Time  Nick           Message
09:36 imi            can I make 2 PRs in top of each other? or instead should I clone a new repo for my new PR?
09:55 pgimeno        you don't have to clone twice, you can create a feature branch, then checkout master, then create another independent feature branch
09:55 pgimeno        there's a reason they are called branches :)
09:56 pgimeno        sending PRs from master is HIGHLY discouraged
10:00 imi            do I do it from github.com or do I do it from commandline?
10:24 Zughy[m]       as you prefer
10:24 Zughy[m]       why do we have a "WIP" label for PRs when people can just turn their PRs into drafts?
10:25 Zughy[m]       sfan spotted #13092
10:25 ShadowBot      https://github.com/minetest/minetest/issues/13092 -- Lua on each mapgen thread by sfan5
10:26 sfan5          🤷
10:26 sfan5          our label predates the github feature
10:29 imi            I want to know where this line of log is generated:
10:29 imi            2023-06-11 12:24:27: ACTION[Server]: Imre moves default:chest to chest at (0,7,-5)
10:29 imi            grep -HRniF "to chest at" yielded no result
10:30 imi            as far as I can tell that means no result in translation files, no lua, no cpp
10:48 pgimeno        I'd consider the possibility that it's parametrized everywhere: %s moves %s to %s at (%d,%d,%d)
10:52 imi            hm...
10:53 imi            I wouldn't have guessed that based on the fact that for a protected chest the same line appears
11:01 imi            for grep -HRniF ' moves ' there are lots of results, so I checked grep -HRniF ' moves ' _CPack_Packages as well, nothing useful (despite having a _CPack_Packages/Linux/TGZ/minetest-5.7.0-linux/games/minetest_game dir)
11:04 imi            ok I found it, was harder than expected, thanks for the help
11:05 imi            _CPack_Packages/Linux/TGZ/minetest-5.7.0-linux/games/minetest_game/mods/default/functions.lua line 747
11:51 imi            aren't there any bots announcing if someone makes a PR against minetest game? anyhow, I made this: https://github.com/minetest/minetest_game/pull/3040 -- I didn't know what else to say as description given the fact it's a 8 liner, with commit messages explaining what's going on
11:54 pgimeno        there are no auto bots, you can announce it yourself by writing game#3040
11:54 ShadowBot      https://github.com/minetest/minetest_game/issues/3040 -- Bone logging by imre84
11:54 pgimeno        the bot catches that and posts the URL and title, which lets everyone know what it's about
12:00 Desour         merging #13348 in 15
12:00 ShadowBot      https://github.com/minetest/minetest/issues/13348 -- Move the platform-dependent stuff in renderingengine.cpp to irrlicht by Desour
12:00 imi            also, I'm going to ask this as well without spending too much effort on this one: I'm planning on making a chatcmd_parse_coords function to remedy the situation that everytime there are coords involved in a chatcommand it's a different syntax: sometimes it's x,y,z, sometimes it's x y z, sometimes it's (x,y,z) sometimes tilde notation is supported sometimes it isn't. I'm planning on making a function like this: local
12:00 imi            coords,param_rest=chatcmd_parse_coords(param,base_pos) -- on failure it should return nil,param untouched otherwise return with coords parsed and anything that is left behind in param after parsing the coords (therefore you can have further parameters to your chatcommand). what do you think? I think from az admin perspective this would be a gamechanger (provided that chatcommands use this)
12:01 imi            is it a good idea? if so I might be doing it
12:05 Desour         imi: AFAIK there's a function in builtin that does at least parts of this. making it a public API would be useful. but be warned that trying to add helpers to builtin can be very troublesome
12:06 imi            I've seen that it supports (x,y,z) with tilde notation included
12:07 imi            /deleteblocks uses that
12:12 MTDiscord      <josiah_wi> Desour is a core dev? Is that recent?
12:12 Desour         yes. yes.
12:12 MTDiscord      <josiah_wi> Congratulations.
12:12 Desour         thx
12:13 MTDiscord      <josiah_wi> I reviewed part of your activeobjectmgr cleanup and it inspired me to make #13577 in addition to a couple others. This one is going to conflict with your PR; do you want it merged before or after?
