Time  Nick       Message
00:27 Zughy[m]   sfan5: so are we basically waiting for a backtrace in #12487 ?
00:27 ShadowBot  https://github.com/minetest/minetest/issues/12487 -- minetestserver crashes following random CPU spike
00:40 MTDiscord  <savilli> I would say, not just for a backtrace, but for a backtrace with debug symbols
00:42 MTDiscord  <savilli> for me it doesn't look like a LuaJIT bug tho
00:43 MTDiscord  <savilli> LuaJIT bugs don't produce such a clear backtrace
00:49 MTDiscord  <savilli> for example #12594 definitely looks like a LuaJIT bug
00:49 ShadowBot  https://github.com/minetest/minetest/issues/12594 -- Raspberry pi Segmentation fault
08:05 sfan5      Zughy[m]: a meaningful backtrace and/or test results with plain Lua or up-to-date LuaJIT
08:20 sfan5      merging #12607 in 10m
08:20 ShadowBot  https://github.com/minetest/minetest/issues/12607 -- [nosquash] Update credits tab by sfan5
10:26 lionkor    hey, quick question: the .clang-format seems to break lines in main.cpp that were not broken before. what's up with that?
10:26 lionkor    should I use it anyways? or just ignore the clang-format? the column=90 limit seems not enforced in this file.
11:16 lionkor    okay, i just did the formatting manually now, i hope it's good that way and I didn't miss any clang-format rules (pr #12641)
11:16 ShadowBot  https://github.com/minetest/minetest/issues/12641 -- Add handling of CLICOLOR, MINETEST_COLOR environment variables (#7553) by lionkor
11:59 sfan5      the clang-format is close but not exactly the intended code style, which is why we stopped using it in CI
15:27 rubenwardy I think "game platform" or "game creation platform" is a better descriptor for Minetest than "game engine"
15:38 MTDiscord  <Warr1024> Both are good, though I worry that the inclusion of "creation" puts too much emphasis on users being expected to create, as if there isn't a lot of already-created content available for playing.
16:00 MTDiscord  <wwar> "game creation platform, you can also play though"
17:33 x2048      Can we add https://irc.minetest.net/ to the chat topic? Would be so easy to navigate to the logs...
18:03 kilbith    so 5.6 when?
18:03 kilbith    if you keep doing long freezes on too frequent releases (<= 6 months), you end up getting nowhere and it'll bore everyone
18:24 rubenwardy we have provisions to use a dev branch but we never do
18:26 kilbith    yeah a freeze should never stop the dev, just like a x.y.1 release
18:27 kilbith    we lose 3 weeks already, during which x2048 could have merged his shaders
18:27 kilbith    3 weeks of almost nothing interesting
18:27 rubenwardy don't the shaders still have issues?
18:29 x2048      They still have, there's some work to be completed on the PR, and I've been helping with the fixes for 5.6.0
18:30 x2048      I am working on the stuff, keep the branches fresh, so I don't think we are totally stuck, even if there is no big movement right now.
18:31 kilbith    releasing may get people back to reviews
18:32 kilbith    I just wish there won't be too much releases like that during the year
18:33 kilbith    one time a year is enough
18:33 rubenwardy I'd rather have 2-3 releases a year with a short feature freeze, like a week
18:34 kilbith    that'd work too
18:34 rubenwardy it allows a much quicker cycle when it comes to release features and iterating
18:34 rubenwardy Feature freezes are only supposed to be a week long
18:37 sfan5      in this case it looks like the freeze was the right size to catch relevant issues
18:38 sfan5      I'm sure it could be condensed though if people had more time
18:39 sfan5      releasing once a year is also how you get people filling twenty bugs during freeze because they never test games/mods with dev versions
18:40 rubenwardy yeah. I lost time just as the freeze started
18:40 rubenwardy doesn't https://irc.minetest.net/minetest-dev/ make more sense
18:41 MTDiscord  <luatic> We need to encourage people to test dev versions more
18:41 rubenwardy guess it doesn't matter that much
18:41 rubenwardy Yeah, I failed a bit with the RC release
18:42 sfan5      can we release now then?
18:43 rubenwardy Our last blocker is done, so I guess so
18:49 sfan5      I'll prepare the commits and windows builds then
19:46 sfan5      builds uploaded and tested
19:49 MTDiscord  <Warr1024> hype
20:01 sfan5      well, time to push?
20:02 Krock      should be fine, I think.
