2018-03-17 GnuCash IRC logs
00:10:11 *** pilotauto has quit IRC
00:48:34 *** marusich has quit IRC
00:48:54 *** marusich has joined #gnucash
00:52:44 *** marusich has quit IRC
00:52:52 *** marusich has joined #gnucash
01:57:07 *** Aussie_matt has joined #gnucash
02:13:09 <fell> @tell Mechtilde in guide/currency you link en.wikipedia.org/wiki/ISO_4217. Are there reasons not to use de.wikipedia?
02:13:09 <gncbot> fell: The operation succeeded.
02:16:40 *** bertbob has quit IRC
02:18:22 *** bertbob has joined #gnucash
02:24:27 *** Cuare has joined #gnucash
02:41:02 *** storyjesse has joined #gnucash
02:45:59 *** fell has quit IRC
02:47:00 *** fell has joined #gnucash
02:48:21 *** marusich has quit IRC
03:01:40 *** gncbot sets mode: +o fell
03:18:44 *** Mechtilde has joined #gnucash
03:31:10 <Mechtilde> fell, Guten Morgen
03:31:10 <gncbot> Mechtilde: Sent 1 hour and 18 minutes ago: <fell> in guide/currency you link en.wikipedia.org/wiki/ISO_4217. Are there reasons not to use de.wikipedia?
03:31:40 <fell> Einen wunderschönen
03:31:50 <Mechtilde> fell, danke für Deine Mail auf gnucash-de
03:32:12 <fell> Oh, das hört man gern.
03:32:35 <fell> Was hat dir denn gefallen?
03:33:33 <Mechtilde> wie du beschreibst, warum die Sachen geradde so sind, wie sie sind
03:34:11 <Mechtilde> Der erste poster hatte mich auch in Chemnitz auf das "schlechte" Programm GnuCash angesprochen
03:34:31 <Mechtilde> ihn hatte ich nochmal auf die Liste verwiesen
03:34:46 <fell> Tja, ist auch immer so ein Balanceakt, nicht zuviel Frust rauszulassen.
03:35:05 <fell> um niemanden zu verschrecken
03:37:03 <Mechtilde> und das mit dem Wikieintrag: Da habe ich wohl nur den Link übernommen
03:37:52 <fell> dann änder ich das zusammen mit dem Katital, das ich noch gefunden habe.
03:38:10 <Mechtilde> danke
03:39:24 <fell> Wäre ja auch möglich gewesen, daß du die Seite qualitativ zu schlecht fandest. In so einem Fall sollten wir einen Kommentar einfügen, warum.
03:41:04 <Mechtilde> es ist her, dass ich das zunächst ohne den link zu beachten übersetzt habe, und dann den Link nihct mehr betrachtet habe
03:41:52 <fell> Alle anderen in dem Kapitel hast du angepaßt sagt mir mein neuer Freund grep
03:44:07 <fell> Ach und einmal hast du im Bild USD und im Text EUR als Fremdwährung. Das könnte verwirren ;-)
03:47:12 <Mechtilde> wo genau?
03:47:39 <Mechtilde> das heißt im Text ist es falsch
03:47:47 <Mechtilde> kannst Du das mit ändern?
03:48:34 <fell> Ziemlich zu anfanf der Wechselkurs geschichte.
03:48:48 <fell> Klar
03:53:12 *** storyjesse has quit IRC
03:53:24 <fell> Ich glaub, ich werde noch ein bischen in https://wiki.gnucash.org/wiki/Documentation_Update_Instructions schreiben zu den verschiedenen docbook tags, was sie bedeuten und wo welche verwendet werden sollten.
03:55:27 <Mechtilde> das sollte ich dann wohl auch nochmal lesen, bevor ich daran weiterarbeite
03:55:31 <Mechtilde> ;-)
03:56:24 <fell> Man muß sie sich auch in den verschiedenen Endprodukten ansehen:
03:57:05 <fell> <varlist> sieht in yelp bescheiden aus, aber in HTML gut,
03:59:03 <fell> <appendixinfo><releaseinfo> wird in HTML ausgedruckt und in yelp unterdrückt.
