Time  Nick             Message
05:34 MTDiscord        <fatalerror420> False
05:37 MTDiscord        <GreenXenith> ^this is in reference to mazes "did I observe this right?". @The lesser shitposter used the Discord reply feature which is not bridged
05:38 MTDiscord        <fatalerror420> Hehe
05:46 erlehmann        is the “let's store all the media cache in a sqlite database” thing dead or do i need to convince anyone that it's not a good idea for the purpose of cache cleanup?
05:46 erlehmann        #11567
05:46 ShadowBot        https://github.com/minetest/minetest/issues/11567 -- Use an sqlite database for the media cache by Desour
05:51 erlehmann        people probably thought that relatime = atime is useless
05:51 erlehmann        but that is not true!
05:53 erlehmann        the mainline linux kernel will update atime with relatime with a 24h granularity
05:53 erlehmann        it is hardcoded, you can see here https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/fs/inode.c#n1719
05:53 erlehmann        “Is the previous atime value older than a day? If yes, update atime”
05:54 erlehmann        and “With relative atime, only update atime if the previous atime is earlier than either the ctime or mtime or if at least a day has passed since the last atime update.”
05:54 erlehmann        this is hardcoded and unlikely to be changed
05:58 erlehmann        #11499
05:58 ShadowBot        https://github.com/minetest/minetest/issues/11499 -- [Draft] Night vision by x2048
05:59 erlehmann        so here hecktest says “I do have plans for dropping support for the fixed pipeline and it's likely to happen, the set of machines that can run Minetest comfortably while not supporting at least OpenGL 2 is basically empty.” https://github.com/minetest/minetest/pull/11499#issuecomment-890270979
05:59 erlehmann        ; glxinfo |grep 'OpenGL version'
05:59 erlehmann        OpenGL version string: 1.4 Mesa 18.3.6
06:00 erlehmann        does that mean i will not be able to run minetest?
06:00 erlehmann        of am i misunderstanding something here?
06:01 erlehmann        this is on the computer on which i do most of my private development
06:01 erlehmann        which includes minetest
06:02 erlehmann        mods
06:10 erlehmann        i get around 25 fps in mineclonia with minetest 5.4
06:11 erlehmann        in the title bar it says OpenGL 1.4
06:11 erlehmann        i am really anxious about someone who probably has a powerful gaming computer to make judgements like that
06:12 erlehmann        people with supercomputers should really ask people who do not have supercomputers if their performance assessments are accurate
06:12 erlehmann        or at least test it!
08:05 celeron55        erlehmann: exactly what hardware do you have?
08:05 celeron55        and which OS
08:09 celeron55        you have to understand even 10 years ago when i started MT it was considered ridiculous to support opengl 1 in a game
08:13 celeron55        if you want a really quick test to see whether you can run opengl 2 in practice, try running some webgl stuff in a modern web browser, like https://webglsamples.org/multiple-views/multiple-views.html
08:16 erlehmann        celeron55, i have several thinkpads with debian
08:16 erlehmann        celeron55 i am writing you from a thinkpad T60
08:17 erlehmann        i am trying the thing you showed me in chromium now
08:18 erlehmann        > It does not appear your computer supports WebGL. Click here for more information.
08:18 erlehmann        wow websites now have OpenGL requirement? o.0
08:18 celeron55        core duo? it's the same hardware i used 10 years ago
08:18 celeron55        it will be dropped, i am sad to say
08:18 erlehmann        what no
08:18 erlehmann        why
08:19 erlehmann        i can help maintaining it
08:19 erlehmann        i know a bit of C++
08:19 celeron55        well, you'll have to write a fixed pipeline rendering backend when the time comes
08:20 erlehmann        i do not see why not just keep the old code? is it such a burden?
08:20 celeron55        core duo is literally the only cpu that is barely usable these days but with no support for opengl 2 in the integrated gpu
08:20 celeron55        or, well, the gpu is in the chipset i think (?) but whatever
08:21 erlehmann        i think it has nothing to do with core duo, more with intel integrated graphics yes
08:21 celeron55        i have 7 years old intel integrated graphics in this laptop, and i have opengl 3
08:21 erlehmann        so i do not get it, why would support need to be removed? the code works right now!
