2020-06-22 GnuCash IRC logs
00:02:38 *** angel has joined #gnucash
00:05:09 *** David has quit IRC
00:07:21 *** David has joined #gnucash
00:07:43 *** warlord has quit IRC
00:08:01 *** warlord has joined #gnucash
00:24:57 *** fell_laptop is now known as fell
00:42:14 *** angel has quit IRC
00:46:37 *** Mechtilde has joined #gnucash
00:56:37 *** omnireq has joined #gnucash
01:13:03 *** dtux has quit IRC
01:19:48 *** sbluhm has joined #gnucash
01:19:48 *** ChanServ sets mode: +v sbluhm
01:53:42 *** fell_laptop has joined #gnucash
01:53:43 *** ChanServ sets mode: +o fell_laptop
01:55:25 *** fell has quit IRC
02:08:16 *** fell_laptop is now known as fell
02:24:49 *** Mechtilde has quit IRC
02:38:05 *** jervin has quit IRC
02:39:15 *** jervin has joined #gnucash
02:47:35 *** jervin has quit IRC
02:55:51 *** chris1 has quit IRC
03:00:26 *** gjanssens has joined #gnucash
03:00:26 *** ChanServ sets mode: +o gjanssens
03:01:49 *** omnireq_ has joined #gnucash
03:03:01 *** omnireq has quit IRC
03:08:02 <gjanssens> .
03:15:53 *** omnireq_ has quit IRC
03:26:24 <fell> jralls: Sorry, i got summoned yesterday and later was too tired. I am in between wondering, why you tried to backport uk.po-3.905 into 3.10.
03:27:31 <fell> gjanssens: do you really want to wait for 5.0 with the guile version?
03:27:54 <fell> IMHO it is a dependency change.
03:28:36 <fell> And some distributions already replaced guile2 by guile3.
03:31:44 *** Mechtilde has joined #gnucash
03:45:27 <gjanssens> fell: distros that upgrade that aggressively can carry the guile 3 patch until we're ready to make it mainstream
03:45:44 <gjanssens> While reviewing the patch I missed two errors that chris later found.
03:46:25 <gjanssens> To my mind the upgrade was just a simple find and replace (s/_/G_/g), but clearly there are subtleties there.
03:46:59 <gjanssens> Had chris not found these two exceptions, we would have shipped a broken gnucash, not only with guile 3 but with all supported guile versions.
03:47:29 <gjanssens> If we were earlier in the 4.0 release process, I would have added it in as there was still time for user feedback.
03:47:55 <gjanssens> But I'm very reluctant to apply such a patch to the final release.
03:49:47 <gjanssens> Simply because I don't know if we didn't miss any other subtle changes.
03:50:29 <gjanssens> I can see it applied to a later release in the 4.x series, though preferably after some real user testing on those aggressive distros.
03:51:45 <gjanssens> Of course I'm just expressing my view. If the other devs feel confident enough to apply it still, I will accept (and would even be willing to do the work).
05:05:23 *** Aussie_matt has quit IRC
05:05:45 *** User has joined #gnucash
05:12:03 *** storyjesse has joined #gnucash
05:19:17 *** sbluhm has quit IRC
05:21:16 *** sbluhm has joined #gnucash
05:21:16 *** ChanServ sets mode: +v sbluhm
05:22:47 *** jost has quit IRC
05:35:15 <fell> OK, perhaps jralls should mention the patch in the 4.0 release notes.
05:45:48 *** jost has joined #gnucash
06:56:51 *** suukim has joined #gnucash
07:12:49 *** phoenix has joined #gnucash
07:56:34 *** chris1 has joined #gnucash
08:06:15 *** sbluhm has quit IRC
08:37:33 <chris> CDB-Man_ cannot reproduce bug 797796 SPY=$0.00 using your register.
08:40:35 *** KaiForce has joined #gnucash
09:23:13 *** storyjesse has quit IRC
09:25:45 *** Unhammer has joined #gnucash
09:25:51 <chris> CDB-Man_ you know how gnucash handles currencies and commodities natively; there was half-finished work to create a book-currency to which all commodities were exchanged. unfortunately that developer has been inactive for 6 years and counting.