04:08:39 <fell> Was war eigentlich in Chemnitz für eine Veranstaltung, Mechtilde?
04:08:56 <Mechtilde> CLT Chemitzer Linux Tage
04:27:10 <fell> Mechtilde: Eine Änderung hab ich noch für irgendwann: statt "CURRENCY" heißt es jetzt "Währung", aber dann müßte man auch die Bilder aktualisieren oder beim ersten Vorkommen eine Anmerkung einfügen.
04:28:38 <Mechtilde> ich habe hier noch andere Projekte, die ich abarbeiten sollte
04:28:52 <Mechtilde> ändere das, was Du für notwendig hälst
04:29:44 <fell> Keine Panik, war mir nur vorhin aufgefallen und weil ich doch so vergesslich bin ,,,
04:31:11 <jotrago> I left a comment yesterday regarding Close Book not zeroing out TX entered on the Close Date. Chris commented. After digging in bugzilla I found Bug 789674 (2017-10-31 ) which appears to cover the issue. I have added my comments confirming the symptoms and adding sample files.
04:35:38 <fell> Good Morning, Jotrago!
04:37:40 <fell> Thanks, Goodto have an example file. I /believe/ it is one of the typicel problems mixing dates and timestamps.
04:41:27 <fell> Jotrago, are the files compressed or plain xml?
04:50:37 *** storyjesse has joined #gnucash
05:09:20 *** PaulW2U has joined #gnucash
05:14:29 *** O01eg has quit IRC
05:21:10 *** harshitaneja has joined #gnucash
05:21:14 *** benoitg has joined #gnucash
05:24:35 *** harshitaneja has quit IRC
05:25:18 *** harshitaneja has joined #gnucash
05:47:47 *** benoitg has quit IRC
06:10:23 *** jonas has joined #gnucash
06:31:43 *** storyjesse has quit IRC
06:42:57 *** harshitaneja has quit IRC
06:45:53 *** benoitg has joined #gnucash
06:49:24 *** PaulW2U has quit IRC
07:00:27 *** ncv has joined #gnucash
07:05:27 *** benoitg has quit IRC
07:20:05 *** benoitg has joined #gnucash
07:30:04 *** harshitaneja has joined #gnucash
07:33:55 *** harshitaneja has quit IRC
07:48:56 *** oozer has joined #gnucash
08:12:51 *** robin has quit IRC
08:28:24 *** Aussie_matt has quit IRC
08:48:11 *** fabior has joined #gnucash
09:07:18 *** oozer has quit IRC
09:21:57 *** fabior has quit IRC
09:29:17 *** benoitg has quit IRC
09:34:53 *** Jimraehl1 has joined #gnucash
09:37:40 *** Jimraehl1 has left #gnucash
09:44:10 <ajm> jralls: (Regarding bug #793768 - gnucash file corruption)... I had to revert to a stable version of GC in the meantime but set up a virtual machine with which I was going to build from git and investigate further without endangering my main machine. Sadly I ran out of time shortly after so haven't followed through. This bug you've found though would only affect the troubleshooting approach you gave me in comment 5, rather than be the
09:44:10 <ajm> cause of my original problem, did I understand that right?
09:45:17 *** harshitaneja has joined #gnucash
10:08:19 *** chris has quit IRC
10:18:19 *** chris has joined #gnucash
10:57:47 *** harshitaneja has quit IRC
11:24:18 *** Robert1 has joined #gnucash
11:26:28 *** Robert has quit IRC
11:33:51 *** Derperperd has quit IRC
11:34:32 *** Derperperd has joined #gnucash
11:37:05 *** oozer has joined #gnucash
11:42:25 *** valdas has joined #gnucash
12:04:07 *** valdas has quit IRC
12:04:29 *** valdas has joined #gnucash
12:07:37 *** valdas has quit IRC
12:19:13 *** Cuare has quit IRC
12:32:41 <Mechtilde> I read the release schedule. Can i hope of a fix of https://bugzilla.gnome.org/show_bug.cgi?id=792452
12:54:38 <fell> Mechtilde, can you add the aqbanking logs (see my mail last night)
12:55:33 <Mechtilde> last night?