20:05 Krock      no bugs in the 5.6.0 milestone so we're good on track, even though I didn't have time to check all the recent reports
20:12 rubenwardy The release forum topic should just link to the blog. Finishing up the post now
20:12 rubenwardy Could do with a gif of the animated particle spawners
20:13 rubenwardy Hopefully there's something in devtest
20:16 sfan5      repos tagged and pushed, will post a topic now
20:28 sfan5      uhhhhhhhhhhh
20:28 sfan5      did we bump the protocol version this release
20:29 sfan5      particles I guess
20:29 rubenwardy no you didn't
20:29 rubenwardy or maybe CDB is wrong
20:30 sfan5      63fc728a84 src/network/networkprotocol.h (Wuzzy               2021-06-24 18:21:19 +0000  213) #define LATEST_PROTOCOL_VERSION 40
20:30 rubenwardy nope, 5.5.1 and 5.6.0 are both 40
20:30 rubenwardy is this in the checklist
20:30 sfan5      yes
20:30 sfan5      there's also a comment in the cmakelists.txt
20:31 sfan5      but it's very pointless because editing that is handled by a script
20:31 rubenwardy need to add a big red one to bump_version.sh
20:31 sfan5      the checklist is kinda all over the place
20:32 sfan5      https://0x0.st/oetN.png I started doing stuff from here
20:33 sfan5      so do we want to quickly erase 5.6.0 or is this 5.6.1 now?
20:33 sfan5      or just not do it
20:34 rubenwardy having multiple versions with the same name will cause confusion
20:34 rubenwardy but idk, it's only the protocol version
20:34 rubenwardy could wait a few days and see what issues come up, and then fix them as part of 5.6.1
20:35 sfan5      that's not going to work as well as it sounds but I'll agree to everything that doesn't involve me doing the thing over again today
20:36 rubenwardy what will have used the tag so far? It's been 20 minutes
20:37 sfan5      the builds I uploaded obv
20:37 rubenwardy do distros watch for tag creation?
20:38 rubenwardy I imagine they poll it
20:38 sfan5      yes but also why does that matter, updates are not automatic
20:40 rubenwardy Might cause conflicts in their systems. I'd say that it doesn't matter too much if there's confusion over the two versions as it's just a protocol version change, so don't mind an erase too much
20:41 sfan5      ok then
20:55 sfan5      new tags pushed and stuff
20:55 sfan5      (win builds too)
20:59 Zughy[m]   congratulations, people ❤️
21:01 Zughy[m]   oh no, now every PR results updated 18 minutes ago
21:02 rubenwardy github bug when force pushing :(
21:03 MTDiscord  <Ronoaldo> Would it worth planning an script to automate the release steps? Just to avoid pressure on the already busy dev folks ?
21:06 rubenwardy there is already a script but a lot is also manual
21:06 rubenwardy doing continuous deliver would be good, but there's a lot we'd need to do
21:07 MTDiscord  <Ronoaldo> I mean, reading the steps from sfan print I see some of these as things that could be automated
21:07 sfan5      the "update version in source" part is
21:08 sfan5      already
21:08 rubenwardy Ideally, all the builds would be done by CI and automatically added to the release
21:08 rubenwardy not sure how to handle the apk signing, I'd probably need to download and sign it and reupload
21:09 MTDiscord  <Ronoaldo> makes sense
21:09 rubenwardy if you want to be extra fancy you can deploy to places like Google Play and Itch from CI
21:10 MTDiscord  <Ronoaldo> I'm probably automation freak so don't listen too much but I got sort of tired of doing some stuff manually and more used to an "automate first" approach as much as possible.
21:10 MTDiscord  <Ronoaldo> i recall one script that did a lot of stuff which I called "release-the-kraken.sh" to tag/release/deploy all the things for instance
21:13 rubenwardy You can tell that Android Studio is compiling because my laptop is lagging
21:14 rubenwardy running optpng at the same time probably doesn't help
21:15 sfan5      can't you literally download the CI build and resign it
21:15 sfan5      unlike gitlab where the version is wrong (it will say 5.6.0-abcdef instead of 5.6.0) the android artifct should be literally the same
21:15 sfan5      the windows builds on gitlab*
21:16 sfan5      anyway: website updated
21:16 sfan5      I leave the rest in the list to someone else
21:25 MTDiscord  <ROllerozxa> the CI apks don't have minetest game or irrlicht shaders bundled last I checked, maybe something to look into for the next release
21:27 sfan5      mtg: ah could be; irrlicht: this should not happen though
21:59 rubenwardy Android build crashes when testing
22:00 rubenwardy will debug in a bit
22:22 MTDiscord  <savilli> in irr::video::COGLES1Driver::updateVertexHardwareBuffer?
22:28 Zughy[m]   Does it make sense to keep #11698 open, considering we're already two versions away?
22:28 ShadowBot  https://github.com/minetest/minetest/issues/11698 -- Async "mapblock version mismatch error" seems to be stuck in deadlock sometimes