08:22 erlehmann        i was using a similar computer, probably even weaker, when i first discovered minetest
08:22 erlehmann        when i develop stuff for mineclonia i do it on this machine
08:22 celeron55        to get better performance on practically all hardware available today (used and new), the rendering has to operate differently, not just how it calls opengl but it needs to construct its data differently in order to be able to call opengl in a more modern way
08:23 celeron55        you can use opengl 1.4 then too of course
08:23 erlehmann        but?
08:23 celeron55        but it won't be any faster on modern hardware then
08:23 erlehmann        i just want to continue to play minetest
08:24 celeron55        then your goal is to get an opengl 1.4 backend written along with the 2.0 backend that is planned
08:24 celeron55        there's also a 3.0 or newer backend planned for the actual gaming computers
08:24 erlehmann        from a very uneducated point of view the rendering of minetest sucks big time not because of a lack of shader magick
08:25 erlehmann        but because ppl do stuff like “implement particle support in a way that you can not make worse even if you tried” and something like that
08:25 celeron55        your gma950 gpu can render stuff just about as fast as the cpu can send stuff to it, no matter how ineffectively it's send
08:25 erlehmann        celeron55, is openGL ES an out for me? https://www.jwz.org/blog/2012/06/i-have-ported-xscreensaver-to-the-iphone/
08:25 erlehmann        yes ok?
08:25 celeron55        go forward a year or two in gpu development and the cpu can't keep up anymore without being very specific in how it sends stuff
08:25 celeron55        it has to send it in bulk
08:25 erlehmann        is this a lacture about shaders
08:26 celeron55        webgl uses opengl es 2.0
08:26 celeron55        you just attempted to use it on the website
08:26 celeron55        that i linked
08:26 erlehmann        oh ok
08:26 erlehmann        oh but glxinfo gives me 2 more version strings!
08:26 erlehmann        OpenGL ES profile version string: OpenGL ES 2.0 Mesa 18.3.6
08:26 erlehmann        OpenGL ES profile shading language version string: OpenGL ES GLSL ES 1.0.16
08:26 erlehmann        does this mean i can use minetest still
08:27 celeron55        not sure, we'll have to consult with hecks
08:27 erlehmann        i am really really anxious about this, minetest is about the only game i play a lot :/
08:28 erlehmann        and i think i am somewhat an exception in that ppl with old or bad computers usually are not developers. but the T60 is the last computer i found with a 4:3 matte screen that i fast for almost everything (once you put an SSD into it lol)
08:28 celeron55        i understand that, but the thing is you could get more recent hardware for almost no money
08:28 celeron55        definitely less money than a new battery costs for the t60
08:28 celeron55        than what*
08:28 erlehmann        i got an MNT reform 2 recently
08:28 erlehmann        like days ago
08:29 erlehmann        > you could get more recent hardware for almost no money
08:29 erlehmann        that actually was true until about february/march
08:29 erlehmann        then the prices for used thinkpads went through the roof
08:30 erlehmann        i have a bunch of machines for spare parts. guess what, they all have either that same chipset from intel or some jank that runs way too hot.
08:30 celeron55        the mnt reform will run opengl 2 very nicely
08:31 erlehmann        is the jwz post relevant?
08:31 erlehmann        i mean jwz is a very smart guy
08:32 erlehmann        and implemented OpenGL 1.3 in terms of OpenGL ES 1.1?
08:32 celeron55        i don't think it has any relevance
08:32 erlehmann        but you need openGL ES for mobile anyway or not?
08:33 celeron55        MT will be targeting opengl es 2.0 on mobile, not 1.1, because all phones that have the power to run MT have 2.0
08:33 erlehmann        i thought the whole “the thing worked for 10 years but you need to get a better computer now” thing was purely a fleckenstein/mineclone2 thing
08:33 erlehmann        guess i thought wrong
08:34 celeron55        anwyay make sure to talk with hecks and tell him to at least make sure it's possible to write an opengl 1.4 backend, or i guess we could call it a gma950 backend
08:35 erlehmann        what exactly is hecktest doing btw?
08:35 erlehmann        can i see some roadmap?
08:35 erlehmann        let's say i'm *probably* not the only person affected
08:35 erlehmann        i encounter it often that ppl who play games on not-too-recent computers just aren't devs
08:35 erlehmann        devs have the newest and greatest
08:36 erlehmann        there is, for example, this thing with minetest games and minetest 5.3 compat – all because minetest 5.4 is harder to get to run on raspberry pies.