09:29:46 *** Mechtilde has quit IRC
09:45:38 <warlord> A long long time ago every transaction (or maybe split) had a "commodity" and "currency" associated with it. It's been so long I honestly don't recall how it was used, but I remember arguing at the time that removing it would make balancing accounts harder. Here we are, 20 years later, still dealing with imbalance issues.
09:46:03 <warlord> (or maybe it was an account that had a commodity and currency?)
09:46:18 <warlord> I suppose I can go look at the code history from 20 years ago to see..
09:47:03 <chris> warlord: is that before or after single-entry was made 'illegal'?
09:49:32 <warlord> chris, define "made illegal"
09:51:24 <chris> when transactions were corrected to always balance?
09:52:46 <warlord> Ah. That was Linas' doing. The removal of the commodity/currency duality was earlier.
09:54:01 <CDB-Man_> [08:37:33] <+9d54b3chris> CDB-Man_ cannot reproduce bug 797796 SPY=$0.00 using your register.
09:54:08 <CDB-Man_> ... well then
09:54:32 <CDB-Man_> You're running average cost, right?
09:58:22 <chris> sure.
09:59:25 <CDB-Man_> Well this is very odd, I wonder what else contributes to my zero dollar spy
09:59:28 <chris> must be something about the 9.95 SPY txn.
10:00:04 <CDB-Man_> Did you recreate a register using those exact 3 entries in my
10:00:09 <CDB-Man_> Screenshot?
10:00:28 <chris> from your imgur/Bdp74ZG
10:01:35 <CDB-Man_> Maybe we need to print to trace file every calculation that's happening for my register then
10:02:02 <CDB-Man_> You're running account summary right?
10:02:48 <CDB-Man_> Also: return of capital is Dr cash, CR SPY, zero units, zero price
10:03:09 <chris> ^yes acc-summary
10:03:16 <chris> what's return-of-capital in layman terms
10:03:43 <CDB-Man_> The company is returning a portion of your original investment to you
10:03:53 <CDB-Man_> Rather than paying a dividend from retained earnings
10:04:12 <CDB-Man_> It's therefore a reduction of cost basis
10:04:30 <chris> I have never seen this case documented in gnucash documentation?
10:04:55 <CDB-Man_> https://www.investopedia.com/terms/r/returnofcapital.asp
10:04:55 <chris> ah it is in tutorial
10:05:54 *** phoenix has quit IRC
10:06:07 <CDB-Man_> Anyways, on the $0 spy issue, if you replicated the entries verbatim, along with the $9.95 split, idk what else to say
10:06:36 <CDB-Man_> Since I have $9.95 on other trades for other stocks but they don't suffer this problem
10:07:09 <CDB-Man_> If you have some debug code I can add to the report to print something more verbose, I can do that
10:07:26 *** phoenix has joined #gnucash
10:08:31 <CDB-Man_> Anyways, time for some breakfast
10:10:49 <CDB-Man_> [09:45:38] <warlord> A long long time ago every transaction (or maybe split) had a "commodity" and "currency" associated with it. It's been so long I honestly don't recall how it was used, but I remember arguing at the time that removing it would make balancing accounts harder. Here we are, 20 years later, still dealing with imbalance issues.
10:11:08 <CDB-Man_> As in each account effectively kept track of 2 totals?
10:11:51 <CDB-Man_> Rather than right now where stock accounts "only" track share units?
10:12:47 <CDB-Man_> [09:47:03] <+9d54b3chris> warlord: is that before or after single-entry was made 'illegal'?
