Time  Nick       Message
01:02 Hijiri     Should I put m_ before member variable names
01:02 Hijiri     for new classes
01:03 Vexyl      "The prefix m_ is discouraged for public members in newer code as it is a part of the class' interface, but sometimes needed for consistency when adding a member to older code. " https://dev.minetest.net/Code_style_guidelines
01:04 Hijiri     ok
01:04 Hijiri     thankerino
01:05 Vexyl      fine for private members?
01:05 Hijiri     This will be a public member
01:07 paramat    yes m_ is fine for private members
01:07 Hijiri     fine or encouraged?
01:07 Hijiri     I am actually adding a private member too
01:09 paramat    hmm, probably encouraged for consistency
01:09 paramat    but not totally sure
01:12 Hijiri     Should I avoid using auto
01:13 Hijiri     is dev wiki self-signed?
01:13 Hijiri     I am getting a security warning
01:25 rubenwardy Dev wiki doesn't support ssl
01:25 rubenwardy Well, not properly
01:30 paramat    i suggest only using auto if you really need to, as it makes code less readable and clear
02:05 paramat    auto is used in MT code so it seems fine when needed
03:47 Hijiri     I don't absolutely need to
03:47 Hijiri     it would just be standing in for std::pair<thing, other_thing> from a map
04:11 Cornelia   I'd recommend a typedef or a using statement in that case... Well.. that's how I'd do it.
04:19 Hijiri     It's only used in one place though
05:20 Hijiri     PR #7269
05:20 ShadowBot  https://github.com/minetest/minetest/issues/7269 -- Physics modifiers by raymoo
05:23 Hijiri     It's a fairly small feature so hopefully I don't have to mergebeg too long
20:07 opal       going to prepare a patch
20:17 opal       oh, request-pull works differently than i assumed
20:20 opal       https://paste.debian.net/hidden/7292d252/ (direct: https://paste.debian.net/downloadh/7292d252)
20:21 sfan5      you shouldn't be using any pthread types outside of #if USE_POSIX_THREADS
20:21 opal       `git am` will apply the patch
20:21 opal       i dont think i was
20:22 sfan5      >// Ignore unused variable warning ifndef USE_POSIX_THREADS
20:22 sfan5      implies that you do
20:22 paramat    if you're unable to submit a PR you could post this code in a github issue for someone else to apply
20:22 opal       paramat: for the third time, i do not have a github account
20:22 p_gimeno   opal, maybe you're looking for format-patch?
20:23 opal       please stop suggesting that
20:23 opal       p_gimeno: yeah, thats what i ended up using
20:23 paramat    i didn't know you didn't and won't
20:24 opal       ok editing the patch to wrap the one var declaration in an ifdef
20:24 opal       escaped my head when i was writing it
20:25 opal       fixed: https://paste.debian.net/hidden/a3b157b4/ and raw: https://paste.debian.net/downloadh/a3b157b4
20:26 paramat    sorry, i don't remember who does or does not have github accounts :)
20:26 opal       i said this yesterday, shouldnt be something you have to go out of your way to remember
20:26 opal       but i'll stop talking about it
20:28 opal       the patch compiles and i have tested this version for a good couple hours yesterday
20:28 opal       no crashes
20:31 opal       note that i used glibc's 8mb stack size because i haven't profiled minetest's stack usage
20:32 opal       nor do i trust myself to make a good decision on that
20:43 paramat    well it's due to the large amount of issues and people i deal with each day, no-one can expect devs to remember trivia of one person :)
20:53 Vexyl      You could, you know, make a github account.
20:57 rubenwardy <opal> and before someone says "make a github account", no, i dont agree to their terms of service
20:57 opal       thanks rubenwardy
20:57 rubenwardy wow, word for word
20:57 opal       lmao
20:58 opal       there's also the secondary and tertiary matters of: github itself does not release its software in the foss realm, and github exploits a decentralised protocol (git) by accompanying centralised features around it
20:58 opal       and then people believe github == git after a while, and people who dont use github are ostracised by people like Vexyl
21:00 paramat    those seem non-issues
21:00 opal       they're issues to me
21:00 opal       the ToS issue is enough to back my decision against github though
21:00 opal       the others you dont have to agree with
21:00 paramat    ok
21:00 opal       :p
21:01 paramat    #7267 is ready
21:02 ShadowBot  https://github.com/minetest/minetest/issues/7267 -- Dungeons: Mostly fix missing stair nodes by paramat
21:12 nerzhul    opal, then propose the PR on gitlab if you are so constraint
21:12 nerzhul    there is a mirror there, but it's readonly
21:12 opal       i selfhost git
21:14 nerzhul    but if you want to contribute you should use standard tools used by the maintainers. there is a CI it's not for nothing
21:14 opal       furthermore im not uploading a 50mb forked repo anywhere on a limited LTE connection
21:14 opal       i made a patch file, please use it
21:14 opal       git is a standard tool
21:15 opal       git was made to produce and accept patches from various locations (email, plaintext) easily
21:15 nerzhul    and github is the standard tool for MT contribution. Don't be so closed, it's not a debate :)
21:15 opal       so
21:15 opal       you're going to deny a fix just because of stupid politics?
21:16 opal       i created a formatted patch that someone has to just `git am` on and it would be done, but you decide to argue on irc about it?
21:16 opal       now how backward is that?
