Time  Nick       Message
06:38 erlehmann  i really don't get why you don't add in the previous support and then discuss the replacement https://github.com/minetest/minetest/issues/11611
06:39 erlehmann  nearly every project i know that ripped out stuff and THEN discusses replacement did never get around to providing a sufficient replacement
06:42 erlehmann  you can see this with systemd btw
06:43 erlehmann  rubenwardy, while you can still have pixel fonts with ttf, it is quite difficult to make sure they are crisp on every configuration
06:45 erlehmann  rubenwardy in fact, whenever i switch resolutions (i.e. get a new display), i have to figure out which combination of resolution and fontsize make it so that no antialiasing is happening. i have not yet found a way to automate that – disabling antialiasing at the wrong combination of resolution & fontsize will just result in an ugly mess.
06:45 erlehmann  (i use GNU unifont in almost all applications)
06:47 erlehmann  i am trying to get an ancient github account back
06:47 erlehmann  i hate github, it is among the worst platforms for this (it was one of the best once, but ie and chrome were once the best browsers, amigaOS was once the best OS, you get it)
06:57 erlehmann  got it
07:10 MTDiscord  <Sublayer plank> what do bitmap fonts do that ttf fonts can't do, practically speaking?
07:21 erlehmann  Sublayer plank from a user side: guaranteed crisp font display with predictable character widths (not necessarily monospace). from an artist/developer side: extremely simple to modify, as they often are loaded by a program in a form in which they can be edited with a text editor and/or a program as simple as bitmap(1) or ms paint.
07:22 erlehmann  Sublayer plank then there are the performance benefits of not having so much bloat, but i am not sure if the slowness that Desour noticed and i reproduced has anything to do with it.
07:23 erlehmann  disclaimer: i have made glyphs for a pixel font, modified a pixel font, implemented naive rendering code for a pixel font in some other project, and tried to generate TTFs myself and failed at it
07:24 erlehmann  but i am by no means an expert – so if anyone else has done similar or ideally more work in that space, i'd love to know their opinion on it
07:27 erlehmann  Sublayer plank from a rendering perspective, ttf provides a lot more features, but all of those make editing more cumbersome and can make rendering less predictable. for example, true type hinting distorts the glyphs so that the rasterizer produces less ugly output. this is very user-friendly, but also requires a lot of developer work.
07:29 erlehmann  Sublayer plank maybe it is easier to consider these things if you approach the problem from a “does this make it more likely that people will convert or modify or create a font for use in minetest” point of view
07:29 erlehmann  i.e. a custom format that *did not* exist before makes it less likely than one that did exist before, a more complex format makes it dependent on the tooling if people do it
07:29 erlehmann  granted, there is no lack of fonts right now
07:30 erlehmann  but that's how i approach it
07:30 erlehmann  i shall add that thought to the issue
07:44 MTDiscord  <Sublayer plank> well I can agree to the fact bitmap fonts are easier to make, I made a bitmap number font for a löve game like two days ago
07:46 erlehmann  Sublayer plank and would you have any idea how to create a TTF from it without help?
07:47 erlehmann  (by which i mean a working version, just saying “i'd probably use fontforge” doesn't cut it)
07:48 erlehmann  i mean … i'd probably use fontforge. ;)
07:48 erlehmann  (but i have no idea what to do after starting it hehe)
07:50 erlehmann  btw, part of the discussion reminds me of the “but you can also use BMP instead of TGA” argument, which was probably only made by ppl who never had a real need to choose between implementing writers for those formats.
07:50 erlehmann  i.e. it makes sense until you get into the specifics
07:50 MTDiscord  <Sublayer plank> I'd have absolutely no idea how to turn it into a TTF, other than "something something fontforge" lol
07:51 erlehmann  see
07:51 erlehmann  and you and me are devs, we know probably more about it than some random kid with paint 3d
07:52 erlehmann  (yes, texture packs and the like are made by random kids. i know of mods created in windows notepad using just the rubenwardy book.)
07:53 erlehmann  (they are not necessarily on contentdb, ppl do it for their friends. a dev will probably never see any of that.)
08:50 erlehmann  i created an issue for the TGA thing https://github.com/minetest/minetest/issues/11613
08:50 erlehmann  (i found none and wanted to summarize the story so far)
08:59 sfan5      "Criticism [...] that they rely on TGA support was initially brushed aside due to lack of understanding by Minetest engine developers"
08:59 sfan5      always our faults eh
09:00 sfan5      how about you chalk that up to MCL devs not properly explaining that they need TGA loaded in existing worlds
09:05 erlehmann  sfan5 well it is true, they did not explain it, i can add that.