10:12:54 *** phoenix has quit IRC
10:12:55 <warlord> CDB-Man_, yes
10:13:06 <CDB-Man_> Hmm, at one point entries were not forced to balance? Interesting
10:13:40 <warlord> CDB-Man_, it would allow you to "commit" an imbalanced transaction (and you'd get transactions with that little box in the corner)
10:14:02 <CDB-Man_> warlord: yes, that double track would be quite useful. Though without the ability to lookup historical exchange rate, I can see why dual commodities were impractical
10:14:45 <CDB-Man_> I don't think I've ever seen this "little box"
10:15:04 <CDB-Man_> Impractical +back then
10:17:05 <CDB-Man_> Chris, how many hours until you head to bed? If very shortly, then I'll save that transaction report trial for this evening
10:17:13 <chris> <1hr :)
10:17:23 <CDB-Man_> Okay, this evening it is then
10:17:39 <CDB-Man_> Hopefully I'll have run it by the time you wake up
10:18:24 <CDB-Man_> I will also include the SPY TRADING account in the report
10:18:27 <chris> here's what it should look like: https://pastebin.com/raw/E59YjPC1
10:18:40 <CDB-Man_> And also all accounts in the future that holds SPY
10:19:09 <CDB-Man_> Okay, the reference helps
10:22:35 <CDB-Man_> chris: to ensure I don't muck it up, can you also attach the patched version of trep?
10:22:53 <CDB-Man_> Like how the other day my trace file produced that error
10:23:56 <chris> https://pastebin.com/raw/FVjcedDb
10:30:11 <CDB-Man_> Thanks
11:11:59 *** chris1 has quit IRC
11:17:40 *** chris1 has joined #gnucash
11:18:06 *** omnireq has joined #gnucash
11:24:55 *** omnireq has quit IRC
11:44:58 *** guak has joined #gnucash
11:46:15 *** sbluhm has joined #gnucash
11:46:15 *** ChanServ sets mode: +v sbluhm
11:57:03 <jralls> fell, no, I'm not going to mention that OpenSuSE broke GnuCash by dropping Guile 2.2. There was no need for them to do that, the Guile team goes to great lengths to ensure that multiple major versions are side-by-side installable.
11:57:46 <jralls> That said, I don't see any reason not to merge Guile 3 support into 4.x when it's working.
12:00:21 *** Mechtilde has joined #gnucash
12:00:29 *** sbluhm has quit IRC
12:02:21 *** jervin has joined #gnucash
12:10:11 *** ArtGravity has joined #gnucash
12:10:11 *** ChanServ sets mode: +v ArtGravity
12:19:12 *** CDB-Work has joined #gnucash
12:19:12 *** ChanServ sets mode: +v CDB-Work
12:22:45 *** dtux has joined #gnucash
12:40:43 <jralls> CDB-Man, the little box warlord refers to is the gray box-with-an-x that appears in every amount and value cell in the split view of an unbalanced and uncommitted transaction.
12:41:01 *** sbluhm has joined #gnucash
12:41:01 *** ChanServ sets mode: +v sbluhm
12:41:34 <jralls> CDB-Man_ & CDB-Work ^^^^^^
12:46:11 <jralls> gjanssens, I'm not very happy about BobIT's removal of per-account register formats. I think it's unnecessary and will annoy a lot of users, many of whom will decide not to upgrade.
13:01:02 <CDB-Work> ah, I'm pretty good with never having unbalanced transactions, so that's probably why I haven't noticed it :)
13:02:12 <CDB-Work> jralls: i saw briefly your response on the ticket; before I start drafting an answer in detail, did you consider the fact that the cumulative column rolls into the per unit ACB column? and the per unit ACB column, while it uses cumulative as tne numerator, it uses the cumulative # units as the denominator
13:02:22 <CDB-Work> therefore the per unit amount does not change after a sale
13:06:51 *** User has quit IRC
13:11:13 <CDB-Work> > removal of per-account register formats <-- did he merge all registers into a single general journal??
13:14:00 *** sbluhm has quit IRC
13:17:42 *** gotnone has joined #gnucash
13:19:23 <gotnone> Hi, can anyone tell me how many days back the generic import transaction matcher looks at when importing OFX data? Is this configurable in settings or a config file somewhere?
13:24:17 *** sbluhm has joined #gnucash
13:24:17 *** ChanServ sets mode: +v sbluhm
13:39:51 <fell> jralls, FYI: https://bugzilla.opensuse.org/show_bug.cgi?id=1173232
14:02:53 <gjanssens> jralls: wrt the removal of per-account register layouts, BobIT did this on my suggestion. I think there are good use cases to fall back to a default. But honestly I don't have enough energy left for gnucash to take on such discussions with users :(
14:03:13 <gjanssens> So if you prefer to revert feel free to do so.