21:16 nerzhul    no, i just said your opinion about our tool is just bad for everybody who just want your fix :)
21:16 opal       and who are you talking to about "closed"
21:16 opal       you are closed about someone who takes time out of whatever theyre doing to contribute to your project, even if they dont have the exact same setup as you do
21:17 nerzhul    i like the contribution, especially when they pass in our CI which is our quality gate for a merge
21:17 opal       i was banned from github. happy? creating a new github account would be against TOS on there
21:18 opal       the hell is a CI
21:18 opal       i have gone this far in my programming / development life without needing such a thing
21:19 rubenwardy CI = continuous integration
21:19 Vexyl      can't someone pull from your git repo and push to origin? /shrug, I'm still new to git.
21:19 rubenwardy it's a web server which compiles our code on most supported platforms to make sure it works
21:19 rubenwardy well, make sure it compiles
21:20 opal       Vexyl: they could `git am` it just the same and it would work fine
21:20 nerzhul    maybe you are a genious, but genious guys should not be so rude and our CI define the quality. With the debian patch, how can i be sure it builds correctly on the 10 tasks we have on the CI ? i should do it myself ?
21:20 nerzhul    musl libc is good for docker, else useless :p
21:21 opal       nerzhul: *i'm* rude?
21:21 opal       you know what, who are you in relation to minetest?
21:21 paramat    i appreciate the patch, if it's useful we might apply it ourselves
21:21 nerzhul    i think it's too late for me to have a debate with a debian integrist. Libre integrist are a cancer for the whole Libre. Libre has win, stop being integrist
21:22 opal       what?
21:22 nerzhul    opal: i'm a coredev
21:22 opal       then make sense?
21:22 nerzhul    but if you are banned from GH you cannot see the team, anyway the team has IRC grants, then you can see who is who :p
21:22 opal       glibc isnt the fucking standard, POSIX is, and POSIX does not define an 8mb stack size
21:22 nerzhul    MT runs fine on BSD libc
21:22 opal       im not using BSD now am i
21:23 opal       glibc is still not the standard
21:23 nerzhul    maybe, but the code works well on BSD libc, which is not fucking glibc
21:23 nerzhul    and it works in clang libc too
21:23 opal       clang doesnt have its own libc
21:23 nerzhul    exact, just libc++ :p
21:24 opal       CI sounds like some github lock-in bullshit
21:24 opal       if you find it useful, ok
21:25 opal       but dont expect people to go out of their way to sign right up to github just to pat your back with the CI thing
21:25 opal       that's an unprofessional and unrealistic expectation
21:26 nerzhul    opal, then compile on minw32, mingw64, macosx, clang 3.6, clang 5.0, gcc 6, gcc 4.9 and give us the results if you are a genious :)
21:27 opal       well i got gcc 6.4.0 down, i can do clang 5.0.1
21:27 nerzhul    we are not paid for this work, and CI permits to win some review time. Many libre project sshould have not failed if correctly tested or with a CI... don't expect a project to be honest without quality gates
21:27 opal       but dont tell me that github has *no* method of going through this process with an external commit?
21:27 opal       if it's impossible then perhaps it's a flaw in the CI design
21:27 nerzhul    travis requires a github repository integration
21:27 opal       my issue isnt with integrity testing, i get that
21:28 nerzhul    you can also fork the project on gitlab.com and test it on the gitlab CI direclty if you want
21:28 opal       my issue is "how can your integrity testing not accept a commit from an outside source"
21:28 nerzhul    i don't see the point with the stack size here
21:28 rubenwardy we could create a PR on your behalf for that
21:28 rubenwardy it's a non-issue
21:28 opal       why couldnt someone have suggested this earlier before i had to get into a fight
21:28 opal       thanks rubenwardy
21:29 nerzhul    haha
21:29 nerzhul    i don't want to do it, and if you are not on the PR discussion, how can we easily find the discussion and you to battle about it ?
21:29 rubenwardy I don't understand the stack size issue however, so will have to read up on that
21:29 Vexyl      If you self-host git why can't one of the coredevs pull from you and then push to origin? Wouldn't that go through CI?
21:30 opal       pulling the git repo once (50+mb gone from my monthly data limit) is enough for me, making a <1kb patch and uploading it; i'm not gonna reupload 50+mb just for a <1kb change
21:30 opal       i mean, if you want to pay for me to have high-speed cable here, then please by all means
21:30 opal       but i have my reasons too
21:30 nerzhul    do want you want it's not our problem :p
21:30 opal       lol
21:30 Vexyl      It wouldn't only pull the changes? :(
21:31 nerzhul    don't forget you just need a full sync only once on git :)
21:31 opal       Vexyl: if i'm forking a repo, i have to re-upload the full repo
21:31 Vexyl      Oh, I figured you cloned the repo already
21:31 opal       nerzhul: i take your comment against musl as a practical libc personally
21:31 opal       Vexyl: cloning is downloading, not uploading
21:34 opal       nerzhul: fwiw i can run seamonkey, chromium, gimp, and quite a few other programs just fine under a musl environment. i had issues in the past with glibc's unicode rendering under terminal so i had to switch
21:34 opal       i'm not going to use something that pushes broken versions when there's something non-broken available
21:36 nerzhul    no problem for me, i use alpine image on kubernetes in many softs and it works just fine, but not all