2020-04-20 GnuCash IRC logs

00:02:36 *** omnireq has quit IRC
00:02:38 *** omnireq_ has joined #gnucash
00:23:48 *** omnireq_ has quit IRC
00:23:52 *** Gerd has quit IRC
00:23:58 *** omnireq_ has joined #gnucash
00:26:23 *** bertbob has joined #gnucash
00:26:24 *** ChanServ sets mode: +v bertbob
00:40:42 *** bertbob has quit IRC
00:46:23 *** bertbob has joined #gnucash
00:46:24 *** ChanServ sets mode: +v bertbob
00:57:03 *** Mechtilde has joined #gnucash
01:00:05 *** Mechtilde has quit IRC
01:04:22 *** sbluhm has joined #gnucash
01:06:53 *** fell has quit IRC
01:08:12 *** fell has joined #gnucash
01:08:12 *** ChanServ sets mode: +o fell
01:19:18 *** Mechtilde has joined #gnucash
01:22:14 *** sbluhm has quit IRC
01:36:35 *** suukim has joined #gnucash
01:37:10 *** jervin has joined #gnucash
01:38:56 *** sbluhm has joined #gnucash
01:38:56 *** ChanServ sets mode: +v sbluhm
01:39:19 *** frakturfreak has quit IRC
01:43:55 *** jervin has quit IRC
01:53:50 *** frakturfreak has joined #gnucash
01:53:50 *** ChanServ sets mode: +v frakturfreak
02:05:08 *** Gerd has joined #gnucash
02:18:44 *** JayC has quit IRC
02:20:53 *** JayC has joined #gnucash
02:20:53 *** ChanServ sets mode: +v JayC
02:30:58 *** Mechtilde has quit IRC
02:34:33 *** Han has joined #gnucash
02:34:34 *** Gerd has quit IRC
02:51:32 *** hussam has joined #gnucash
02:51:32 *** ChanServ sets mode: +v hussam
03:15:18 *** Han has quit IRC
04:10:37 *** Mechtilde has joined #gnucash
04:15:02 *** Gerd has joined #gnucash
04:44:38 *** Gerd has quit IRC
04:45:59 *** Gerd has joined #gnucash
05:00:05 *** Aussie_matt has quit IRC
05:11:20 *** Gerd has quit IRC
05:36:18 *** Aussie_matt has joined #gnucash
05:47:05 *** gjanssens has joined #gnucash
05:47:05 *** ChanServ sets mode: +o gjanssens
05:49:34 <fell> jralls, about https://wiki.gnucash.org/wiki/index.php?title=Setting_up_OFXDirectConnect&diff=next&oldid=16127: IIRC it was fixed in 6.1.4.
05:50:00 <fell> in git after 6.1.4
05:54:27 <chris> gjanssens: quick question... is it difficult to generate a flat package for adrien.m?
05:55:05 <warlord> chris, what do you mean by "a flat package"?
05:55:14 <chris> would I need to generate a binary in code?
05:55:17 <chris> flathub/flatpak
05:55:41 <chris> wish to do further testing without necessarily polluting the git repo with numerous beta commits
05:55:43 <warlord> chris, we have a daily flatpak build already... is that not sufficient?
05:55:55 <warlord> oh, for a branch..
05:56:27 <chris> i.e. don't want to push my beta commits too often
05:56:46 <chris> I'm sure you'd be annoyed if I commit/revert/commit/revert repeatedly
05:56:53 <warlord> code isn't set up to build arbitrary branches. Or rather, the gnucash-on-flatpak can support it, but there is no infrastructure to allow submission of a request to build an arbitrary branch
05:57:35 <warlord> I could set it up, but you'd still have to have the code pushed to the code repo to work.
05:57:56 <warlord> Or you could just use the gnucash-on-flatpak repo yourself to build a package?
05:58:58 <chris> the 2nd option would require the beta tester take a leap of faith and trust my builds?
05:59:22 <chris> and the 1st option would simply require code to generate builds from an arbitrary branch?
06:00:16 <chris> it's all for https://bugs.gnucash.org/show_bug.cgi?id=797659
06:15:38 <gjanssens> chris: the first option is equally a leap of faith.
06:15:51 <gjanssens> The beta tester has to trust warlord and his infrastructure.
06:16:16 <gjanssens> I think setting up your own flatpak build would be the better approach here.
06:16:49 <gjanssens> It's clearly for testing which means a beta tester is aware of the risk involved
06:17:05 <chris> ok, it means generate a large binary and they get from me?