12:13 ShadowBot      https://github.com/minetest/minetest/issues/13577 -- Extract and pull up assignFreeId from CAO/SAO by JosiahWI
12:17 Desour         I'd like to get mine in first, tbh, as it contains bugfixes. doesn't really matter though, as long as rebases are easy
12:18 MTDiscord      <josiah_wi> Excellent, I'll keep this one marked as draft and rebase it once yours is in.
17:25 Zughy[m]       meeting in 35 minutes, but there are no points. What do you want to do?
17:26 sfan5          who is even there?
17:29 MTDiscord      <josiah_wi> I have 3 small PRs up that are ready for review; one has a core dev approval.
17:32 Desour         o/ hi, I'm here
17:35 Krock          I'd be around if a meeting is supposed to take place
17:39 Desour         why was it again that we skip the gcc_5 and clang_3_9 CI builds?
17:40 sfan5          github got rid of the ubuntu 18.04 image
17:41 Desour         ah right. (also found the commit now)
17:52 nrz            Time to add Debian 12 to our builds
17:54 Desour         github only offers ubuntu 20.04 and 22.04, and non-posix, it seems. no debian. https://docs.github.com/en/actions/using-github-hosted-runners/about-github-hosted-runners#supported-runners-and-hardware-resources
17:56 Desour         looks like ubuntu 18.04 has ran out of support (apart from security updates) this year (according to <https://ubuntu.com/about/release-cycle>)
17:59 sfan5          docker is usable via gh actions to use any distro but it caused some weird behaviour regarding shell which means it isn't a drop in replacement
18:00 sfan5          irrlichtmt uses docker to still get 18.04 builds for example
18:00 srifqi         hello
18:39 Desour         (made an issue #13583)
18:39 ShadowBot      https://github.com/minetest/minetest/issues/13583 -- [discussion] Define a policy for raising our dependencies' versions
18:45 srifqi         i agree in creating a policy/strategy for dependencies, but not sure about the implementation
18:45 srifqi         for example, the Android dependencies are updated as soon as we noticed that a new version is available
18:48 Desour         my idea was that if someone wants to upgrade dependencies, they can apply the policy (aka look at what the supported distros have) to see if the change will be accepted
18:49 Desour         we don't need to update as soon as something updates
18:54 srifqi         uh ... my example is not a good example since we basically follows Google Play guidances
18:54 srifqi         how old of a system do we support?
18:55 rubenwardy     Android dependencies are different as we are directly in control of them
18:55 Desour         (idk anything about android development, so I might've misunderstood you)
18:55 Desour         we support a range of ages. idk which is the oldest
18:57 srifqi         oh, yeah. we do have minetest_android_deps repo
19:02 srifqi         i was going to check e.g. the architecture supports for gcc 5.5, but it is still a lot of architectures: https://gcc.gnu.org/onlinedocs/gcc-5.5.0/gcc/x86-Options.html
19:03 srifqi         i guess the easiest way is to check for versions on every supported distros and oses
19:03 srifqi         (as proposed)
19:04 Desour         hardware archs are probably not too relevant for us
19:04 Desour         it's not defined which distros are supported. this is the issue
19:05 srifqi         i'm not familiar with Linux distros (i mainly use Windows)
19:05 srifqi         (ok, maybe also Lubuntu in a VM)
19:07 srifqi         if Ubuntu family is really conservative, i guess that should help in deciding "minimal" version
19:07 srifqi         in Windows, we should just use whatever vcpkg install
19:08 MTDiscord      <josiah_wi> Debian oldstable may be reasonable as well.
19:10 srifqi         i'm guessing this one: https://wiki.debian.org/DebianBullseye
19:10 srifqi         https://wiki.debian.org/DebianOldStable
19:11 Desour         debian also has lts https://wiki.debian.org/LTS
19:12 srifqi         is the difference only in the handler/maintainer?
19:13 MTDiscord      <josiah_wi> I think the current LTS is also Debian oldstable.
19:14 Desour         lts is 10, oldstable is 11
19:14 srifqi         it is Buster (the release before Bullseye) in this article: https://wiki.debian.org/LTS
19:15 Desour         lts is only community supported, according to wiki
19:15 MTDiscord      <josiah_wi> Interesting. I did not realize there's already been another Debian stable release.