08:36 celeron55        well, i stopped trying to run games on my core duo laptop in the year 2014, and thinking about the transition away from those as a normal distribution, we're going quite far to the "almost nobody" territory today
08:37 erlehmann        oh yeah, ppl often say that. in the meantime, emacs has not broken for me once. or dosbox. or openttd.
08:37 erlehmann        or battle for wesnoth, for that matter.
08:37 erlehmann        (the rpi story is not my personal experience)
08:38 celeron55        those are 2d programs, 2d rendering hasn't changed much... or well, if you take modern 2d games (commercial ones you can buy on steam or such), the fact is almost all of them require opengl 2 or 3
08:38 erlehmann        ok well but openarena is basically quake3 and it runs?
08:39 celeron55        it makes the logistics of rendering so much nicer when you can move data around without transformations or effects on the cpu and do the final transformations and effects on the gpu
08:39 erlehmann        oh yeah, ofc that's what its all about, i get that
08:39 celeron55        quake was literally made for opengl 1
08:39 erlehmann        it is about making it easier for developers, not users
08:40 celeron55        (if you take the new improved re-release of quake that was made very recently, it won't run on opengl 1)
08:40 erlehmann        let's say i often take care to make my software run in a way that it runs on the machines of ppl who complain, not on what optimizes for theoretical purity
08:41 celeron55        many, many users of modern hardware complain that MT isn't optimized for them
08:41 erlehmann        oh yeah, but they probably have no idea why it sucks
08:41 erlehmann        like the particle lagfest
08:41 celeron55        i'm thinking this mostly in terms of world rendering range
08:42 erlehmann        waspsaliva, the probably most advanced cheat cient (seriously, try it) has a noparticles setting ^^
08:42 erlehmann        ah ok
08:42 erlehmann        well everything lags if i put rendering range to infinite. on my thinkpad T60 i get like 13 fps or so.
08:42 erlehmann        probably less if i load a lot of the world
08:43 celeron55        yeah, it just can't fill the polygons faster than that no matter what
08:43 erlehmann        nah, it also lags the same way on faster machines lol
08:43 erlehmann        >  The code is in hacks/glx/jwzgles.c. If you want to use it to port your own legacy code, just include jwzgles.h. Let me know if it works!
08:43 erlehmann        lol
08:43 celeron55        that's because MT sends the stuff to the gpu so inoptimally
08:43 erlehmann        i have to look at that
08:44 erlehmann        but why have not different rendering backends?
08:44 celeron55        somebody has to write and maintain them, and you can create a lot of extra work when some backend starts requiring preprocessing of data that others don't need
08:45 erlehmann        btw, core duo … i think on a single core machine it ran about the same. is minetest using more than one core?
08:46 celeron55        in a singleplayer game it can use 3 cores to some extent, one for the client, one for the server and one for the mapgen
08:47 celeron55        actually 4, also one for voxel mesh generation on the client
08:48 erlehmann        btw, i read that you like voxelareas, so maybe you like that cora and me ported it to work with mcl2-derived games https://git.minetest.land/Mineclonia/mcl_meshnode
08:48 celeron55        it's enough that you should see a difference if you test disabling and enabling a second core
08:48 erlehmann        hmm ok
08:49 erlehmann        i did not know minetest was cpu bound
08:49 erlehmann        can i reach hecktest on irc or via email?
08:49 celeron55        but it won't be a huge difference, mostly a second core gets rid of some lag spikes probably
08:49 celeron55        you'll probably have to wait for him to pop up on #-dev
08:49 erlehmann        not in mineclone5 lol
08:49 erlehmann        it's full of lag spikes ^^
08:49 erlehmann        oh i should get there again
08:50 erlehmann        it has been a loooooooong time
08:51 celeron55        you can also write an issue on github asking about opengl 1.4 support. it would be most important to list computers in use today that don't support 2.0
08:51 celeron55        the decisions will be made based on whether there is a considerable amount of users needing opengl 1.4 or not
08:52 erlehmann        well i must say the minetest community is pretty good at keeping out specific voices
08:52 erlehmann        cora was not able to register at the forums either with or without tor
08:52 erlehmann        bc sussy ip
08:53 celeron55        you can always contact a forum moderator on IRC to register on the forum if you look too much like a spammer to the forum
08:53 erlehmann        celeron55 i hope you do realize that the set of users who go to github and the set of users who play on old computers are not exactly overlapping?