09:06 erlehmann  sfan5, i mean “no one knew how this was used” was true or not?
09:06 erlehmann  (on the engine side)
09:07 sfan5      do you mean "no one knew that anyone uses TGA" or "no one knew how MCL uses TGA"
09:07 sfan5      both are true anyway
09:07 erlehmann  sfan5, i added “(MineClone* developers did not explain that in detail)”
09:07 erlehmann  is that enough?
09:07 sfan5      sure
09:08 sfan5      not like that detail of the story matters, I just wanted to point out that blaming the coredevs is too easy
09:08 erlehmann  i mean it is obvious to anyone who has played mineclone2 or mineclone5 this year, but i already told you that some of them think minetest enigne devs are idiots, which mostly means that they do the minimal work in interacting
09:08 erlehmann  not sure if those who spoke up think that though
09:08 erlehmann  they would not admit it anyway ;)
09:08 erlehmann  they are hard to work with
09:10 erlehmann  sfan5 i think details matter for the future: minetest engine devs will probably not research stuff if someone yells “this breaks my use case” and does not explain it in-depth
09:10 erlehmann  do you agree with that?
09:10 erlehmann  i mean especially bc you were quick to point out there is an alternative
09:11 erlehmann  (even though BMP and PNG both do not solve that problem, there are others that could have been solved, maybe)
09:13 erlehmann  sfan5, is anything missing on that issue i should add? do you feel anything else is wrongly represented?
09:14 sfan5      no
09:15 erlehmann  btw, i do agree with the sentiment often repeated on the forums that core devs seem at times out of touch with the minetest game devs / mod devs / server admins. however, most of the “evidence” forum users present is “paramat or sfan5 shot down my totally stupid proposal that would have hilariously bad consequences for speed / compatibility / gameplay / development of mods”, so while i have my own r
09:15 erlehmann  easons, i think it makes total sense to not take most of that feedback seriously.
09:16 erlehmann  like, sending the server what platform is used, lol
09:16 erlehmann  what do they wanna do with it, make a mod display “this server is not for mobile users”?
09:27 erlehmann  i read https://github.com/minetest/minetest/security/policy – i was told to create issues for the security vulnerabilities i found (UB is always a security issue), so i will file them publicly
09:28 erlehmann  if i misunderstood this, tell me please
09:41 erlehmann  sfan5 celeron55 rubenwardy this is not a good issue, but is it a good start? https://github.com/minetest/minetest/issues/11614
12:04 pgimeno    why has the irrlichtmt repository not been renamed yet?
12:07 erlehmann  if you rename it, will it break links to existing issues there? will it break “git pull”?
12:07 erlehmann  i hope not
12:07 erlehmann  i verified the TGA thing as good as i could btw https://github.com/minetest/minetest/pull/11598
12:07 pgimeno    https://github.com/vrld/Quickie example of renamed repository
12:08 erlehmann  ah
12:08 erlehmann  1. can be cloned with original name
12:09 erlehmann  2. https://github.com/vrld/Quickie/issues/86 is redirected
12:09 erlehmann  3. https://github.com/vrld/Quickie/issues/86#issue-988269356 is redirected
12:10 erlehmann  i guess renaming is not a big problem then but why
12:10 erlehmann  do it
12:10 erlehmann  ?
12:11 erlehmann  what is the new name anyway
12:17 pgimeno    irrlichtmt is what the repo title says
12:18 pgimeno    when cloning inside minetest/lib, currently you have to be careful to specify irrlichtmt as the destination
12:19 pgimeno    if renamed, that would not be a problem; also that repository no longer contains Irrlicht, and that might be a misrepresentation of the brand
12:21 MTDiscord  <Sublayer plank> github has redirection in place for renamed repos, if you rename it to minetest/irrlichtmt there will be a redirect  irrlicht -> irrlichtmt if you pull with the old remote
14:59 sfan5      pgimeno: you had a PR to allow text output for /help again, right? where was that?
15:02 sfan5      erlehmann: well given it's inside irrlicht I think you filed that on the wrong repo
15:02 sfan5      (don't bother closing, we can transfer it)
15:03 erlehmann  sfan5 damn :/
15:03 erlehmann  sfan5 but is it ok to file this publicly? you literally told me right after that trace to file issues
15:05 MTDiscord  <Jordach> sfan5 re #11600 I'm going to need some guidance on manually building zstd from source for multi arch
15:05 ShadowBot  https://github.com/minetest/minetest/issues/11600 -- Possibility of an "early" 5.5 release
15:06 MTDiscord  <josiah_wi> Function move_for_backend in terminal_chat_console.cpp calls a move function on line 59. Someone on Discord has reported a case of the function being undeclared at compile time. Is this the curses function for moving the cursor, perhaps? Or what should I do to find out what's going on? I grepped for "* move(int*" in the whole project and didn't find anything, so I'm thinking it's a library function.