14:04:36 <jralls> gjanssens, I'd rather fix than revert. I'll point out that falling back to a default is not the same as deleting the saved layout and using only the default. The current implementation does the latter.
14:06:15 <gjanssens> Sure. I did write at more length when David H. first brought this up about what I think are the challenges of a good user experience with 3 levels of layout management (calculated default, user set default, per register settings).
14:07:38 <gjanssens> It requires users to understand all three levels and be able to switch between them whenever appropriate. To me that's just too much for most users.
14:08:13 <gjanssens> Anyway there are two use cases at the extreme
14:09:01 <gjanssens> 1. Users that currently have to tweak each and every register because their chosen font/language doesn't nicely calculate good defaults (dates get truncated, number columns are too small)
14:09:37 <gjanssens> 2. Users that wish to tweak each register to have optimal widths for the description and transfer column.
14:09:43 <gjanssens> Both are fair use cases.
14:10:38 <gjanssens> The users of use case 1 were the ones complaining during the 3.x cycle. The users of case 2 will be complaining for the 4.x release unless we find a solution.
14:12:34 <jralls> Right. I don't see a big problem with it. When opening a register check if the account has a saved layout, as in <=3.11. If not, check if its type group does and if not that then use the calculated defaults.
14:13:09 <gjanssens> That's the developers view and yes, that's easy.
14:13:23 <jralls> How is that hard for users?
14:14:15 *** suukim has quit IRC
14:16:06 <CDB-Work> by layouts, are you talking about things such as column width?
14:16:33 <CDB-Work> I for one have the split account name column set at 5x the normal width, so that I can see the FULL account name
14:16:55 <jralls> Yes, that's exactly what we're talking about.
14:17:19 <jralls> And that's exactly what 3.90x breaks.
14:18:21 <gjanssens> Go reread my mail messages. it's way to long to repeat in irc - they are part of this list thread: https://lists.gnucash.org/pipermail/gnucash-devel/2020-June/045175.html
14:19:14 <gjanssens> jralls: it doesn't break in 3.90x. It works differently. You can still set the split account name column to 5x the normal width. You just have to tell gnucash you want that to be the default.
14:19:30 <gjanssens> But whatever.
14:19:43 <jralls> The very first sentence is false. No effort is made to reuse what was there before.
14:20:40 <CDB-Work> Hmm, if there was the ability to set default column width in current gnucash, I didn't know about it and wish that I did. I would definitely set default to 5x for my own use
14:20:41 <jralls> What was there before is discarded, *even if you had the tab open when you closed GnuCash 3.10*.
14:21:50 <jralls> CDB-Work, that's the new feature: You can set the default widths per register type (currency, stock, etc.).
14:22:14 <CDB-Work> Ah, okay, so I was not particularly blind then!
14:22:17 <gjanssens> Then *that's* a bug. If the tab was still open, the available column widths should have been retained.
14:23:04 <gjanssens> I suspect this is because it's now stored differently and BobIT didn't add code to migrate.
14:23:47 <jralls> That's what it looks like. The old way is [Register b2fb098c6822d77fcf2ffc09e5ce73ff] and the new way is AccountGuid=b2fb098c6822d77fcf2ffc09e5ce73ff.
14:24:10 <gjanssens> As for registers that weren't open when last saved, there's no way to map individual saved layouts to a common default. So yes those would inevitably get lost.
14:24:15 *** frakturfreak has joined #gnucash
14:24:15 *** ChanServ sets mode: +v frakturfreak
14:25:20 <jralls> And that's the part that I don't understand and that your email to David doesn't explain: Those register layouts are stored in the .gcm, why discard them?
14:26:23 <jralls> While thinking about this plenty of corner cases came up that would either be confusing or cumbersome for the user
14:26:38 <jralls> What corner cases?
14:27:17 <jralls> Sorry, my IRC client didn't handle that right. "While..." is a quote from your email.
14:29:07 <gjanssens> https://lists.gnucash.org/pipermail/gnucash-devel/2020-June/045182.html
14:30:51 *** Hamaryns has joined #gnucash
14:30:51 *** ChanServ sets mode: +v Hamaryns
14:32:55 <gjanssens> jralls: bug 797815 complains about a poorly configured gnucash-launcher.cmd file on Windows.