08:53 erlehmann        also can you pls answer the question in the query
08:53 celeron55        yes, but our methods are a bit limited
08:54 celeron55        modern software would performan an automatic statistics collection for this purpose on all user's computers with would allow making a very quick and precise decision
08:54 celeron55        -an
08:54 celeron55        which*
08:54 erlehmann        > automatic statistics collection
08:55 erlehmann        lol
08:55 erlehmann        automatic statistics collection can be found from debian popcon maybe?
08:55 celeron55        the most we can do is make a poll on the forum, pretty much
08:55 erlehmann        does it record the machines specifics?
08:55 celeron55        i think it only records a list of packages
08:56 erlehmann        it is funny, debian shows kinda how unreliable such stats are for the lower end https://popcon.debian.org/
08:56 erlehmann        like there is 1 arm response or 8 sparc
08:57 erlehmann        basically its stats are only useful for x64 and x86
08:58 celeron55        maybe we should include a questionnaire on a release simply asking "the minetest developers would like to know which opengl version your graphics card supports in order to make development decisions, send it? yes/no"
08:59 celeron55        then it would send a http post to an endpoint on my server that appends a line in a file
09:00 erlehmann        yes
09:00 erlehmann        but the thing is
09:01 erlehmann        if you make it imposible or hard to use minetest on an older thinkpad, i bet i am not the only one who will stop playing or developing mods. usually ppl just are disappointed and go away.
09:01 erlehmann        i have seen this a bunch of times in the lats 15 years
09:01 erlehmann        last
09:01 erlehmann        often when i complain about such stuff and *know* other users have the same problem
09:01 erlehmann        devs say “you are the only person who ever complained, so it will be alright”
09:02 erlehmann        because the vast majority of users never even interact with the devs
09:02 celeron55        i always say people should complain more
09:02 celeron55        people think authors don't care, but the problem more often is information flow
09:03 celeron55        being silent when problems occur is a silly culture
09:03 erlehmann        na, it is not that they even get the idea
09:14 celeron55        erlehmann: btw, which browser did you try for the webgl test website? and does it have a setting for enabling/disabling webgl?
09:14 celeron55        possibly, does it have a setting to force-enable webgl?
09:14 celeron55        i think some browsers disable webgl on some gpus "just to be sure"
09:15 celeron55        also, does anyone know of a easy to install opengl 2.0 test program
09:15 celeron55        an*
09:16 celeron55        i think hecks has mentioned in the past which specific opengl extensions included in 2.0 are the important ones, you could check for the existence of those in glxinfo
09:16 celeron55        but, we need the list of those first
09:19 celeron55        ARB_shader_objects at least
09:19 erlehmann        celeron55 i used chromium
09:19 celeron55        but that just means shaders, if you can enable shaders in current MT that means ARB_shader_objects is supported, and gma950 definitely supports that on linux
09:19 erlehmann        ; glxinfo |grep ARB_shader
09:19 erlehmann        GL_ARB_separate_shader_objects, GL_ARB_shader_objects,
09:20 erlehmann        well shaders are dog slow
09:20 erlehmann        also i don't really see why any browser should allow random websites to compile code on my graphics card, isn't that like, a security issue?
09:21 celeron55        that's what webgl does
09:21 celeron55        well, it can be a bit of a problem on old drivers where the author didn't expect to be fed code from the internet
09:23 celeron55        on that particular site the shader source is in plain sight on the main document itself so you can just view source
09:27 sfan5            "all because minetest 5.4 is harder to get to run on raspberry pies." <- what is this about?
09:27 celeron55        erlehmann: for shits and giggles, you can try enabling webgl.force-enabled in chromium's settings
09:30 erlehmann        sfan i *suspect* it is probably just that minetest 5.3 is in debian so it is easier to get on raspis with raspian. it came up once in mineclone2 development when ppl wanted to break compat with mineclone 5.3.
