saw your PR
12:15:56 *** User has quit IRC
12:17:15 <codesmythe> gjanssens: I'm open to doing this a better way. I don't think assembling would work because it is not the case that we just need to specify the gncmod that the particular scheme file needs. The way gnc-module.c works, it creates a situation that every scheme files is implicitly dependent on every libgnc.*so file. I think. :-)
12:19:19 <codesmythe> I could move the definition of GNCMODULE_TARGETS to the toplevel CMakeLists.txt file to make it more visible.
12:20:04 <codesmythe> gjanssens: Hold on ... I misunderstood what you meant about assembling ... let me think about that.
12:26:24 <gjanssens> codesmythe: yeah as a non-native speaker I sometimes express myself in odd ways...
12:27:34 <codesmythe> gjanssens: No, in this case it was me having a predefined notion of what 'assembling' should mean without carefully reading what you said. :-)
12:32:33 *** kael has joined #gnucash
12:35:51 <jralls> codesmythe is identifying a legitimate problem, though: The modules are not well isolated. I found that before I'd fixed gwen to use Gtk3 that libgncmod-aqbanking was getting linked to everything else and causing Gtk2 vs. Gtk3 runtime errors.
12:39:22 <codesmythe> gjanssens: On assembling ... let's say CMake is processing the GNC_ADD_SCHEME_TARGETS macro in src/app-utils/CMakeLists.txt. That is where it needs the list of gncmod targets to depend on. But CMake would not have processed any of the directories after app-utils yet so the gncmod list would not be fully populated at that point.
12:39:59 <codesmythe> gjanssens: The DIST assembly stuff works because it runs after all the subdirectories have been processed.
12:41:44 <gjanssens> codesmythe: Ok, understood. So that would not work indeed for the scheme dependencies.
12:42:19 <jralls> codesmythe: The Scheme files need targets that can have dependencies so that CMake knows in what order to build things.
12:45:04 <gjanssens> jralls: the module isolation is poor indeed
12:46:12 <gjanssens> And because each module gets an opportunity to initialize lots of other stuff, including loading scheme modules, it's become practically impossible to untangle the dependency spaghetti we currently have.
12:46:44 <gjanssens> As for targets for dependencies isn't that more or less what codesmythe's PR is doing ?
12:47:01 <gjanssens> It adds all gnc-modules as dependencies to each scheme file being added.
12:47:09 <codesmythe> jralls: The is a dependency mechanism in place as GNC_ADD_SCHEME_TARGETS takes in a dependency list. In the PR, I added all of the gncmod modules as dependencies in that macro, instead of passing the list in on every invocation. I agree with gjanssens that this hides a list deep in the minutia of compiling the scheme files.
12:47:09 <gjanssens> Or would you like more fine-grained control ?
12:48:43 <jralls> I would prefer more fine-grained control.
12:50:05 <jralls> I don't think that GncAddSchemeTarget works the way codesmythe thinks it does. It creates an environment for running guild on the scheme file but AFAICT doesn't pass any dependency info to Cmake.
12:52:36 <codesmythe> https://github.com/Gnucash/gnucash/blob/maint/src/cmake_modules/GncAddSchemeTargets.cmake#L126
12:53:55 <jralls> That's add_custom_command, not add_custom_target. IIUC correctly Cmake builds its dependency tree on targets, not commands.
12:54:30 *** bertbob has quit IRC
12:57:17 <jralls> So it needs an add_custom_target that depends on the shared libs and that runs the custom command built beginning at line 113.
12:58:15 <gjanssens> Hah, masters at work here :D
12:58:16 <codesmythe> From https://cmake.org/cmake/help/v3.2/command/add_custom_command.html, "If DEPENDS specifies any target (created by the add_custom_target(), add_executable(), or add_library() command) a target-level dependency is created to make sure the target is built before any target using this custom command."
12:58:55 *** mrklintscher18 has quit IRC
12:59:26 *** rubdos has joined #gnucash
12:59:31 <jralls> Well, it ain't working.
13:02:09 <codesmythe> I certainly may have the full dependency graph wrong for the scheme stuff, but I think add_custom_command would do what we want if we correctly told it what to do. :-)
13:04:34 <codesmythe> As long as gnc-module.c works the way it does, we would still have these 'file too short' problems even if we had the dependency graph exactly correct.
13:06:05 <jralls> I haven't seen "file too short", I've seen "file not found" because it isn't built yet.