14:33:30 <gjanssens> Do we still have a use case for that file ? I thought we obsoleted it ages ago as gnucash itself knows where to find its dependencies.
14:33:53 <gjanssens> I'm tempted to remove that one unless you have a valid use case for it ?
14:33:54 <jralls> Yeah, I saw that. That file is left over from the old shell-script build system with its goofy paths. It's no longer needed.
14:34:13 <gjanssens> Yay, I'll drop it.
14:34:33 <jralls> Plus it should never have been in the installer, it's for running gnucash from the build dir.
14:36:48 <jralls> On the subject of goofy, did you see cstim's PR about "Change" vs. "Edit"? He claims that we have a policy to always use "Edit" for UI controls, never any synonyms. I don't remember any such policy. Do you?
14:37:44 <gjanssens> Nope
14:38:20 <jralls> warlord, you've been around the longest. Do remember anything like that?
14:39:03 <warlord> ... reading back..
14:39:47 <warlord> Change vs Edit? I honestly don't recall any such policy, but I wasn't paying attention to HIG issues.
14:42:36 <jralls> OK. I'll chalk that up to trying to make an argument on thin ice. I'll leave it to simmer for a bit and see what happens.
14:46:09 <jralls> gjanssens, On the UX with column widths, the confusion would stem from the user not understanding that a saved layout is all of the column widths, and since there are users who complain that scrolling their notebook tabs has gotten too cumbersome even the current scheme will lead there.
14:48:16 <jralls> We have two menu items, Set this layout as the default and reset the default to calculated. A third, reset this register to default, plus actually documenting how all of this works in Help, would I think resolve most of the potential for confusion.
14:49:35 *** dtux has quit IRC
14:59:05 *** Mechtilde has quit IRC
14:59:30 <gjanssens> jralls: ok, you can give that a shot.
15:11:45 <gjanssens> jralls: https://bugs.gnucash.org/show_bug.cgi?id=797803 about flatpaks that try to recompile guile sources. I think that's either a guile or a flatpak bug (and there's a bug against the gnucash on flathub github repo).
15:12:07 <gjanssens> Do you want to track this in our bugzilla as well, or should I just close it as NOTGNUCASH ?
15:12:40 <gjanssens> The link to the github issue is in the comments.
15:13:03 <jralls> You said something earlier about flatpak performing unnatural acts on timestamps. Do you know anything more about that?
15:14:11 <jralls> Is it really flatpak or is it something about the flathub build process?
15:19:38 <jralls> OIC, flatpak build gives build products a 0 timestamp. That's bad.
15:20:17 <jralls> So it's not GnuCash but it isn't really flathub either.
15:21:50 *** Hamaryns has quit IRC
15:25:23 <jralls> And according to mclassen in https://github.com/flatpak/flatpak/issues/926 it's on purpose. Can't we work around it by setting GUILE_AUTO_COMPILE=0 in environment?
15:27:43 *** jervin has quit IRC
15:27:53 *** David has quit IRC
15:28:07 *** David has joined #gnucash
15:28:29 <jralls> Ah, I see in https://github.com/flathub/org.gnucash.GnuCash/issues/24#issuecomment-522767767 that that's a poor option. Hmm.
15:39:09 *** jervin has joined #gnucash
15:48:58 <jralls> gjanssens, you can close it as not gnucash if you want, but since the 0 timestamps are on purpose nobody's going to fix it upstream "won't fix" is probably a more accurate answer. Is there any way we can put a workaround in? Maybe change the timestamp on *.scm to 0 as well?
15:50:24 *** mohave has joined #gnucash
15:50:59 <gjanssens> jralls: these timestamps are set by the build process on *.scm files as well.
15:51:37 <gjanssens> For completeness it appears to only affect Ubuntu so far. On Fedora this works fine.
15:52:41 <mohave> I would definitely vote to maintain the ability to customize column widths on an account by account basis.
15:55:18 <jralls> gjanssens, interesting about Ubuntu. So if the .scm and .go files have the same timestamp it still sets off auto-compile. Hmm, we can patch guile for the build, right? Changing that is probably a one-liner.