09:31 erlehmann        i mean minetest 5.3
09:31 erlehmann        damn mineclone5 is living rent-free in my head
09:32 sfan5            I see, makes sense
09:32 erlehmann        what i am saying is that almost none of the raspi users was aware of the development discussion
09:32 erlehmann        that is the parallel here
09:32 erlehmann        i think minecraft does not run on such old computers as i have
09:32 celeron55        the curse of debian
09:32 erlehmann        minetest does
09:32 sfan5            I did in fact build this thing -> https://kitsunemimi.pw/tmp/minetest-5.4.0-rpi%2B.tar.gz <- once but probably nobody knows
09:33 erlehmann        oh, debian is amazingly fast updating if you work together with the maintainers
09:33 celeron55        erlehmann: minecraft did run on core duo integrated graphics, and linux no less, 11 years ago
09:33 erlehmann        i have drawn about 500 or so emojis for GNU unifont and it was in debian way before it was in arch or something like that
09:33 celeron55        in less than a year they then made it unplayable
09:33 erlehmann        yeah, but minetest *still* runs there
09:33 sfan5            I have also considered that we should provide official RPi builds or something similar but didn't get to that yet
09:34 erlehmann        sfan5 it would probably make it easier
09:34 sfan5            (would be a .deb and hopefully linked on the download page)
09:34 celeron55        official raspi builds would be great marketing, if nothing else
09:34 erlehmann        i mean, i keep saying that minetest 5.3 is a bigger security risk than minetest 5.4 (due to the CSSSM), it is not a joke
09:35 erlehmann        celeron55 nah, if the packaging is *good enough*, debian may take it
09:35 celeron55        i'd rather have distros repackage everything to their taste than take our builds
09:35 erlehmann        yeah working with distros is better
09:35 celeron55        that's the only thing where they have some pride and taste
09:35 erlehmann        usually there is a reason when a program has a deb on their webpage
09:35 celeron55        making packages, that is
09:35 erlehmann        but it is not in the distro
09:36 erlehmann        usually QA issues
09:36 erlehmann        nah, debian also does the whole porting and security stuff
09:36 erlehmann        like they will fix bugs if upstream is shitty
09:38 erlehmann        and if one package has a bug, debian can tell you which others do
09:40 erlehmann        celeron55 i hope if minetest rendering gets refactored that it will be flexible enough to have a rougelike-style rendering too ^^
09:40 erlehmann        then i do no longer need to worry about opengl or 3d
09:40 erlehmann        roguelike
09:40 erlehmann        i mean
09:44 celeron55        well, i mean, just do that if you'd like, but it won't really relate to what's happening now, it would be a different abstraction layer, probably practically requiring forking or rewriting the client and changing almost everything
09:45 celeron55        everything down to pointing at nodes would be different
09:45 celeron55        not just rendering
09:46 erlehmann        well, i guess waspsaliva could already do it
09:46 erlehmann        i mean cora has a macro to build heart shaped nether portals
09:46 erlehmann        not that i really want it
09:54 celeron55        an SDL based minetest client that could run on a 386 would be interesting
09:56 celeron55        not really useful, but interesting
09:57 celeron55        your biggest problem would be memory usage, 3d voxel data is incredibly memory heavy in 1980s standards
09:59 celeron55        fps would probably be non-existent if you didn't pour months or years of development time into optimization
10:11 erlehmann        well
10:11 erlehmann        i wonder if brutal minecraft can run on older computers
10:11 erlehmann        i saw screencasts of it
10:11 erlehmann        it is a doom mod, but minecraft-style levels
12:41 pgimeno          @erlehmann: in my view, if Minetest does not want to die from lack of users, it needs to update to more modern graphic standards, and by that I mean appeal, not capabilities; however the former implies the latter.
12:42 pgimeno          If it only appeals to people who can run it on a potato, it leaves out people who can run it on a more modern machine and feel it's too old looking.
12:43 pgimeno          And the latter group is almost certainly more numerous.
12:45 MTDiscord        <Warr1024> Minetest cannot die from lack of users.  If it could, it would be modding API issues that would do it, not graphics.
13:14 erlehmann        pgimeno basing something on a hypothetical what future users would like and future technologies would enable sounds suspiciously like every other case of “let's break compatibility” i have seen. the actual argument is more honestly written out ass “i care neither about your use case nor your opinion – so let's do what i, personally, think is trendy”.
13:15 erlehmann        pgimeno there may be good reasons to break compatibility, but “some hypothetical ppl i just made will not use the software” sounds like it is not among them.
13:15 erlehmann        i just made → i just made up
13:49 pgimeno          no, currently shadows suck because the graphics subsystem sucks, it's not merely hypothetical
14:08 sfan5            performance issues on certain current hardware are real, too
15:14 rubenwardy       10 years old machines in 2021 were up to date machines in 2011
15:14 rubenwardy       dated graphics code is likely the cause of performance issues
15:14 rubenwardy       along with dumb stuff like n^2 collisions and not batching rendering
15:32 celeron55        saying MT does what's trendy is the most ridiculous argument i've ever heard
15:32 celeron55        everything we do is so far from trendy i forgot the word even existed
15:33 MTDiscord        <Warr1024> Just make a PR that removes all the suck but keeps everything else intact.