13:08:22 <lmat> gjanssens: "between 'k' and 'h'" perhaps "h" in human is close? "ch" in "loch"?
13:08:59 *** bertbob has joined #gnucash
13:10:43 <jralls> A Dutch friend told me that "G" is like "Ch" in yiddish. gjanssens told me when we met F2F that Flemish is softer than Dutch so I think of his G as H with just a touch of k.
13:12:08 *** bertbob has quit IRC
13:12:41 <gjanssens> jralls: I love your description (G=H with a touch of k)
13:12:59 <jralls> But is it right?
13:13:05 <gjanssens> Yes
13:13:43 <gjanssens> The "ch" in "loch" is too soft
13:14:43 <gjanssens> On a (Flemish) spectrum you'd have "h" "ch"(Flemish) "g"(Flemish) "ch"(Yiddish) "k"
13:15:32 <gjanssens> And if you want you could add the French "G" (like in "Guitare") between "ch"(Yiddish" and "k"
13:16:51 <codesmythe> gjanssens: Do you see 'file too short' on Fedora?
13:25:33 <codesmythe> The other question I wanted to ask was about getting the CMake build to work on Travis. As I understand it, Ubuntu 14.04 has cmake 2.8 and we need at least 3.0. There are three ways that we could get cmake 3.0: a) use a PPA such as https://launchpad.net/~george-edison55/+archive/ubuntu/cmake-3.x, (b) download the source from cmake.org and compile it, or (c) use the precompiled binaries on cmake.org, assuming they work on Ubunut 14.
13:25:34 <codesmythe> s there a preferred approach?
13:28:11 *** bertbob has joined #gnucash
13:35:34 <gjanssens> codesmythe: I get "file too short" as well.
13:36:02 <lmat> codesmythe: I would think PPA is the most straightforward option.
13:36:15 <gjanssens> As for getting cmake in Travis, I'm fine with a PPA. I believe this is supported in Travis.
13:37:08 <gjanssens> https://docs.travis-ci.com/user/installing-dependencies/#Installing-Packages-from-a-custom-APT-repository
13:38:10 <gjanssens> I didn't know how to find a PPA that provides cmake. You apparently do.
13:40:13 <gjanssens> In the long run we should consider generating docker files for all the environments we want to test against.
13:40:57 <gjanssens> That's the new approach suggested by Travis. It doesn't appear they will provide a newer environment than Trusty anytime soon
13:41:17 <gjanssens> But I'd be happy if we could run both cmake and autotools tests already
13:41:35 <gjanssens> For cmake, should we test both the make and the ninja generator ?
13:42:04 <jralls> The more tests the better...
13:44:36 <lmat> gjanssens: Yes, docker files would be great so that I can recreate the CI build locally and avoid turnaround time on pull requests ^_^ Not that I do enough of them to matter much, but it was a real hasstle before.
13:47:38 <gjanssens> codesmythe: do you plan to adjust .travis.yml ?
13:48:03 *** Cuare has quit IRC
13:48:04 <gjanssens> lmat: I have no experience yet with docker. So this will become an interesting project ;)
13:49:11 <codesmythe> gjanssens: Yeah, I'll submit a PR for that.
13:50:39 <gjanssens> codesmythe: Wonderful. I was thinking that perhaps we should extract the script line into a separate file in the repo (something like travis-run.sh)
13:50:52 *** User has joined #gnucash
13:51:13 <gjanssens> If we add options for cmake in two variants it would become a bit unwieldy to put it all in one line && together
13:51:37 <codesmythe> gjanssens: Ok, I'll do that.
13:51:50 <gjanssens> Good
13:53:57 <gjanssens> And if I can make one more request: I propose to store in in util or test-core or something similar to avoid cluttering the toplevel directory.
13:54:26 <codesmythe> ok
13:55:02 <gjanssens> Will probably be util for now as test-core is a subdir or common, which is probably hidden too far away
13:55:07 <gjanssens> Tx
13:55:44 <codesmythe> I'm going to withdraw PR #178 and think about it some more.
13:56:37 <gjanssens> Ok
13:57:36 <codesmythe> Thanks all for the discussion.