15:08 sfan5      @Jordach sure, what did you try so far?
15:08 MTDiscord  <Jordach> the more laughable question is it's only makefiles
15:08 MTDiscord  <Jordach> and they haven't documented the cmake lists
15:09 sfan5      @josiah_wi /usr/include/ncurses.h:extern NCURSES_EXPORT(int) move (int, int);				/* generated */
15:09 sfan5      I used cmake for zstd on windows and it worked fine
15:09 MTDiscord  <Jordach> sooner or later macOS builds from me will likely be signed entirely with my developer license
15:10 erlehmann  Jordach do you have a script on how to build thm
15:10 MTDiscord  <Jordach> i just need to build the static library as per usual
15:10 erlehmann  on macoss
15:10 erlehmann  like for brew
15:10 MTDiscord  <Jordach> brew is prebuilt
15:10 MTDiscord  <josiah_wi> sfan5, do you have any idea why ncurses.h would fail to be included?
15:11 MTDiscord  <Jordach> that said
15:11 sfan5      I haven't even looked at the .cpp file you mentioned, so no
15:11 MTDiscord  <Jordach> i'd argue it's not 5.5 we should be releasing
15:11 MTDiscord  <Jordach> but more 6.0
15:11 MTDiscord  <Jordach> we added a new major dependancy - chopped up irrlicht further
15:12 MTDiscord  <Jordach> inbetween releases we've done more or less a huge amount of work that somehow is larger than the previous 5.x and even 4.x releases
15:12 MTDiscord  <Jordach> i'm still working on the experimental hours of the day blending
15:12 sfan5      I get what you mean but the version numbers don't signify progress but compatibility
15:12 MTDiscord  <Jordach> while zstd is backwards compatible
15:13 MTDiscord  <Jordach> it won't be that way forever
15:13 MTDiscord  <Jordach> lest it becomes permanent and we become dependant on supporting ancient ass versions for no reason
15:13 erlehmann  look a major update implies a backwards incompatibility
15:13 MTDiscord  <Jordach> i prefer the term, slightly incompatible
15:13 erlehmann  Jordach are you one of the clowns who wants me to no longer run minetest?
15:13 MTDiscord  <Jordach> no
15:14 MTDiscord  <Jordach> fuck sakes i'm bending over backwards to ensure 2008 macs can still run MT with Core 2 Duos
15:14 erlehmann  well if hecktest get its wishes, soon minetest will not be able to run on whatever hecks makes up is “obsolete” and no longer used.
15:15 erlehmann  network protocol interop is the only thing that will let players play then
15:15 erlehmann  if they are on these machines
15:15 MTDiscord  <Jordach> i should mention that mavericks practically supports every intel mac
15:15 erlehmann  i have no idea of os x sorry
15:15 MTDiscord  <Jordach> since 2006 at minimum
15:16 erlehmann  i only know that new OS X turned some machines from 2013 from a specific run into paperweights
15:16 erlehmann  and also i had to use os x some while back … is it still that bad?
15:16 erlehmann  or has it improved, like linux desktop improved 15 years ago or so in short time?
15:16 MTDiscord  <Jordach> osx isn't bad if you've already used linux in some capacity
15:17 pgimeno    sfan5: https://codeberg.org/pgimeno/minetest/pulls/1
15:17 MTDiscord  <Jordach> just note that osx uses bsd'isms rather than linux'isms
15:18 erlehmann  nah i mean rather stuff like “can't easily tell a window to stay above others”. but this discussion does not belong here, how can i contact you otherwise?
15:18 MTDiscord  <Jordach> i'm not sure if the machine hosting my lounge instance is alive
15:19 erlehmann  pgimeno wow cool
15:21 sfan5      pgimeno: our CI has some complaints https://github.com/sfan5/minetest/runs/3588789880
15:21 sfan5      (I adopted the PR on Github because I remembered today that I wanted this feature)
15:22 sfan5      ah wait logged out users can't see CI results, that's awful from github
15:22 pgimeno    I'll try rebasing
15:22 sfan5      https://a.uguu.se/tyT0tNzfUpEU_.txt
15:23 pgimeno    oh my did I make that rookie mistake?