12:55:37 <Mechtilde> to me?
12:56:05 <fell> on gnucash-de
12:56:51 <fell> or wait
12:57:05 <Mechtilde> I didn't see it
12:57:23 <fell> I wrote something, but to whom?
12:57:35 <Mechtilde> I sent the log files to Geert
12:57:45 <Mechtilde> after FOSDEM
12:59:37 <fell> AFAIK Geert is out of town or otherwise busy almost the rest of the month
13:00:10 <Mechtilde> so I send the mail to you
13:00:27 <fell> OK
13:03:19 <Mechtilde> done
13:03:21 <fell> yes, Instructions were almost at the end of the gnucash-de mail
13:05:17 <Mechtilde> I don't understand... :(
13:05:35 *** fabior has joined #gnucash
13:05:57 <fell> I still didn't revive the laptop with my private key. :-(
13:06:09 <Mechtilde> ok
13:06:23 <Mechtilde> I will resend
13:06:58 <Mechtilde> done
13:12:03 <fell> An der Konfiguration ist mir jetzt nichts aufgefallen. Aber eigentlich sprach ich von den Protokollen ;-)
13:13:19 <fell> Ich muß jetzt erstmal zu Abend essen ....
13:13:27 <Mechtilde> das sind welche Dateien
13:13:45 <Mechtilde> dann bis später
13:50:43 *** fabior has quit IRC
13:52:45 <fell> '/home/frank/.aqbanking/backends/aqhbci/data/banks/de/*/logs' durchstöbern
13:53:43 <fell> Natorlich mechtilde statt frank
13:53:49 <Mechtilde> klar
13:54:03 <Mechtilde> ich schau noch drüber wegen dAten und so
13:54:52 <Mechtilde> dass muss aber ein log sein, bei dem ich das ausprobiert ahbe mit der Überweisung
13:54:54 <Mechtilde> ?
13:55:04 <fell> ja
13:57:03 <fell> kannst ja nochmal nach https://wiki.gnucash.org/wiki/De/HBCI#Fehlersuche die Einstellungen überprüfen
13:57:10 <Mechtilde> nach dem Update auf die 2.7.6 geht es wohl gar nicht mehr
13:57:35 <Mechtilde> Fehler
13:57:42 <fell> Oh, eine regression?
13:57:57 <Mechtilde> my fault
14:00:14 <Mechtilde> It is the same problem in version 2.7.6 as described in the bugreport
14:00:22 <Mechtilde> there is no log file about this
14:02:13 <fell> with "export export AQHBCI_LOGLEVEL=debug .."?
14:02:31 <Mechtilde> where do I find that
14:02:43 <fell> https://wiki.gnucash.org/wiki/De/HBCI#Fehlersuche
14:03:36 <Mechtilde> und wie findet dann Gnucash den anderen Debug Level?
14:04:18 <fell> danach in derselben shell starten
14:05:53 <fell> gnucash --nofile --debug --extra
14:06:31 <fell> hab ich in meiner test.sh stehen
14:07:32 <fell> --nofile willst du wahrscheinlich nicht
14:07:43 <Mechtilde> hab ich mir auch gedacht
14:09:01 <Mechtilde> dass die Eingabe nicht möglich ist, scheint nicht geloggt zu werden
14:10:57 <Mechtilde> I get the message I posted in the bugreport for every charakte I put into
14:11:31 <fell> Nice
14:11:55 <Mechtilde> I demonstrated it to Geert
14:14:19 <fell> something interesting in /tmp/gnucash,*
14:14:27 <fell> ?
