2020-03-29 GnuCash IRC logs
00:02:47 *** omnireq has quit IRC
01:29:42 *** fell_laptop has joined #gnucash
01:29:42 *** ChanServ sets mode: +o fell_laptop
01:31:54 *** fell has quit IRC
01:49:56 *** sbluhm has joined #gnucash
01:49:56 *** ChanServ sets mode: +v sbluhm
01:52:35 *** jervin has quit IRC
02:01:17 *** fell_laptop is now known as fell
02:12:19 *** gjanssens has joined #gnucash
02:12:19 *** gncbot sets mode: +o gjanssens
02:12:19 *** ChanServ sets mode: +o gjanssens
02:29:59 *** User has quit IRC
02:34:23 *** Mechtilde has joined #gnucash
03:00:05 *** storyjesse has joined #gnucash
03:06:25 *** sbluhm has quit IRC
03:10:42 <chris> gjanssens: did you have a good holiday?
03:14:23 <gjanssens> chris: it was great. I really enjoyed it
03:15:12 <gjanssens> And so did the wife
03:15:31 <chris> nice, a break away from all the craziness around!
03:16:33 <chris> and now some more break at home!
03:28:14 <gjanssens> :)
03:28:24 <gjanssens> Yes, time for our morning walk :D
03:48:19 *** Mechtilde has quit IRC
03:49:41 *** Mechtilde has joined #gnucash
03:53:26 *** sbluhm has joined #gnucash
03:53:26 *** ChanServ sets mode: +v sbluhm
03:55:02 *** jervin has joined #gnucash
03:58:21 *** omnireq has joined #gnucash
04:46:37 *** storyjesse has quit IRC
05:22:12 *** Aussie_matt has quit IRC
06:17:33 *** lmat has quit IRC
06:17:37 *** lmat has joined #gnucash
06:19:59 *** Gerd has joined #gnucash
06:22:42 *** sbluhm has quit IRC
06:24:54 *** lmat has quit IRC
06:26:08 *** lmat has joined #gnucash
06:46:53 *** Gerd has quit IRC
07:00:51 *** lmat has quit IRC
07:00:54 *** lmat has joined #gnucash
07:03:35 *** sbluhm has joined #gnucash
07:03:35 *** ChanServ sets mode: +v sbluhm
07:04:47 *** Gerd has joined #gnucash
07:33:28 *** Aussie_matt has joined #gnucash
07:40:08 *** mauritslamers has quit IRC
08:15:55 *** storyjesse has joined #gnucash
08:24:43 *** mauritslamers has joined #gnucash
08:24:43 *** ChanServ sets mode: +v mauritslamers
08:26:05 *** Aussie_matt has quit IRC
08:41:43 *** lmat has quit IRC
08:41:46 *** lmat has joined #gnucash
08:55:09 *** lmat has quit IRC
08:56:25 *** lmat has joined #gnucash
08:58:30 *** User has joined #gnucash
09:04:52 *** Gerd has quit IRC
09:21:41 *** Jimraehl1 has joined #gnucash
09:23:39 *** Jimraehl1 has quit IRC
09:53:30 *** lmat has quit IRC
09:54:12 *** lmat has joined #gnucash
09:59:51 *** sbluhm has quit IRC
10:10:56 *** storyjesse1 has joined #gnucash
10:11:18 *** storyjesse has quit IRC
10:11:18 *** storyjesse1 is now known as storyjesse
10:58:25 *** lmat has quit IRC
11:00:53 *** lmat has joined #gnucash
11:05:28 *** Gerd has joined #gnucash
11:06:58 *** storyjesse1 has joined #gnucash
11:07:22 *** storyjesse has quit IRC
11:07:22 *** storyjesse1 is now known as storyjesse
11:07:58 *** sbluhm has joined #gnucash
11:07:58 *** ChanServ sets mode: +v sbluhm
11:22:04 *** chris has quit IRC
11:24:01 *** chris has joined #gnucash
11:24:01 *** ChanServ sets mode: +v chris
11:25:36 *** lagertonne has quit IRC
11:29:23 *** lagertonne has joined #gnucash
12:02:18 *** delli3 has quit IRC
12:22:14 *** sbluhm has quit IRC
12:55:19 *** Mechtilde has quit IRC
12:56:24 *** sbluhm has joined #gnucash
12:56:24 *** ChanServ sets mode: +v sbluhm
12:56:39 *** Han has quit IRC
12:58:01 *** jervin has joined #gnucash
12:58:33 *** Mechtilde has joined #gnucash
13:16:34 <jralls> chris, re -fanalyzer, it's about time. Clang has had that forever. I ran GnuCash's code through it and cleaned up the failures in fall 2018.