15:24 sfan5      if you rebase it squashing it all into one commit would also be helpful
15:29 pgimeno    sfan5: rebased and pushed
15:29 pgimeno    (and squashed)
15:30 sfan5      I don't see it
15:30 pgimeno    oh right, I pushed to notabug, not to codeberg, one sec
15:32 pgimeno    done
15:32 pgimeno    https://codeberg.org/pgimeno/minetest/src/branch/pg-help-text-option
15:34 erlehmann  what is codeberg?
15:34 pgimeno    a Gitea instance for free software, kinda like Savannah etc but with Gitea
15:35 pgimeno    same as notabug which I also used, but notabug has stayed on gogs and has not updated to gitea, which is more complete
15:36 erlehmann  nice
15:37 pgimeno    (gogs is itself an attempt to clone GitHub's interface, and Gitea is a fork because gogs progressed too slowly)
15:40 dzho       for what it's worth, github didn't invent web interfaces to git hosting
15:41 dzho       it might be more of a quibble to point out elsewhere but when talking about procedurally-generated block sandbox gaming it might land a little differently.
15:41 dzho       (which is to say, minecraft didn't invent the genre so minetest isn't just "an attempt to clone" it)
15:44 dzho       gitorious:gogs/gitea::infinimer:minetest
15:45 dzho       derp
15:45 dzho       infiniminer*
15:45 dzho       I lost track, a little like spelling "banana"
15:48 sfan5      gitorious, man that takes me back
15:49 dzho       ikr
15:49 dzho       I had been reminded of git.or.cz recently and gitorious was sort of the next step.
15:50 pgimeno    sfan5: thanks for adopting
15:50 sfan5      np
15:51 sfan5      the harder part might be waiting for two reviews
15:53 erlehmann  sfan5, what do you think about making the xml glyph data tsv, would it work? https://github.com/minetest/minetest/issues/11611#issuecomment-918002341
15:53 erlehmann  i mean the reader is simple
15:53 erlehmann  just decide if you want to linebreak on *both* \r and \n or only \r (i suggest both bc ppl who use notepad to edit game data)
15:53 sfan5      of course?
15:53 erlehmann  i am definitely willing to test it
15:54 erlehmann  i mean that way the xml is gone
15:54 erlehmann  but the easy to edit pixel font still there!
15:54 MTDiscord  <Jonathon> https://github.com/minetest/minetest/pull/11612 should negate your statement of speed anyways
15:55 erlehmann  i will test it!
16:01 erlehmann  celeron55 sfan5 who do i have to please to become core dev so i can approve stuff i test
16:02 erlehmann  i think i have shown i am much more anal about QA and security than most others here
16:02 erlehmann  so to say
16:03 pgimeno    sfan5: about the approvals, when you adopted it had you reviewed it? I mean, if you don't say explicitly that you haven't, since the PR is under your name, others may assume you've approved it. That's a bit of a shady area.
16:03 pgimeno    s/shady/grey/
16:04 sfan5      I would have added the label and a comment if that was the case
16:04 pgimeno    ok
16:15 pgimeno    as for #11600, I'd prefer if zstd was opt-in for old worlds rather than forced
16:15 ShadowBot  https://github.com/minetest/minetest/issues/11600 -- Possibility of an "early" 5.5 release
16:16 pgimeno    it's too easy to not be able to go back to an earlier version
16:17 MTDiscord  <Jonathon> hasnt every other map update being forced?
16:19 pgimeno    I guess you have a point, it feels somehow different for me
16:23 Pexin      one might claim that "it's how we've always done it before" is never a valid justification
16:23 Krock      map version bumps are quite rare though
16:23 erlehmann  yes
16:23 erlehmann  it is really obnoxious to test the same world with two clients and then realize you have to recreate it
16:23 erlehmann  testing with 5.4 is a thing after all
16:23 Krock      pgimeno: I raised this concern too, but it seems to have drown in the countless discussions
16:25 MTDiscord  <Jonathon> pexin: correct, im just pointing that out
16:25 MTDiscord  <Jonathon> rather than opt in, i would say for one version there should be a option to disable it for testing
16:26 MTDiscord  <Jonathon> because getting users to enable or disable things is annoying
16:29 pgimeno    I guess I can write a map converter, I've done these things before
16:30 erlehmann  > there should be a option to disable it for testing → yes please
16:31 erlehmann  i have sent users world files
16:31 erlehmann  to test stuff
16:31 erlehmann  while i am on 5.5-dev, they are on 5.4 or sometimes even 5.3
17:02 sfan5      well perhaps --recompress can get an option to compress it to the previous version
17:02 sfan5      but I am against adding an option
17:04 pgimeno    what's the current map version? 28 or 29?