15:33 celeron55        developing MT is almost like making not being trendy an artform
16:00 pgimeno          trendy? it's all about improving graphics, it's even in the roadmap
17:47 erlehmann        celeron55 sfan5 i have read up on opengl history and wow what a ride. i think i will probably use the last version of minetest that runs on my main computer and try to backport protocol changes, if the rendering is ruined. or just find something else that is similarly fun.
17:47 sfan5            the thinkpad from 2011 is your main computer?
17:47 erlehmann        yes
17:47 sfan5            huh
17:48 erlehmann        it's the last widely prduced laptop that has a 4:3 matte screen
17:48 erlehmann        and up until this openGL thingy the only problem was disk speed, which you can solve with an SSD
17:49 erlehmann        i may switch to reform, but so far there is no full disk encryption there
17:49 erlehmann        and who knows what else breaks because even less people hve reform 1 or 2 than old thinkpads
17:49 Krock            and after that, your intel iGPU will be the bottleneck
17:49 erlehmann        ?
17:49 sfan5            I'm pretty sure you can make FDE work there if it runs u-boot
17:50 erlehmann        it is just not there yet
17:51 erlehmann        anyways, i intend to make the reform2 my main computer *if it works*, but if minetest is going after “how many people have the hardware” i am *definitely* worse off with reform, i'll have to find something else eventually
17:51 erlehmann        i understand now why everyone and their dog loves shaders
17:52 sfan5            because they're needed to make modern things work?
17:52 erlehmann        i doubt that is the reason, because for most purposes that are not high-performance gaming it is not necessary
17:53 erlehmann        i think the reason is that devs appreciate the mental model much more
17:53 Pexin            mmmm
17:53 sfan5            sounds like whatever you read about OpenGL is outdated then
17:53 Pexin            m.m.m.m.
17:54 erlehmann        sfan5 no i am going by “which apps have 3d will not work without shaders” and the vast majority seem to be games
17:55 erlehmann        i have yet to encounter a screensaver, for example, that does shader stuff, but maybe the era of screensavers was before GLSL
17:55 Pexin            currently typing on a laptop that "should" run MT ok, but nvidia has no updated drivers for this which are compatible with modern linux kernel. and nouveau is.. lacking
17:55 erlehmann        anyway, i get it now. i just hope whatever rendering changes minetest makes are better handled than the shadow mess.
17:55 Krock            Pexin: there's still the legacy drivers
17:56 erlehmann        i.e. not breaking unrelated things
17:56 Pexin            Krock: already pursued. legacy drivers not compatible with this kernel. or at least, refuses to install.
17:56 erlehmann        Pexin how is software rendering lol
17:56 erlehmann        LIBGL_ALWAYS_SOFTWARE=1 on the thinkpad gets me about 1 fps
17:57 Pexin            it's not really software rendering,
17:57 erlehmann        that's like a fifth of opengl ES performance!
17:57 Pexin            but nouveau is missing something
17:57 sfan5            Pexin: did you try the package from your distribution or "on your own"?
17:57 Pexin            sfan5: both
17:58 sfan5            distro?
17:58 erlehmann        well the upside of shaders is much cooler cheat clients i guess
17:58 Pexin            well, there isnt one in the distro
17:58 sfan5            that's a no then
17:58 Pexin            but the ppa yes
17:58 sfan5            just trying to compile the sources from nvidia won't work, but distros often provide the necessary patches with their packages
17:58 Pexin            the official legacy from nvidia has its own windows-style installer which not only didnt work, but screwed my packages
17:59 Krock            huh? the version your package manager comes which should be the most stable one
17:59 MTDiscord        <Jordach> Linux just works™️
17:59 Krock            s/which/with/
18:00 Pexin            Krock: there isnt a compatible legacy driver in the standard ubuntu repo, the official ppa wont install because it hates this kernel
18:00 sfan5            installing vendor packages especially drivers on Linux is a fatal mistake
18:00 erlehmann        Jordach same as OS X, it works until you install some weirdo vendor shit that generally does not play nice with the system and breaks every upgrade
18:00 Pexin            sfan5: yeah. I noticed   :]  managed to fix it though
18:00 MTDiscord        <Jordach> I don’t know getting Minetest purring was straightforward
18:00 erlehmann        a friend of mine has an old macbook with some VJing sotware bc that's sometimes her job and … well
18:01 sfan5            which GPU is it anyway?