14:16:13 *** bob_ has joined #gnucash
14:18:26 <Mechtilde> I will look through it
14:19:24 <fell> Version of aqhbci-tool?
14:20:46 <Mechtilde> aqbanking v5.7.8
14:22:27 <Mechtilde> in /tmp/gnucash.trace wird nur der Start der Funktion aufgezeichnet
14:22:33 <Mechtilde> nicht die Eingabe der daten
14:25:41 <fell> From what I see in src, It tries to get the local (stored) BIC or IBAN of your bank
14:26:56 <fell> what you sent me has a IBAN
14:27:13 <fell> So it should not fail
14:27:50 <fell> I suspect, the path is not correctly set.
14:28:43 <fell> does the tracefile have some initialization of aqb*
14:30:47 <Mechtilde> I send it per mail
14:32:40 <fell> Ha, negative logic!
14:33:18 <fell> if (!localBIC || !localIBAN|| (strlen(localBIC) == 0) || (strlen(localIBAN) == 0))
14:34:05 <fell> In your conf is a IBAN, but no BIC
14:34:31 <Mechtilde> yes
14:34:58 <Mechtilde> because it isn't necessary
14:35:19 <Mechtilde> any more
14:35:38 <Mechtilde> or is it necessary for GnuCash
14:36:01 <fell> I thouhgt so, too
14:37:11 *** valdas has joined #gnucash
14:38:02 <fell> No IBAN or no BIC means IBAN and BIC are both required
14:39:28 <Mechtilde> I will try it
14:40:17 <fell> also the comment says it: // Verify that we have a local IBAN and BI
14:40:25 <fell> C
14:40:57 <Mechtilde> yes then it works
14:41:22 <Mechtilde> So someone changed it with version 2.6.18 I guess
14:42:10 <fell> let me see ...
14:45:11 <Mechtilde> I saw it first in 2.6.18
14:47:19 <fell> Last change 5 years ago
14:49:31 <Mechtilde> 2 years
14:49:44 <Mechtilde> sorry
14:53:33 <Mechtilde> I got version 2.6.18 mid of November
14:54:03 <Mechtilde> 2017
14:54:34 <fell> https://github.com/Gnucash/gnucash/blame/maint/src/import-export/aqb/dialog-ab-trans.c
14:54:50 <fell> line 598
14:56:11 <Mechtilde> I read it but don't understand why I fund this problem in November 2017
14:56:55 <fell> When did you remove the BIC from your conf?
14:57:24 *** mdf has joined #gnucash
14:58:24 <Mechtilde> try to remember wenn I changed from pin/Tan to chiptan
14:59:49 <jralls> ajm: That's correct.
15:02:14 <Mechtilde> Can we mark the field to have to filled
15:03:25 <fell> the construction site is line 708ff
15:03:49 <fell> Hab ich nicht verstanden
15:05:09 <jralls> Mechtilde: In 792452 the error message you quote has the %s substitution characters. Is that because you copied it from the msgid in de.po so that you'd have the English version, or are those characters actually showing up in the dialog?
15:05:40 <Mechtilde> it was copied from the po file
15:05:57 <Mechtilde> for the English text
15:06:09 <jralls> Oh, good. Thanks.
15:06:56 *** frakturfreak has joined #gnucash
15:07:50 <Mechtilde> it was filld with "bank Number and Account Numbr
15:07:56 <Mechtilde> filled
15:08:17 <fell> (localBankCode, localAccountCode)
15:08:29 <Mechtilde> yes
15:10:06 <jralls> Good. I was concerned that there were two problems.
15:10:34 <jralls> So is BIC completely dead, or should we check for it if IBAN is empty?