13:17:56 <jralls> It has its limits, though. It can't see ptr ownership problems and doesn't know about GObject reference counting (which we badly under-use).
13:20:22 <jralls> Unfortunately the author of that piece is severely confused. Static analysis can't catch double-frees unless they happen to be in a flow-of-execution contained in a single translation unit (which usually means a single C file).
13:21:13 <jralls> Clang has another feature for finding those, https://clang.llvm.org/docs/MemorySanitizer.html, but it works at run time, not compile time.
13:30:55 *** sbluhm has quit IRC
13:36:09 <jralls> gjanssens: I'm confused by https://wiki.gnucash.org/wiki/Release_Process#Flatpak_Executable. I ran the script but the only org.gnucash.Gnucash.json in the directory is the one from https://code.gnucash.org/gnucash-on-flatpak. Surely I'm not supposed to push that to https://github.com/flathub?
13:38:10 <gjanssens> jralls: that is slightly out of date :(
13:38:25 <gjanssens> Give me a minute to re-verify what should happen now
13:38:31 <jralls> OK.
13:41:21 <jralls> Meanwhile, do I have push on that repo?
13:44:02 <gjanssens> I don't know. Have you ever requested it ? I am not the owner of that repo and don't have management access myself.
13:45:40 <jralls> I haven't.
13:49:25 <gjanssens> Hmm, then you won't be able to start the flathub build.
13:49:52 <gjanssens> That aside, the following should happen in the flathub repo:
13:50:40 <gjanssens> 1. update gnucash-source.json and gnucash-docs-source.json to point at the new tarballs
13:53:36 <gjanssens> 2. tag the commit with a build id (3.9 for the base, 3.9-x for any subsequent changes in the 3.9 series, like an updated aqbanking release)
13:54:25 <gjanssens> 3. Update gnucash.releases.xml based on the output of git log --date=short --format="<li>%ad - %s (%an)</li>" <major.minor-pkg>..HEAD
13:54:42 *** storyjesse has quit IRC
13:55:13 <gjanssens> 4. Update the BUILD_ID passed to the gncuash build command in org.gnucash.GnuCash.json to match the tag set
13:55:36 <gjanssens> And I think step 2. should actually happen last.
13:56:49 <gjanssens> I propose I'll do it for this release and record more precisely what to do while I'm doing it.
13:57:14 <gjanssens> For the next release we have to get you write access to the flathub repo.
13:58:21 *** jervin has quit IRC
13:58:27 <jralls> OK, that seems reasonable. FWIW the two json files are build-script products, they're not committed to the repo. It would be nice to have gnucash.releases.xml work that way too.
14:01:03 <gjanssens> These two files are build-script products in gnucash-on-flatpak, they are committed in the flathub repo.
14:01:58 <gjanssens> I can see if I can do something similar for gnucash.releases.xml, though it is only meaningful for the flathub repo.
14:02:03 <gjanssens> Time for supper...
14:02:09 <jralls> OK, later.
14:21:20 *** sbluhm has joined #gnucash
14:51:04 *** Gerd has quit IRC
14:56:54 <gjanssens> jralls: gnucash-on-flatpak can only build release packages if 1. gnucash and gnucash-docs are tagged 2 the release tarballs are available on sourceforge
14:57:13 <gjanssens> So far I think you only have tagged gnucash, not gnucash-docs ?
14:57:47 <jralls> And 3 the README.txt file has been created and uploaded to SourceForge.
14:58:07 <gjanssens> Right. Required for the shasums
14:58:07 <jralls> But it shouldn't care about tags if it's working off of the tarballs.
14:58:43 <jralls> Just pushed the docs tag.
14:59:06 <gjanssens> It checks the -r value against existing tags to discriminate between release and development builds
14:59:23 <gjanssens> If the tags exist the script will take this as a release
14:59:54 <jralls> Huh? I thought that's what the -c parameter did.
15:00:35 <jralls> I'd rather not play the tags game we did in the old Windows build scripts. That was a real PITA.
15:00:36 <gjanssens> No. The -c was added as sometimes the release version doesn't match the tarball name.
15:01:02 <gjanssens> Windows has a very unreliable build. Flatpaks don't.