06:17:24 <gjanssens> Almost
06:18:03 <gjanssens> Flatpak works with repositories, so you would upload your build artefacts to a public location (the build artefacts are the repository)
06:18:13 <gjanssens> And the beta tester can install flatpak from that repo then
06:18:19 <chris> to dl from code is slightly more officious lol
06:19:05 <chris> i can try generate build nonetheless
06:30:28 <warlord> gjanssens, chris I dont mind setting up code to build from your branch, but it does mean you need to push your branch to code (for all to see)
06:34:13 <chris> that's the idea, full transparency
06:35:08 <chris> it'd have a long commit trail, and when it's confirmed working, fixup all and commit to maint
06:35:31 <chris> it'be a shame to have this bug still open for 3.x final release
06:39:29 <warlord> I dont mind updating code to build a dev branch. gjanssens -- what say you?
06:39:50 <warlord> chris, note that it's just a daily build, not on-demand
06:41:06 *** User has joined #gnucash
06:42:00 *** phoenix has joined #gnucash
06:43:29 <chris> of course daily is fine... and can be wiped after, say, 30days
06:47:45 <warlord> wiping is not automatic at this point ;)
06:48:00 <warlord> (unless gjanssens added that to the -on-fp repo code)
06:48:58 <warlord> Just tell me the branch name and I'll add it to the script. Currently only maint and master get built.
06:49:19 <warlord> (also note that if a branch fails, it wont build subsequent branches)
06:51:47 <chris> I'd go for e.g. maint-beta-797659 so maint-beta-* and master-beta-*.... or if not, maint-beta
06:52:33 <warlord> I dont think I can do a glob on the name, so I need the exact branch name.
06:53:28 <chris> then a simple beta is best
06:54:58 <warlord> Doesn't matter to me. You and gjanssens can battle it out and tell me the result. Next build is in 18 hours.
06:55:11 <warlord> So you've got ~12 to let me know (before I go to bed tonight)
07:07:05 *** Gerd has joined #gnucash
07:07:32 *** phoenix has quit IRC
07:08:24 *** KaiForce has quit IRC
07:14:48 <gjanssens> I'm not too comfortable adding feature branches to our main repo.
07:14:58 <gjanssens> In the main repo we don't alter history.
07:15:20 *** Aussie_matt has quit IRC
07:16:04 <gjanssens> That beta branch on the other hand would be a candidate for altering history (rebasing, squashing,...)
07:16:27 <gjanssens> That's not something for a public repo.
07:16:47 <gjanssens> Before we do we need at least jralls opinion on this as well.
07:17:34 <gjanssens> I have to leave now for at least 6 hours. I'll see if I find a moment to respond later still.
07:18:00 *** gjanssens has quit IRC
07:20:57 <chris> I actually agree... if the beta branch can't be force-pushed at leisure then it'll be far too messy
07:22:20 <warlord> The alternative is that I could set up another branch build that pulls from github?
07:22:47 <chris> if that works, it'd be great :)
07:23:22 <chris> so I'd be pushing to github.com/gnucash/gnucash~beta
07:23:38 <chris> maybe wait till gjanssens is back or jralls. no rush at all.
07:25:13 <chris> I might as well take the plunge and build flatpak
07:31:43 *** Gerd has quit IRC
07:32:33 <warlord> I am not going to set it up until I hear back.
07:44:49 *** Gerd has joined #gnucash
07:48:03 <warlord> chris, technically I could even pull off github.com/<yourname>/gnucash ;)
07:48:14 <warlord> with whatever branch you want.
07:56:09 *** Mechtilde has quit IRC
07:59:18 <chris> now you're just spoiling me :) -- I think ideally the script would pull github.com/gnucash/gnucash:beta and build, and skip for today if :beta is not present
07:59:41 <chris> but best await gjanssens/jralls to offer views too
08:07:30 <warlord> So gnucash/gnucash-beta would be its own fork of the gnucash repo that is only in github? *ponders*
08:08:57 <chris> no, beta would be a branch in gnucash/gnucash repo?
08:09:58 <warlord> I dont think that would work, because the gnucash/gnucash repo is controlled by code.
08:10:33 <warlord> Stepping away for a bit to go (finally) grab breakfast.
08:19:25 <chris> enjoy breakfast :)
08:19:47 <chris> just created on github.com: branch maint-beta. now will try push a harmless commit in code.
08:24:05 <chris> (I think the branch creation was silent to everyone)
08:26:18 <chris> @tell gjanssens - another major headache is to decide what happens to reports in 4.x -- keep or ditch the old ones so that new-owner-report / new-aging / balsheet-pnl can take over?