17:09 Krock      pgimeno: serialization.h  #define SER_FMT_VER_HIGHEST_WRITE 29
17:09 pgimeno    thanks
17:11 Krock      although the difference of HIGHEST_READ and -WRITE is a mystery to me w
17:12 erlehmann  <sfan5> but I am against adding an option
17:12 erlehmann  why
17:12 erlehmann  it would make regression testing much easier
17:13 Krock      ehm.. --recompress already exists
17:13 Krock      and adding an enum for compression type lookup is a pretty simple thing
17:13 sfan5      map updates were always non-backwards compatible, if an option was added for every change like this it would be a terrible mess
17:14 erlehmann  github javascript makes my computer run 10 degrees hotter than it run normally
17:14 erlehmann  i know again why i hate bloated websites
17:16 specing    Stop using github (:
17:17 Krock      check whether there's enough thermal compound
17:17 pgimeno    erlehmann: may I PM you?
17:17 erlehmann  ok
17:18 erlehmann  pgimeno, do it
18:26 celeron55  any other coredev comments about erlehmann proposing himself as coredev?
18:29 Krock      without a single PR this seems to be coming out of the blue, honestly speaking
18:29 celeron55  or opinions, more like
18:29 Krock      yes I know about the early Minetest days, but I guess at east some things have changed since then
18:30 nrz        For me it's not sufficient
18:30 nrz        Many pr merged should be required
18:31 celeron55  well that's enough, it's then decided erlehmann has to gain trust as a contributor
18:40 Jordach    i would step up for it seeing as i'm crazy enough to care about macs, and by extension iOS
18:45 Jordach    however, my concerns still resolve around finding a valid way of getting the real minetest on both Apple App stores
18:45 Jordach    Alongside including JIT
18:46 MTDiscord  <luatic> I could probably handle the Lua side of things
18:47 MTDiscord  <luatic> Is there something like "MTG maintainer"? A couple of PRs is literally just waiting for a second approval.
18:51 erlehmann  Krock nrz what exactly do you expect me todo?
18:52 sfan5      the expectation is to have existing, relatively recent contribution history
18:52 Jordach    I'm working on cleaning up the sky right now
18:52 sfan5      @luatc we have a team of MTG devs that is a subset of the coredevs, yet nobody really cares anymore
18:52 Jordach    Everyones favourite weeb hasn't been about recently
18:52 sfan5      +i
18:53 erlehmann  sfan5 look, i can't just make features up
18:53 sfan5      hecks said he's busy with work, might be a few weeks
18:53 Jordach    I was told you can't make first person attachments
18:53 Jordach    well that was clearly bunk
18:53 erlehmann  also i think wuzzy has made like 16 PRs or so and i have seen how long those lie around
18:53 sfan5      erlehmann: thankfully other people have already made up features and posted them on the issue tracker
18:54 erlehmann  sfan5 yeah but if i test them what does it help if i can not approve them?
18:54 MTDiscord  <Warr1024> erlehmann, it may just require some mixture of vigilance and patience for the opportunity to come.
18:54 sfan5      erlehmann: no no I meant you could take open feature requests and implement them
18:54 MTDiscord  <Warr1024> You testing things makes it a lot easier for other people to approve them still
18:54 sfan5      as for testing, all testing helps even if you cannot formally approve
18:55 sfan5      for example I don't think I'll be testing the TGA thing because you said you did, so that makes the process easier for me
18:55 sfan5      v-rob: if you get a chance please look at the 5.5 release proposal and also the protocol bump issue (both marked 'Discussion')
18:56 MTDiscord  <Warr1024> You don't need a formal coredev position to have your reputation as a thorough and reliable tester to carry a lot of weight.
18:56 Jordach    the lazier the core dev the more efficient minetest PRs get merged
18:56 erlehmann  <sfan5> for example I don't think I'll be testing the TGA thing because you said you did, so that makes the process easier for me
18:57 erlehmann  sfan5 funnily enough i think you should also test it if you approve it, bc what else would the approval be than a stamp “it works”
18:57 erlehmann  i mean “it works and it looks ok on a source level”
18:58 erlehmann  and i don't think you should believe me if i say it works, bc i could have made a mistake!