15:01:10 <jralls> Sorry, wrong one. I should have said the -r parameter.
15:01:33 <gjanssens> Ah. '-r' stands for 'revision', not release.
15:01:47 <gjanssens> It can be a branch name or a tag name, or even a commit hash
15:02:24 <gjanssens> The script needs something to decide whether it's a release or not.
15:02:28 <jralls> Well, the wiki page says otherwise, but you *did* say that it's out of date.
15:02:46 <gjanssens> For now that's the presence of a tag.
15:03:14 <gjanssens> The script is very much written with full automation in mind at some point.
15:03:28 <gjanssens> But I'd rather not have that discussion tonight.
15:03:47 <gjanssens> The gnucash-on-flatpak release build is running on my local system.
15:04:01 <jralls> Automation is good.
15:04:48 <jralls> I think you'll need to write the flathub admins to get me push on the org.gnucash.Gnucash repo. You should probably add fell too.
15:05:20 <fell> Yep
15:07:34 *** TownsendHardware has joined #gnucash
15:08:03 *** psmst has quit IRC
15:08:03 <jralls> For another we-can-discuss-it-later item, it would be nice if the build script could be in python instead of sh.
15:09:34 *** psmst has joined #gnucash
15:12:37 <gjanssens> Ok. I was initiated in sh when I first got introduced to linux, so I tend to fall back to it easily. But python should be fine as well.
15:12:49 <gjanssens> I have requested write access for both of you.
15:12:54 <jralls> Thanks.
15:13:50 <jralls> Yes, sh is a fine fall-back common denominator that should be always available. So is C. That doesn't make either of them the best choice. ;-)
15:20:53 <gjanssens> :D True
15:25:20 *** frakturfreak has joined #gnucash
15:25:20 *** ChanServ sets mode: +v frakturfreak
15:27:06 <gjanssens> flathub build has been triggered. I'll update the documentation tomorrow - time to leave.
15:30:54 *** jralls changes topic to "Free GPL Personal and Small Business Accounting || Moderation note: To get voice register with NickServ, then re-join. || Please don't ask to ask, just ask and wait! (Possibly a few hours!!) || publicly-logged channel || latest stable: 3.9 || https://www.gnucash.org || https://wiki.gnucash.org/wiki/"
15:31:07 <jralls> Good night.
15:32:05 <gjanssens> jralls: final note: the wiki only focuses on release builds. Perhaps that's what got you to take -r as release number only.
15:32:11 <gjanssens> Anyway see you tomorrow!
15:32:26 <jralls> As it should.
15:32:36 *** gjanssens has quit IRC
15:34:02 <jralls> @tell gjanssens But the wiki says " You can omit -c or -d if they are the same as -r, but always specify at least -r (as that will otherwise default to maint resulting in a development build or a tarball not found error)." That certainly implies that the -r flag is what tells the build script that it should do a release build.
15:34:02 <gncbot> jralls: The operation succeeded.
15:34:49 *** frakturfreak has quit IRC
15:41:33 *** omnireq_ has joined #gnucash
15:42:45 *** omnireq has quit IRC
15:58:01 *** jervin has joined #gnucash
16:25:51 *** sbluhm has quit IRC
16:34:09 *** User has quit IRC
16:36:49 *** oozer has joined #gnucash
17:01:34 *** Gerd has joined #gnucash
17:41:36 *** lmat has quit IRC
17:42:32 *** lmat has joined #gnucash
18:20:12 *** Aussie_matt has joined #gnucash
18:35:51 *** Gerd has quit IRC
18:38:21 *** jervin has quit IRC
18:38:24 *** jervin has joined #gnucash
19:18:29 *** jervin has quit IRC
19:22:02 *** omnireq_ has quit IRC
19:34:32 *** chris has quit IRC
19:35:38 *** chris has joined #gnucash
19:35:38 *** ChanServ sets mode: +v chris
20:05:06 *** oozer has quit IRC
20:20:12 *** bertbob has quit IRC
21:04:35 *** omnireq has joined #gnucash
21:07:07 *** lmat has quit IRC
21:07:09 *** lmat has joined #gnucash
21:57:47 *** bertbob has joined #gnucash
21:57:48 *** ChanServ sets mode: +v bertbob
22:37:24 *** jervin has joined #gnucash
22:37:45 *** jervin has quit IRC
23:21:57 *** omnireq has quit IRC
23:53:05 *** jervin has joined #gnucash
23:57:21 *** jervin has quit IRC