19:15 srifqi         oh, btw, i'm going to bed. see you
19:16 MTDiscord      <josiah_wi> No wonder - it was literally released yesterday.
19:17 Desour         the good thing is the debian lts's gcc (8) already has most c++17 features https://en.cppreference.com/w/cpp/compiler_support/17
19:18 sfan5          IMO we shouldn't bump our minimum requirements "just because"
19:19 sfan5          if there's an usecase then sure
19:19 nrz            yep, C++17 is clearly cool, but let's wait a bit for debian 12 to be more used before removing old things
19:19 Desour         the new c++17 features have useful usecases
19:19 nrz            i think unsupported officially distro should be removed
19:19 sfan5          for gcc and clang it's relevant which versions we can reasonably test
19:19 nrz            c++17 std::filesystem can permit to drop all our filesystem code, or nearly, yes
19:20 Desour         nrz: we don't have to wait for debian 12. debian 10 already has gcc 8
19:20 nrz            gcc 8 has complete c++17 support ? if i remember it wasn't the case no ?
19:20 nrz            for ubuntu, what is the oldest LTS used, i don't talk about 10 years support LTS
19:20 sfan5          but for example there is no point in raising the minimum sqlite version, MT probably works with any 3.x you can find
19:21 nrz            we should just have supported version no ?
19:21 nrz            we don't use advanced feature, only supported version is sufficient
19:21 sfan5          (we don't even check the sqlite3 version in cmake)
19:22 Desour         I agree, there's no reason to upgrade the version without reason. but a proper policy would allow easily bumping the version when someone finds it very useful
19:23 MTDiscord      <josiah_wi> Minetest does not work with SQLite on Debian Jessie.
19:23 nrz            at least we should ensure we set the minimum upstream supported version i'd say no ?
19:23 nrz            i don't know where sqlite3 is or other libs
19:23 sfan5          anyway for estimating which compilers or libraries we can require I have suggested in the past to check the oldest still-supported Ubuntu, Debian and maybe Fedora
19:24 nrz            i agree
19:24 nrz            did we evicted debian 9 yet ?
19:25 MTDiscord      <josiah_wi> Debian 8 apparently still has extended LTS.
19:25 Desour         debian 9 is outscheduled, according to https://wiki.debian.org/LTS
19:25 sfan5          in what sense? we don't have debian builds
19:25 nrz            on gitlab.com we had some before, it's deprecated ? 🙂
19:26 sfan5          the gitlab pipelines have been reduced to a minimum to not waste compute minutes
19:26 sfan5          they also haven't ran since 2 months so they're basically broken
19:26 nrz            ok
19:27 Desour         btw. for reasonable testing, if we don't want to use docker in the ci, github has some markdown file here with a list of supported versions: https://github.com/actions/runner-images/blob/main/images/linux/Ubuntu2004-Readme.md
19:27 Desour         (min gcc 9.4.0)
19:31 sfan5          github adds lots of custom stuff to their runners so that's not entirely indicative of what ubuntu can support
19:32 sfan5          e.g. https://packages.ubuntu.com/search?keywords=gcc-7 on focal
19:34 sfan5          anyway if someone wants C++17 then gcc-7 has most features (gcc-8 for minor ones), or clang-5
19:36 Desour         gcc-7 is probably harder to test for us if github advertises gcc-8, but not 7
19:37 sfan5          you can apt install it
19:37 Desour         ah, ok then
20:30 schwarzwald[m] Do we have a way to see the test failure when a unit test fails? Scrolling way up to read the failure message is pretty inefficient.
20:31 Desour         only your terminal emulators search function, sadly
20:51 schwarzwald[m] Maybe we could start converting unit tests to Catch.
20:57 schwarzwald[m] Or have an option to immediately stop the test run if one fails.
22:24 imi            we'll hopefully talk about this tomorrow, but until then I'll just post the link here for anyone to see and/or play with it: https://onecompiler.com/lua/3zbav7368
23:03 imi            it keeps evolving, so find it here instead: https://github.com/imre84/chatcmd_parse_coords