08:26:18 <gncbot> chris: The operation succeeded.
08:28:24 <chris> warlord aha maint-beta was ceremoniously deleted by code
08:36:55 *** Hamaryns has joined #gnucash
08:36:55 *** ChanServ sets mode: +v Hamaryns
08:38:55 *** Gerd has quit IRC
08:39:37 *** Mechtilde has joined #gnucash
08:46:41 *** Yotson has quit IRC
08:47:26 <warlord> chris, right. So like I said, I don't think we can use gnucash/gnucash for this
08:49:11 *** Yotson has joined #gnucash
08:49:51 *** Hirppa has quit IRC
08:54:36 *** Hirppa has joined #gnucash
08:56:30 <fell> chris: there are many possible ways:
08:57:35 <fell> you could build it locally and send the flatpakref to adrien, if nobbody else is interested.
08:59:02 <fell> If more are interested: I do not like a new beta branch in gnucash.
09:06:38 <warlord> Like I said, I can set up a (temporary) flatpak builder off an arbitrary git repo.
09:06:51 <fell> warlord could create builds/flatpak/special/clam
09:11:02 <fell> based on tocreate https://github.com/christopherlam/gnucash/gnucash-on-flatpak
09:11:11 <chris> that's a lot of ceremony to just create some builds
09:12:00 <fell> wheer you can do adjustments to pull a branch from https://github.com/christopherlam/gnucash/gnucash
09:20:11 *** sbluhm has quit IRC
09:29:59 *** Gerd has joined #gnucash
09:39:17 *** chris has quit IRC
09:51:17 *** Hamaryns has quit IRC
09:54:14 *** storyjesse has quit IRC
10:00:09 *** Hamaryns has joined #gnucash
10:00:09 *** ChanServ sets mode: +v Hamaryns
10:32:07 *** chris has joined #gnucash
10:32:07 *** ChanServ sets mode: +v chris
10:33:37 *** Gerd has quit IRC
10:35:18 *** omnireq_ has quit IRC
10:35:54 *** omnireq_ has joined #gnucash
10:39:47 <chris> to jralls re: assistant-loan unable to handle 2,000,000 - I saw that in Windows 3.10 today, but not in Linux
10:40:42 <chris> in windows the loan being 2mil gives a single-unicode char in the pmt formula; loan being 1mil gives a _ in the pmt formula
10:47:45 <chris> so, not sure exactly where the bug will lie
10:51:33 *** Gerd has joined #gnucash
10:53:04 *** Han has joined #gnucash
10:56:48 *** omnireq_ has quit IRC
10:56:58 *** omnireq_ has joined #gnucash
11:07:34 <chris> plus https://bugs.gnucash.org/show_bug.cgi?id=797673#c18 agree that split->posted_date == date should be included in GetBalanceAsOfDate
11:09:38 *** jervin has joined #gnucash
11:09:55 <chris> I believe that my small refactoring had copied the old logic so this boundary bug would've existed for a long time
11:14:06 *** guak has joined #gnucash
11:18:10 <warlord> chris, Oops!
11:18:24 <chris> oops?
11:18:39 <chris> there's lots of oops here
11:19:48 <warlord> Well, copying a fencepost error is an oops.
11:21:25 <chris> I didn't want to fix something that could break other things :)
11:21:47 <chris> (in truth I didn't think it was an error)
11:22:25 <warlord> ok
11:24:02 <chris> if jralls agrees we could/should fix this fencepost error soon
11:25:15 <chris> the 1mil bug is interesting tho
11:26:02 *** Hamaryns has quit IRC
11:39:11 <chris> w.r.t. the custom-flatpak-build issue -- I would be happy with a pull from christopherlam/gnucash:beta but wouldn't really want to use up too much of your space. so if the script can build when the branch is present, or skip if the branch is missing, then it would be perfect.
11:39:24 <chris> now gtg sleep:D
11:45:39 *** Gerd has quit IRC
11:57:38 <jralls> fell, *what* was fixed in AQB 6.2?
11:59:13 <fell> AQOFX_LOG_COMM=1 is a switch again and not the filename
12:00:07 *** jervin has quit IRC
12:03:13 <fell> jralls: commit eb8b83a AqOfxConnect: Improved handling of ENV var AQOFX_LOG_COMM
12:03:23 <fell> This var can now also take the path of the log file to use.
12:03:24 <fell> If you just use "1" as most have before the previous default path
12:03:26 <fell> "/tmp/ofx.log" will be used. Otherwise the given file path will be used.