18:58 MTDiscord  <Warr1024> 2 approvals really just means that 2 people trusted by the project are confident that the thing is correct enough ... there's no rule that says coredevs can't delegate some of that trust though.
18:58 celeron55  i don't think you need to make PRs, but you do need to show your ability and responsibility
18:59 MTDiscord  <Warr1024> More testing is better, but there isn't exactly some particular obvious threshold where it becomes "enough" so we kind of just have to pick an arbitrary one or else nothing would ever get done.
18:59 Pexin      how to define a good gauge of familiarity with the current state of the codebase
18:59 celeron55  (that's only my opinion though, if others agree that code contributions are 100% essential, i will go with that)
18:59 erlehmann  in the light of the fact that wuzzy has a lot of PRs open i am kinda wary of making more PRs
19:00 Krock      @Warr1024 it's a matter of who might have the broadest knowledge about the code to catch mistakes. we already had this discussion about the influence of contributor reviews and they're a good gauge to ease the review process
19:00 celeron55  eh
19:00 erlehmann  there clearly is a need of more review than more PRs?
19:00 celeron55  look at the closed pull request list rather than the open one
19:00 erlehmann  ah ok
19:00 celeron55  it's a lot longer
19:00 celeron55  no need to get overly pessimistic
19:01 MTDiscord  <Warr1024> Knowledge of the code is a double-edged sword when it comes to catching mistakes, and I'd never really feel quite comfortable that something is "well-tested" unless it's tested by both people familiar AND unfamiliar with the code.  Knowing the code too well leads to biases and assumptions.
19:01 erlehmann  i see
19:01 erlehmann  i agree with Warr1024
19:01 erlehmann  knowing stuff too well leads to people never trying out things
19:02 Krock      I don't see any big influence there, but it might be possible, yes.
19:02 erlehmann  i mean it is not surprising to me that all the dupes i have been told except two were found by ppl having no idea of why the engine behaves that way in the first place
19:03 erlehmann  i think if i want to contribute it would be CI work
19:03 MTDiscord  <Warr1024> Also it's easier to put real and consistent effort toward testing a thing when you're not already burned out from dealing with the code for it.  :-D
19:03 erlehmann  making --random-input predictable
19:03 erlehmann  Warr1024 that is true
19:03 Krock      @Warr1024 that's basically me right now
19:04 erlehmann  Krock nrz i am sorry i am *way* better poking holes in things than fixing them
19:04 Krock      erlehmann: I'm used to that. It just won't change anything. What's needed are proposals. - best in form of code to fix it for once and all
19:05 erlehmann  but i can give you some new holes, let me open issues
19:05 nrz        erlehmann, no issue, as a contributor it's fine, like wuzzy does mainly, and after that he found time to contribute a bit with PR
19:05 erlehmann  give me a few minutes to find a new bug or something to show my willingness to contribute!
19:05 erlehmann  :D
19:07 rubenwardy MTG devs don't need to be a subset, they can be distinct
19:08 sfan5      (that just the current situation)
19:08 sfan5      but not like we need more devs to occasionally merge a fix and reject features
19:10 MTDiscord  <luatic> I'd just like to give a 2nd approval to pretty much all one approval thingies of MTG after having had a thorough second look at them
19:12 sfan5      hm
19:13 MTDiscord  <Warr1024> It seems like if you just start functioning as a core dev in every way other than formal status, i.e. you give productive, useful feedback in reviews, do reliable and thorough testing, and your comments are generally well-formed an insightful, then you'll already be a core dev in all but name and it becomes very easy for someone to just close that gap and make it official :-)
19:13 MTDiscord  <luatic> Except game#2567 - that still needs the rebase
19:13 ShadowBot  https://github.com/minetest/minetest_game/issues/2567 -- Lower frequency of standing/sitting animation by An0n3m0us
19:13 MTDiscord  <Jonathon> sounds like lmd is looking to infect mtg with modlib on a more serious note, 2 approvals are still needed for mtg, adding lmd to mtg wouldn't be a bad idea since its essentially abandoned(aside from the brief attention because of core dev meeting)
19:14 MTDiscord  <luatic> I can't add modlib all alone ;)
19:14 MTDiscord  <luatic> I also won't, considering it has plenty of features MTG doesn't need
19:14 MTDiscord  <luatic> If I was to infect something with modlib, it would probably be the engine
19:14 MTDiscord  <luatic> Cue builtinlib
19:17 MTDiscord  <Warr1024> I think one rather important feature for a coredev is being at least willing/able to use the communication media that the project officially uses, such as IRC and GitHub.  If you, say, have serious objections to using GitHub, then that would be a serious obstacle to coredevship, since it's an important central communication medium for the project.