15:12:06 <Mechtilde> for transactions in Germany you didn't need BIC if you have the IBAN
15:12:20 <fell> It worked, when BIC was in the configuration, too
15:12:29 <fell> right, Mechthilde?
15:12:33 <Mechtilde> yes
15:13:01 <Mechtilde> so I think it is easier to have a must field for the local BIC
15:13:12 <fell> how did you call aqhbcitool, to set it
15:13:13 <Mechtilde> that should be a problem for the user
15:13:28 <jralls> Sure, I can understand that having both isn't a problem, BIC is ignored if IBAN is available.
15:13:36 <Mechtilde> if (s)he knows to inut it
15:14:26 <Mechtilde> to set what fell ?
15:14:32 <fell> BIC
15:14:35 <Mechtilde> s/inut/input
15:14:57 <Mechtilde> I use the dialog in gnucash
15:15:59 <Mechtilde> under Tools - Onlinebanking - Aqbanking-Assistent or similar
15:16:44 <Mechtilde> there I can Edit the account
15:17:19 <fell> all required params?
15:17:28 <jralls> So can we change the condition on the test to (shortened psuedocode) if ( !iban && !bic)... ?
15:17:58 <fell> or have we still to run aqhbci-tool to set which one?
15:19:18 <Mechtilde> I in put all parameters in the assistent
15:21:06 <fell> So we can split the condition and say used tools ... to set Iban , ... set BIC
15:21:27 <jralls> Right. A new string, sigh.
15:22:11 <fell> uncovering a hidden feature
15:23:18 <fell> But this part is only for de_DE and I hope, Mechtilde will translate in timee. ;-)
15:23:25 <jralls> Not hidden, it's pretty clearly in the Edit Account dialog: Under Account Info, IBAN and under Bank Info, SWIFT BIC.
15:24:25 <fell> But the current string say exit GC and run aqb tools
15:24:37 <Mechtilde> there are both but you can let one empty
15:24:52 <Mechtilde> fell where?
15:25:01 <jralls> Right, because cstim wrote it when he first did the UI and then forgot about it when he updated it to include IBAN & BIC.
15:25:02 <Mechtilde> ok in the error message
15:25:49 <Mechtilde> but it is enough if you change it in the assistent
15:25:50 <jralls> That's very far indeed from the only instance of error messages that are out of sync with code in other parts of the program.
15:26:41 <Mechtilde> give me a hint and I will send a corrected message in German
15:27:12 <jralls> Right, so the error message can simply be: "Your online account has neither IBAN nor SWIFT BIC set. One or the other is needed for transfers, so please edit the account in Online Banking Setup to include one of them."
15:27:23 <fell> jralls, For SF: uncover hidden features
15:27:59 <jralls> fell: Do you mean for the Project of the Month thing?
15:28:06 <fell> Yep
15:28:14 <Mechtilde> you should set both
15:28:19 <Mechtilde> IBAN AND BIC
15:28:49 <jralls> Wait, I thought you said that you don't need BIC anymore?
15:28:54 <Mechtilde> in the past I set IBAN but no BIc
15:29:12 <Mechtilde> bur Gnucash need IBAn and BIC
15:29:27 <Mechtilde> but
15:29:43 <jralls> Well, that error message triggers if you don't have both, but I was going to fix that.
15:29:43 <fell> In theory not, but there is the contruction site below
15:29:55 <Mechtilde> when I added BIC in the assistent all things works
15:31:16 <jralls> Mechtilde, can you disable the error in code and rebuild, then try with IBAN only?
15:31:40 <fell> early intl transfer went over SWIFT with BIC, SEPA uses IBAN
15:32:27 <Mechtilde> I never build gnucash successfully
15:32:48 <Mechtilde> jralls, no sorry
15:33:43 <jralls> OK. I'd sure like to test that before committing it, but its Germany only.
15:37:03 <fell> Wo hakt es denn beim bauen?