12:05:43 <fell> was fixed in commit 966c921 OFX: Fixed my stupid bug.
12:07:05 <jralls> @tell gjanssens chris, warlord I like the idea of pulling from chris's github repo. I agree that it's better to not have feature branches in the canonical repository and having branches on the github mirror that aren't mirrors might get messy.
12:07:05 <gncbot> jralls: The operation succeeded.
12:11:44 <fell> My first line was wrong. since 6.1.3 AQOFX_LOG_COMM is for the path and 6.2.0 catches the old behaviour replaces 1 by "/tmp/ofx.log".
12:12:34 *** omnireq_ has quit IRC
12:13:55 <jralls> chris, I just tested on 3.10/MacOS, it took $2,000,000.00 quite happily.
12:16:07 <jralls> chris, as for 797673 we did fix the fencepost error by reverting the change that caused it to matter.
12:17:14 <jralls> fell, Ah, that had *nothing* to do with the change you linked. Good that Martin thinks he fixed it, but has anyone with an OFX account actually tested?
12:17:28 *** omnireq has joined #gnucash
12:25:14 <fell> you could ask our flatpak-user with versions since Apr 15, 2020
12:27:00 <fell> my PR for win is still pending.
12:28:12 <jralls> What user would that be?
12:28:30 *** Mechtilde has quit IRC
12:28:35 <fell> users
12:29:34 <jralls> No, user. I don't think anyone runs nightlies unless explicitly asked to. What user did you ask to test that feature?
12:33:56 <fell> I had not announced it because I thought the windows nightlies would follow soon.
12:36:30 <jralls> chris, I tested the loan assistant in Linux and Windows as well. It happens only on windows and as a preliminary guess I suspect it's because of the commas. 999,999.00 works, 1,000,000.00 doesn't.
12:42:19 <fell> ^ That would also depend on LANG. I could enter 1.000.000,00 (or today with spaces or in CH with apostrophe)
12:43:30 <jralls> fell: OK, I suspect it's the thousands delimiter. Happy?
12:44:10 *** calvinct has joined #gnucash
12:44:43 <fell> I suspected a buggy implementation of the delimiter.
12:59:30 *** bertbob has quit IRC
13:01:58 *** bertbob has joined #gnucash
13:01:59 *** ChanServ sets mode: +v bertbob
13:06:59 *** PowaBanga has quit IRC
13:20:35 *** PowaBanga has joined #gnucash
13:23:40 *** omnireq_ has joined #gnucash
13:24:36 *** Hamaryns has joined #gnucash
13:24:36 *** ChanServ sets mode: +v Hamaryns
13:24:52 *** omnireq has quit IRC
13:41:57 *** omnireq_ has quit IRC
13:42:30 *** omnireq_ has joined #gnucash
13:45:46 *** ChattyMan has joined #gnucash
13:47:06 *** Gerd has joined #gnucash
13:47:24 <ChattyMan> Hi Gerd
13:48:36 <ChattyMan> I'm looking for where I get some technical information about my Gnucash app. Am I in the correct place.
13:50:26 <fell> ChattyMan: the IOS app?
13:50:35 *** Mechtilde has joined #gnucash
13:50:51 <ChattyMan> yes
13:52:23 <fell> I thought, in the wiki the python examples might be a good starting point for how to use the gnucash engine.
13:53:13 <ChattyMan> ok, thanks ..... that's a start
13:53:18 *** ChattyMan has quit IRC
13:55:48 <warlord> jralls, chris, I'm fine setting up a secondary flatpak repo that pulls from chris' github and builds branch beta. I can get that set up today and, if given the 'go' before ~8pm US/ET tonight I can set it up to run tonight.
14:03:55 *** suukim has quit IRC
14:04:42 <jralls> warlord, as long as gjanssens has no objections I think you can have it start running. I don't see any downside so long as the resulting logs and binaries are clearly named.
14:08:27 *** omnireq_ has quit IRC
14:08:38 *** omnireq_ has joined #gnucash
14:12:55 <warlord> I don't know how/where it'll be copied..
14:13:11 <warlord> So I'll set up the directoy structure but wait to fire it off.
14:29:27 *** omnireq_ has quit IRC
14:29:38 *** omnireq_ has joined #gnucash
14:42:05 *** sbluhm has joined #gnucash
14:45:18 *** calvinct has quit IRC
14:50:16 *** angel has joined #gnucash
15:02:57 *** Hamaryns has quit IRC
15:05:41 *** Hamaryns has joined #gnucash
15:05:41 *** ChanServ sets mode: +v Hamaryns
15:11:30 *** Hamaryns has quit IRC
15:27:53 *** angel has quit IRC
16:17:06 *** gjanssens-travelling has joined #gnucash
16:17:40 <gjanssens-travelling> .