14:00:49 <gjanssens> You're welcome, thanks for working on the build system :)
14:02:25 *** frakturfreak has joined #gnucash
14:20:32 *** rubdos has quit IRC
14:25:53 *** kael has quit IRC
14:26:59 *** fekepp has quit IRC
14:27:02 *** fekepp has joined #gnucash
14:27:42 *** hoijui has joined #gnucash
14:29:04 <jralls> gjanssens: It's time for me to go and I didn't find the problem, but I did determine that it's being caused by the first boost_filesystem change. I've pushed everything to code including fixing the bogus branch in gnucash.modules. Be back Monday.
14:29:12 *** jralls is now known as jralls_afk
14:50:56 *** fabior has quit IRC
14:53:48 *** hoijui has quit IRC
14:54:19 <gjanssens> Have a nice trip jralls_afk
15:13:01 *** hoijui has joined #gnucash
15:25:01 *** Mechtilde has quit IRC
15:32:59 *** frakturfreak has quit IRC
15:34:08 *** frakturfreak has joined #gnucash
15:36:46 *** ArtGravity has joined #gnucash
15:37:53 <ArtGravity> Can the Mortgage & Loan Repayment Assistant be used to generate scheduled transactions on a loan account that has already been serviced manually?
15:52:48 <warlord> ArtGravity: I believe it can; you just have to tell it how many payments are left (or how many have already been made). Note that it does NOT handle over-payments.
15:54:48 <ArtGravity> Neat. Thanks. I'm pretty excited to test it out.
16:01:03 *** hoijui has quit IRC
16:15:08 <ArtGravity> Any idea what I should do when the amount paid each month is showing too low in the verify screen before committing?
16:15:34 <ArtGravity> I think there is some cushion payment that just stays in the escrow account that isn't being factored into the repayment schedule.
16:19:23 <warlord> How much too low is it?
16:23:23 <ArtGravity> actually the more I look at it the more confused I am
16:23:56 <ArtGravity> the payment is short by roughly the escrow amount
16:24:32 <ArtGravity> not sure if I am setting up the tax and insurance payments wrong possibly
16:25:01 <ArtGravity> tax gets paid into escrow 12 times a year and comes out of escrow twice a year
16:25:11 <ArtGravity> insurance goes in 12 out 1
16:26:36 <ArtGravity> figuring out which amount goes int the escrow transactions is unclear to me
16:26:54 <ArtGravity> 50% of the tax since that is what gets paid out?
16:27:15 <ArtGravity> 1/12 and just set is at part of payment transaction?
16:27:45 <ArtGravity> but then where do I see the payment schedule for it to come out of escrow?
16:29:04 <ArtGravity> or maybe it doesn't do the split transaction the way I am expecting it
16:29:27 <warlord> I *think* you need to supply your monthly escrow input.
16:29:46 <warlord> (it's been a long time since I've looked at that assistant)
16:29:46 <ArtGravity> Right now I do I split entry by hand every month for principal/interest/escrow
16:30:28 <ArtGravity> OK, so maybe it isn't calculating my payments out of escrow
16:31:44 <ArtGravity> I expect to still need to verify amounts every month. I'm mostly just looking to to get a a templated transaction entered automatically
16:33:22 <ArtGravity> I should probably just run one all the way through on a clean set of accounts to see what the final product looks like
16:39:56 *** gjanssens has quit IRC
16:47:43 <ArtGravity> That was actually pretty educational, and I see why it was confusing me.
16:48:19 <ArtGravity> It routes all money though the escrow account, which differs from how I split the transactions
16:51:44 <warlord> Could be.
16:51:50 * warlord has to run...
16:51:55 <warlord> have a good (holiday) weekend.
16:51:58 <ArtGravity> thx for the help
16:52:04 <ArtGravity> I think I got it
16:52:09 <ArtGravity> you too
17:18:21 *** mrklintscher18 has joined #gnucash
17:41:41 *** fekepp has quit IRC
17:56:52 *** rickoehn has quit IRC
18:02:20 <codesmythe> @tell gjanssens Turns out we don't need a PPA for Travis/trusty. We can just install the cmake3 package.
18:02:20 <gncbot> codesmythe: The operation succeeded.
18:13:24 *** tuxd00d has quit IRC
18:58:44 *** frakturfreak has quit IRC
19:52:52 *** codesmythe has left #gnucash
20:29:57 *** ArtGravity has quit IRC
21:39:06 *** fiddlerwoaroof has quit IRC
22:30:41 *** fiddlerwoaroof has joined #gnucash
22:48:05 *** fiddlerwoaroof has quit IRC
23:38:28 *** User has quit IRC
23:50:38 *** O01eg has quit IRC