15:38:28 <fell> jralls, we could send the patch to gnucash-de and ask to test it
15:38:54 *** fekepp1 has joined #gnucash
15:39:27 *** fekepp has quit IRC
15:39:28 *** fekepp1 is now known as fekepp
15:41:02 <Mechtilde> fell Zeit und Energie ganz sauber die Buildumgebung aufzusetzen auf der einen Seite
15:41:34 <Mechtilde> und auf der anderen Seite das Wissen, aus den Fehlermeldungen die Schlüsse für korrekturen zu ziehen
15:41:35 <jralls> fell: OK. dialog-ab-trans actually calls six functions, always in BIC/IBAN pairs. Oddly, the only place it calls GetLocalBic and GetLocalIban is in that one test. Other places it calls GetBic/GetIban and GetRemoteBic/GetRemoteIban.
15:42:26 <Mechtilde> I have no experience with compiling code
15:44:00 <fell> Als ich anfing, hab ich hier die Fehlermeldungen gepostet und warlord hat mir gesagt, was ich noch installieren müßte. Nach ~2h waren wir durch.
15:46:11 <fell> jralls: Get LocalX initializes the sender address from aqb conf.
15:46:48 <jralls> As a side effect? That's naughty.
15:46:59 <fell> The user enters the receiver adress in the dialog.
15:48:10 <fell> The Starting point is one of your bank accounts.
15:48:45 <fell> you might have a bunch of them at different banks.
15:49:42 <fell> so basedon your current bank account, it looks for the address data in the aqb conf.
15:50:00 <jralls> OK. And the BIC is the bank's SWIFT Id, so it will be the same for all. But AQBanking is just sending a request to the bank to do the transfer and the bank knows its own BIC, right?
15:51:58 <jralls> It doesn't look like the logic in the function cares about the BIC, we just need to make sure that nothing breaks if it's nullptr as long as AQB will put through the transfer request without it.
15:52:42 <fell> No, the BIC of my coop bank starts with DEGENO, my accout at the municipal savings bank starts with DELB..
15:54:13 <jralls> fell: Of course. They're different banks.
15:54:56 <jralls> Oh. I meant that if you have several accounts at one bank all of those accounts will have the same BIC.
15:55:08 <fell> yes
15:55:38 <fell> But aqb stores one record for each account
15:56:00 <fell> in it's conf
15:56:52 <fell> IIRC
15:57:35 <fell> at some other point KTOBLZ is used as addressbook
15:58:12 <fell> for the national routing stuff
16:00:02 <fell> And I assume, recyceld, because german IBANS have the form DEcc<BLZ><accountnr> with cc checksum
16:01:12 <jralls> And what is a BLZ?
16:01:25 <fell> national routing number
16:02:32 <fell> BankLeitZahl
16:03:47 <jralls> So an IBAN is essentially the same as the US Routing ID and Account # scheme with a two-digit checksum thrown in because someone doesn't trust TCP's error correction.
16:04:37 <fell> +Hamming distance
16:04:57 <fell> for error detection/correction
16:06:34 <fell> While you mistyped, it will often say: Sorry, but ....
16:06:57 <jralls> Yeah, I understand error correction. But there's little point in doing your own if you're communicating over an already error-corrected channel, and most channels these days are.
16:07:01 <fell> No valid vumber
16:07:49 <jralls> Instead of just looking up the number? After all, it's not hard to construct a bogus number and calculate the right checksum.
16:08:23 <fell> The target is more the security, while entering or OCR
16:09:32 <fell> I can co with a hand filled paper form to the teller machine, which does OCR and transmits the order
16:11:11 <fell> and I am always amused over the text on my bank statements
16:11:51 *** valdas has quit IRC
16:12:13 *** valdas has joined #gnucash
16:12:38 <jralls> OCR on handwriting is not an easy problem. ;-)
16:12:39 *** valdas has quit IRC
16:12:47 *** Mechtilde has quit IRC
16:12:54 *** valdas has joined #gnucash
16:14:46 <jralls> I wrote a patch, but I realized that requiring either IBAN or BIC doesn't make sense: BIC alone isn't enough for the transfer. So the test should just be for IBAN, right?