16:18:32 *** User has quit IRC
16:20:57 *** omnireq_ has quit IRC
16:21:13 *** omnireq_ has joined #gnucash
16:30:40 * gjanssens-travelling is ok with a separate flatpak build pulling from chris' github repo
16:33:15 <gjanssens-travelling> warlord you can configure where you want te files to be uploaded
16:33:45 <gjanssens-travelling> That's a configuration option somewhere
16:34:43 <gjanssens-travelling> I don't know offhand whether you can pass this on the command line or via the settings file
16:35:38 <warlord> gjanssens-travelling, I'll go take a look before I set it off to run. As it is I need to duplicate the whole checkout tree.
16:38:29 <gjanssens-travelling> In custom.sh you can play with host and host_public
16:39:32 <gjanssens-travelling> There's no way to pass it on the command line
16:40:57 *** gjanssens-travelling is now known as gjanssens
16:41:07 <gjanssens> .
16:41:07 <gncbot> gjanssens: Sent 8 hours and 14 minutes ago: <chris> - another major headache is to decide what happens to reports in 4.x -- keep or ditch the old ones so that new-owner-report / new-aging / balsheet-pnl can take over?
16:41:08 <gncbot> gjanssens: Sent 4 hours and 34 minutes ago: <jralls> chris, warlord I like the idea of pulling from chris's github repo. I agree that it's better to not have feature branches in the canonical repository and having branches on the github mirror that aren't mirrors might get messy.
16:42:27 *** omnireq_ has quit IRC
16:42:38 *** omnireq_ has joined #gnucash
16:52:09 *** mib_gvv289 has joined #gnucash
16:58:25 *** mib_gvv289 has quit IRC
17:05:19 *** sbluhm has quit IRC
17:08:27 *** omnireq_ has quit IRC
17:09:24 *** omnireq_ has joined #gnucash
17:25:49 *** Mechtilde has quit IRC
17:44:01 *** oozer has joined #gnucash
17:58:37 *** Gerd has quit IRC
18:01:46 *** JayC has quit IRC
18:51:27 *** omnireq_ has quit IRC
18:51:38 *** omnireq_ has joined #gnucash
18:52:04 <chris> warlord: thanks - will it be ok then if you can pull from christopherlam/gnucash:beta
18:52:18 <chris> (if it doesn't exist then it skips?)
18:52:27 <chris> (and tries again the next day?)
18:53:26 <chris> jralls the fencepost error is not triggered anymore, however as a matter of principle GetBalanceAsOfDate "should" include splits with posted_date == date desired
18:55:24 *** User has joined #gnucash
19:06:20 *** Aussie_matt has joined #gnucash
19:23:31 *** User has quit IRC
19:58:38 *** tonysoar has joined #gnucash
20:23:57 *** omnireq_ has quit IRC
20:24:08 *** omnireq_ has joined #gnucash
20:58:16 <warlord> chris, I dont know if the script is smart enough to exit if the branch doesn't exist. I didn't actually get it set up today, but I'll work on setting it up tomorrow. I need to get to bed; meeting starts in 7h30m
21:05:08 *** oozer has quit IRC
21:06:52 *** mykeylynx has joined #gnucash
21:07:54 <mykeylynx> for USAA "location" under "select a Bank"
21:08:09 <mykeylynx> what should we type
21:23:36 *** guak has quit IRC
21:28:31 *** hussam has quit IRC
21:33:31 *** jervin has joined #gnucash
21:38:00 *** tonysoar has quit IRC
21:43:57 *** omnireq_ has quit IRC
21:44:08 *** omnireq_ has joined #gnucash
22:05:27 *** omnireq_ has quit IRC
22:07:10 *** omnireq_ has joined #gnucash
22:53:27 *** omnireq_ has quit IRC
22:53:38 *** omnireq_ has joined #gnucash
23:11:58 *** chf has quit IRC
23:16:39 *** chf has joined #gnucash
23:34:57 <fell> mykeylynx>, you saw https://wiki.gnucash.org/wiki/Setting_up_OFXDirectConnect and https://wiki.gnucash.org/wiki/OFX_Direct_Connect_Bank_Settings ?
23:37:46 <mykeylynx> Thank you USAA downloaded I had to add all accounts in the accounts tab. I only had one account.
23:39:41 *** angel has joined #gnucash
23:51:35 *** TownsendHardware has joined #gnucash