15:55:54 *** o01eg has quit IRC
15:56:14 <gjanssens> What would you want to patch in guile ?
15:57:24 <jralls> The auto-compile condition so that if the .go and .scm timestamps are the same the .go is OK to use.
15:58:14 <jralls> Or maybe just turn it off so that the .go is *always* OK to use.
16:00:52 *** User has joined #gnucash
16:05:17 <gjanssens> Oh right. We could do that indeed
16:09:17 *** sbluhm has quit IRC
16:13:20 *** jervin has quit IRC
16:13:27 *** jervin has joined #gnucash
16:14:32 *** mohave has quit IRC
16:18:56 <jralls> The function in question, compiled_is_fresh in libguile/load.c, returns true if the times are equal, but it compares mtime.tv_nsec. I guess that the simplest thing would be to change the last line of that function to return 1.
16:31:59 *** User has quit IRC
16:37:14 *** jralls_afk has joined #gnucash
16:37:14 *** ChanServ sets mode: +o jralls_afk
16:38:26 *** jralls has quit IRC
16:40:15 *** David has quit IRC
16:40:40 *** David has joined #gnucash
16:50:46 *** gjanssens has quit IRC
16:56:34 *** phoenix has joined #gnucash
17:00:10 *** frakturfreak has quit IRC
17:12:21 *** jervin has joined #gnucash
17:13:52 *** jervin has quit IRC
17:37:05 *** David1 has joined #gnucash
17:37:50 *** David has quit IRC
17:37:50 *** David1 is now known as David
17:45:59 *** David1 has joined #gnucash
17:46:50 *** David has quit IRC
17:46:50 *** David1 is now known as David
18:03:02 *** ArtGravity has quit IRC
18:13:47 *** ArtGravity has joined #gnucash
18:13:47 *** ChanServ sets mode: +v ArtGravity
18:13:47 *** ArtGravity has quit IRC
18:40:45 *** TownsendHardware has quit IRC
18:49:24 *** Aussie_matt has joined #gnucash
18:55:39 *** fell_laptop has joined #gnucash
18:55:39 *** ChanServ sets mode: +o fell_laptop
18:55:39 *** fell has quit IRC
19:37:00 *** shaggy1 has joined #gnucash
19:37:22 *** shaggy has quit IRC
19:38:53 *** jervin has joined #gnucash
19:45:14 <linas> my name in lights!
20:12:34 <chris> all hail linas
20:31:30 *** omnireq has joined #gnucash
20:39:05 *** jervin has quit IRC
20:39:18 *** jervin has joined #gnucash
20:44:28 *** phoenix has quit IRC
20:45:31 <warlord> Hi linas.
20:53:07 <CDB-Work> chris: per jralls, my cost basis calculator is already how gnucash works, except for the fx part. jralls was also saying, referring to comments you made earlier, that the fx part will likely need to be done lower level than scheme, ie done in c++. thoughts?
20:53:29 <CDB-Work> except for the fx part + the commissions part**
20:59:16 *** guak has quit IRC
21:11:23 *** angel has joined #gnucash
21:23:05 *** codesmythe has quit IRC
21:23:41 *** warlord has quit IRC
21:23:48 *** warlord has joined #gnucash
21:23:58 *** jost has quit IRC
21:37:40 *** jost has joined #gnucash
21:47:47 *** storyjesse has joined #gnucash
22:04:24 *** omnireq has quit IRC
22:06:39 *** omnireq has joined #gnucash
22:06:39 *** ChanServ sets mode: +v omnireq
22:08:02 *** omnireq has quit IRC
22:12:28 *** gotnone has quit IRC
22:41:25 *** angel has quit IRC
22:51:11 *** angel has joined #gnucash
22:56:31 *** fell_laptop is now known as fell
23:22:31 *** angel has quit IRC
23:25:38 *** angel has joined #gnucash
23:27:03 *** angel has quit IRC
23:28:19 *** angel has joined #gnucash
23:30:33 *** David1 has joined #gnucash
23:31:05 *** David has quit IRC
23:31:05 *** David1 is now known as David
23:41:51 *** angel has quit IRC
23:43:13 *** storyjesse has quit IRC