16:15:35 <fell> we have 2 methods:
16:15:59 *** valdas has quit IRC
16:16:06 <fell> traditinal over swift with BIC+account number
16:16:22 <fell> SEPA with IBAN
16:17:52 <jralls> The test is guarded by if(gnc_ab_trans_isSEPA(td->trans_type)).
16:17:53 <fell> single transfer and SEPA single transfer
16:20:32 <fell> just for completeness: there is a 3. traditional national BLZ+account number, but thats is not what we do now.
16:21:55 <jralls> OK, I've put a patch on the bug. Please ask on gnucash-de for someone to test it.
16:22:17 <fell> OK
16:22:40 <jralls> BIAB
16:31:58 *** valdas has joined #gnucash
16:44:11 *** fekepp has quit IRC
17:13:15 *** fekepp has joined #gnucash
17:20:15 <fell> Bedtime for me ,,, zzz
17:20:19 *** fell is now known as fell_afk
17:24:20 *** valdas has quit IRC
17:24:46 *** valdas has joined #gnucash
17:27:52 *** valdas has quit IRC
17:31:18 <jralls> luc14n0: Your PR disables compiling gschemas completely, not just at installation. Are you sure that's what you want to do?
17:45:32 *** User has joined #gnucash
17:50:13 *** User has quit IRC
17:56:00 *** bob_ has quit IRC
18:04:06 <luc14n0> jralls: Yes, gschemas are automatically handled by glib. I don't know how Debian based distros manage to do it, but I'm sure they don't ship the .compiled file since it's based on what gschemas you have in your system.
18:04:52 <luc14n0> I will only change the name of the option and we will be good to go.
18:05:20 <jralls> luc14n0: No, gschemas are emphatically *not* automatically handled by glib.
18:06:02 <luc14n0> sorry, I meant in rpm based distros.
18:07:00 <jralls> You explained yesterday that the RPM system has hooks to automatically recompile when new schemas are added. I expect that other package managers like dpkg and pacman do too.
18:07:54 <jralls> But that's separate from building GnuCash. We've set up the build system so that one can run from the build directory for testing, but in order for that to work the schemas must be compiled.
18:08:42 <jralls> So are you really sure that you want to prevent compiling the gschemas in the build stage as well as the install stage.
18:08:45 <jralls> ?
18:11:50 <luc14n0> As only who manually build GnuCash will want to compile the gschema at all, yes. Who builds it manually won't must have to do anything, only who is packaging GnuCash will pass the option =OFF.
18:13:25 <luc14n0> I mean, who manually build GnuCash in his system but not to package it, for personal use.
18:13:31 <luc14n0> builds*
18:14:42 <jralls> Do you run unit tests in your package build scripts? If so, they'll break if we ever add tests that rely on the schemas.
18:19:20 <luc14n0> This kind of test no. We run a check for this kind of file that is not handled by the package itself but by our triggers/post scripts.
18:19:24 *** User has joined #gnucash
18:20:11 <jralls> No, as in do you run make/ninja check as part of the scripted build?
18:21:28 <luc14n0> Oh, no, currently lot of them are failing here (and I still haven't looked into it)
18:21:54 <luc14n0> So they will usually fail if we disable at all the compilation?
18:22:09 <luc14n0> gschema compilation*
18:23:14 *** User has quit IRC
18:24:37 * jralls is testing that now...
18:25:15 <luc14n0> I ran that modification in my Travis and it passed. So I guess it's ok.
18:26:15 <luc14n0> Oh, but I ran only in openSUSE, I deleted the Arch/Ubuntu ones.
18:26:33 <jralls> I don't think we currently do have any unit tests for the preferences, which are the primary users of gschemas.