19:19 sfan5      isn't that a bit ironic to say over a discord bridge
19:20 Krock      haaha
19:20 MTDiscord  <Jonathon> i mean, it is using irc, just abstracted
19:20 erlehmann  github sucks
19:20 MTDiscord  <Jonathon> more obvious than using say matrix to communicate to irc
19:20 erlehmann  i am only using github because i feel ppl won't read it if i don't use their facebook for coders
19:20 * appguru  proceeds to show good will by joining on IRC
19:20 Pexin      ponders cli full github frontend
19:21 appguru    "Facebook for coders" - lel
19:21 appguru    sadly kind of true
19:21 v-rob      I agree with Warr1024 that the best way to judge core dev material is if people are already doing the sorts of duties a core dev would and are responsible. That's basically how I became one -- make a bunch of useful PRs and do reviews. That's also why people like Wuzzy have been proposed (although he refused), and that's kinda what I think the criteria should be.
19:21 MTDiscord  <Warr1024> I am not unwilling or unable to use IRC; I am, in fact, actually lurking on the network.  I'm just not in the channel directly right now because dealing with self-pings is a PITA, as is juggling apps, and I'm already responsible as a moderator for Discord so I might as well be here.
19:21 erlehmann  appguru especially with the thumbs-upp-button next to posts it is hilarious
19:21 erlehmann  like there are some projects that get mentioned on orange website and then hordes of ppl come and give thumbs up or down
19:22 sfan5      I wasn't implying that, just for clarity
19:22 MTDiscord  <Warr1024> If MTE were to institute a policy like "all core dev meetings must be NOT through some unreliable bridge bot" then it wouldn't really be a problem for me to hop over for that.  I'd just hate to have to mess with complex notification/mute rules on a day to day basis.
19:22 sfan5      question: do you participate here from your phone?
19:22 v-rob      Eh, I can understand being on Discord. IRC's cranky (or maybe it's my client, I dunno), but it does funky things to me from time to time.
19:23 appguru    The Discord relay being online 24/7 also handles the bouncer hassle for you.
19:23 appguru    I have considered importing the logs found on the website to Thunderbird through some script, but I'm too lazy
19:24 MTDiscord  <Warr1024> IRC and mobile still mostly suck for each other.  I use TheLounge client, which makes IRC on mobile a lot better ... but unfortunately it's kinda crap when it comes to customizability, and especially not being able to customize highlights/pings on a per-channel or per-sender basis is a big pain.  If I lurk in the channel and ever also type via Discord, then the discord bot's messages will highlight and ping me on the IRC side.
19:25 MTDiscord  <GreenXenith> The relay is supposed to insert a special character somewhere to prevent self-mentions but as usual, the bridge doesnt like to work
19:25 MTDiscord  <Warr1024> TheLounge at least does handle the "bouncer" case for you (you can self-host an always-on client on your own infrastructure) and the multi-device sync thing, so e.g. shared experience with mobile and desktop.
19:26 MTDiscord  <GreenXenith> actually it may have gotten disabled, hold on
19:26 MTDiscord  <GreenXenith> Ah nope, its on
19:27 appguru    The problem is: I don't have my own infrastructure
19:27 MTDiscord  <Warr1024> Yeah, that does kinda suck.
19:27 Pexin      I just run irssi/screen on my fileserver at home
19:28 appguru    Who hosts https://irc.minetest.net BTW? Downloading all logs wouldn't kill it, right?
19:28 MTDiscord  <Warr1024> People don't like the discord thing because bridges are unreliable, and because not free software.  With a matrix bride, though, you could at least only have to worry about one of those.
19:28 sfan5      me but please don't
19:29 appguru    sfan5: then how can I obtain the full logs?
19:29 sfan5      well why do you want that
19:29 MTDiscord  <Warr1024> I mean, you should be able to bulk-download as long as you use a gentle enough speed limit...
19:29 erlehmann  if there was only an established protocol where clients could transparently sync with a multi user chat on rejoining that had support from a lot of clients for years! (like xmpp)
19:30 MTDiscord  <Warr1024> Haha, yeah, I loved XMPP in its heyday, even though it never quite got its act together for mobile (it got pretty close though).  Sadly it's all xkcd 927 all the way down.
19:30 appguru    sfan5: dunno, it just seems kinda convenient to always have full context available
19:31 appguru    and as it's pretty much plaintext, it shouldn't be all that large
19:31 MTDiscord  <Warr1024> FYI there is a literal plaintext version of each capture, I think, i.e. without any additional markup.
19:31 appguru    Yeah there is
19:31 erlehmann  Warr1024 “in its heyday” wdym, it is not only still used, it works better than before.
19:31 sfan5      it's rare that I need anything older than a week and if so I could also use google
19:32 sfan5      though being able to grep .local/share/hexchat/logs is pretty convenient
19:33 appguru    Well, how large is it? Could you just provide a full .tar for download?
19:33 erlehmann  Warr1024 250 million fortnite players use xmpp, that's probably … a lot
19:33 erlehmann  if the chatlogs were in a git repo, you could just do git pull hmm
19:34 sfan5      mysql apparently uses <400MB for that table
19:34 sfan5      if you want to download all -dev logs I won't stop you but a rate limit will slow you down
19:43 erlehmann  who decided the actual values for --random-input? the digging time and the mouse movement does not seem very good for testing in my opinion.
19:44 erlehmann  however, if you set digging to 200 or 300 and divide mouse movement by 10 it kinda works
19:59 sfan5      does anyone want to review #11605 or should I merge it as trivial
19:59 ShadowBot  https://github.com/minetest/minetest/issues/11605 -- Fix GLES2 discard behaviour (texture transparency) by sfan5
19:59 sfan5      maybe I should be less impatient, it's not an important fix
20:03 Krock      if it's tested, feel free to merge. I don't see how it could cause issues on any other render
20:05 erlehmann  sfan5 this looks like it affects rendering on the MNT reform? https://github.com/minetest/minetest/pull/11605/files
20:05 sfan5      yes
20:05 erlehmann  so is the bug also on the MNT reform?
20:06 sfan5      I guess
20:07 erlehmann  ok i told mntmn about it
20:08 erlehmann  i think he fixed that originally
20:09 sfan5      it's not like you need to test it specifically on there, it affects normal GLES rendering and can be tested on any PC (which I did)
20:09 erlehmann  well the GPU on the reform had some discard bug
20:09 erlehmann  i think
20:10 erlehmann  but from the comment it seems to be the same
21:22 erlehmann  <erlehmann> mntmn, can you test if it breaks your previous fix?
21:22 erlehmann  <mntmn> erlehmann: not currently
21:22 erlehmann  <mntmn> next week
21:23 sfan5      there is no need to test that, it doesn't
22:14 erlehmann  why was this merged https://github.com/minetest/irrlicht/pull/64 but this https://github.com/minetest/minetest/pull/11598 rot yet?
22:15 sfan5      do you see it having two approvals
22:15 erlehmann  oh i see, irrlichtmt does not need that
22:15 sfan5      now if you were to correctly point out that bugfixes can be merged without two approvals the answer is: I haven't bothered
22:18 erlehmann  well it seems to me anything can be committed to master without two approvals if the stars align. the TGA removal for example does not look like there was even a PR regarding it https://github.com/minetest/minetest/commit/1e2b6388818fec0d4cdc52f796850bfb7ec3a22e
22:19 erlehmann  i really don't understand those rules (and i prefer projects where no one ever can commit to master, but let's not go there)
22:21 sfan5      pgimeno: I looked at your code earlier and the only two comments I have are very minor: 1) I wouldn't use `opts[2] ~= nil` to check for presence but rather compare `#opts` 2) the `local cmd_marker` and the following if can be merged into an inline condition
22:27 pgimeno    sfan5: huh, that should be args instead of opts now that I look again
22:30 Desour     wouldn't it be easier to use a {opt = true} like set for opts instead of a string?
22:31 sfan5      I considered too but opts:find("x") isn't that hard
22:32 pgimeno    it would be faster, but it's not really performance-critical and that's sort of clean
22:32 pgimeno    I don't care either way
22:32 pgimeno    I had in mind the getopt options string where you pass all options in a single string
22:33 Desour     if someone wants to extend the getopts function at some time to do "--bar=foo" stuff, it could also be more easily changed to that
22:34 pgimeno    hm, it would better be extended with another argument to not mix -c and --c
22:36 pgimeno    anyway it's local to the file, I had reasons to do that when I wrote it but I no longer remember, but it means it can't be used for anything but the functions in this file
22:36 pgimeno    which are... do_help_command and nothing else :)
22:37 pgimeno    if it's ok as is, the review changes are pushed, otherwise it will take longer
22:40 pgimeno    I think it was that the options are unlikely to be useful to any other command, and if the time comes, it can be moved around/published/whatever