18:26:52 <jralls> But it will take only another minute or two to know for sure.
18:27:19 <luc14n0> Actually I didn't delete Arch/Ubuntu stuff.
18:27:39 <luc14n0> Those are mine: https://travis-ci.org/luc14n0/gnucash
18:32:24 <jralls> But you didn't add -DCOMPILE_SCHEMA=OFF to https://github.com/luc14n0/gnucash/blob/unstable/util/ci/commonbuild so you didn't actually test the no-schemas build.
18:32:48 <jralls> However, it passed on my system, so you're OK for now.
18:33:39 <luc14n0> Hehe, you're right, I didn't :D
18:34:30 <jralls> What I'm concerned about is down the road we get something that needs gschemas under test and the test starts to fail in your build scripts. You write a bug. None of us remember that you disabled compiling the schemas. Much wasted effort results.
18:36:10 <luc14n0> We have change entries in our *.changes file in every package for that.
18:36:31 <luc14n0> What kind of test you think could be added some day?
18:39:37 <luc14n0> Since the only think that option disable is the gschemas compilation, the test could check for the option to conditionally run the test, no?
18:40:18 <luc14n0> I will run the test here on OBS to remember what test are failing.
18:40:31 <luc14n0> tests, better saying
18:46:21 <jralls> IIUC any unit tests of libgnucash/app-utils/gnc-gsettings.c would require compiled schemas. libgnucash/app-utils/CMakeLists.txt has no reference to COMPILE_SCHEMA so someone working in there isn't going to know to make the test conditional.
18:47:17 <jralls> (BTW, when you change the option name as fell requested, could you make it "COMPILE_GSCHEMAS"? After all there are several.)
18:48:17 <jralls> But if such tests get written wouldn't you prefer to run them along with all of the other tests as QA before pushing your RPM into distribution?
18:53:45 <luc14n0> Actually we have our own tests. We use openQA, and as we package GnuCash since 2000 and it occasionally fails is our tests nowadays GnuCash some test of its own.
18:56:25 <jralls> Keeping up with our development seems like rather a lot of work. OTOH, if you have coverage of areas we don't we'd like to get those tests upstream...
18:57:19 <luc14n0> Let me see if I can find GnuCash's openQA tests. But I don't know what exactly those tests are.
18:58:43 <luc14n0> Our openQA test are different from post build checks though. We simulate openSUSE installation and a variety of user actions are made.
19:00:39 <jralls> OIC. That kind of test is a *major* headache to maintain. I don't think we want to add them after all.
19:01:22 <luc14n0> You would need your own openQA instance
19:01:38 <luc14n0> have a look on this: https://openqa.opensuse.org/
19:06:21 * luc14n0 is still trying to find GnuCash's openQA tests.
19:09:58 <jralls> Very nice display, but I don't see GnuCash getting tested.
19:29:11 <luc14n0> Well, I can't find it, later I will ask the folks who take care of openQA tests. For now I will send a new PR.
19:29:54 <jralls> OK.
19:32:04 <luc14n0> I will put a notice in the libgnucash/app-utils/CMakeLists.txt about the new option as well.
19:52:19 <luc14n0> jralls: what do you think of: +# Short notice: When designing a unit test for gnc-gsettings,
19:52:19 <luc14n0> +# don't forget to consider the COMPILE_GSCHEMAS option for
19:52:20 <luc14n0> +# conditional run of the test.
19:52:20 <luc14n0> SET (app_utils_SOURCES
19:52:20 <luc14n0> calculation/expression_parser.c
20:31:40 <jralls> luc14n0: It could be a bit more idiomatic, maybe "NB: Unit tests which require GSchemas should be made conditional on COMPILE_SCHEMAS."
20:32:47 <luc14n0> jralls: yeah, much better indeed.
21:38:35 *** frakturfreak has quit IRC
22:14:43 *** oozer has quit IRC
23:56:22 *** ncv has quit IRC