18:01 erlehmann        TIL: minetest is a cat type pokeman
18:01 Krock            Pexin: hmm. I don't know about 5.x kernels, but at least early 4.x kernels surely worked well with it
18:01 erlehmann        LOL am i getting better performance right now than some nvidia card?
18:02 Pexin            dang, I forget what chipset. compaq presario v6000 anyway
18:02 Krock            you'll easily reach the performance of a Riva TNT 2 card... which is an ancient nvidia card
18:02 Krock            > either NVIDIA GeForce Go 6150 or Intel X3100 depending on cpu family or NVIDIA GeForce Go 7150M
18:03 Krock            that's truely an ancient system
18:03 Pexin            yeah
18:03 erlehmann        celeron55 sfan5 so what's the plan, when do i have to stop playing minetest and writing stuff for mineclonia until i found a new computing platform i like, is it this year or next year?
18:04 erlehmann        like when is openGL 1.4 support dropped
18:04 Pexin            (I don't do real work on this machine, don't worry)
18:04 erlehmann        Pexin i have found that having an older machine make whatever you code buttery smooth on current-gen machines
18:04 erlehmann        bc you literally can't cheat on performance
18:04 sfan5            dunno, this year? you'll find out when it happens
18:04 Pexin            friend gave it to me on the off chance I could make it do something useful. the old windows install had broken
18:05 erlehmann        i am already not happy with such a vague threat of breakage
18:05 erlehmann        i guess "when it's done", like everything
18:06 erlehmann        i wonder if cheat client authors will love it or hate it
18:06 Pexin            OH and when i try to use it for zoom, after a minute video and input both freeze completely, but sound keeps going
18:06 erlehmann        i mean on the one hand fullbright and brightnight will have to be rewritten prob
18:06 sfan5            neutral, rewritten rendering is irrelevant to them
18:06 erlehmann        on the other hand, it's probably not much
18:06 Pexin            especially when the meeting goes into screensharing
18:07 erlehmann        i mean fullbright is easy
18:07 rubenwardy       erlehmann: shaders are needed to make the best use of the GPU, which results in greater rendering performance - even with an iGPU. They're not just for fancy effects
18:07 sfan5            rubenwardy: not with an igpu from 2011 though ;)
18:08 rubenwardy       hmm, didn't they start to get good around then?
18:08 sfan5            dunno
18:08 sfan5            few years later I'd say
18:10 erlehmann        i have a thinkpad T60 as my main machine. i have several broken thinkpad T60s in storage in case the one breaks. otherwise i have a thinkpad R60 and a reform 1 (imx6) and reform 2 (imx8). i never really expected to *have* to use the reforms, but it might soon be the only computers i actually own that can run minetest.
18:11 erlehmann        if anyone else knows a modern computer with a 4:3 matte screen and a non-shitty keyboard that has replaceable parts, speak up. it *probabld* does not get better than reform or old thinkpad.
18:11 erlehmann        probably
18:11 sfan5            what's wrong with 16:10?
18:13 erlehmann        that the item bar is not at the side of the screen or something?
18:13 erlehmann        manufacturers have abandoned 4:3 in favor of the newer 16:9 widescreens around 2012 or sso
18:14 Krock            you guys have item bars?
18:14 erlehmann        Krock are you playing node core or what
18:14 Krock            just F1
18:14 erlehmann        sfan5 actually i want two terminals beside each other that are much higher than wide
18:15 erlehmann        i can kinda get a similar effect on 16:9 with three terminals besside each other, but it's a waste of screen real estate
18:15 erlehmann        also i actually desire a minimum *height* of the screen
18:15 erlehmann        and guess what, 16:9 screens are not that tall
18:16 Pexin            I only minetest in windowed mode anyway
18:16 erlehmann        well, regardless
18:17 MTDiscord        <Jordach> what's wrong with 16:10
18:19 erlehmann        look i don't care about your tastes in screen dimensions. i usually do not play proprietary games becaue they keep breaking, after 5 or 10 or 15 years they are often unplayable unless source is released or someone re-implements the engine.
18:19 MTDiscord        <Jordach> do you live in an office or something
18:20 erlehmann        no but i do work from home and have only a door for the bathroom so technically yes?
18:20 erlehmann        correction, there are two doors
18:20 erlehmann        but one is for the space before the bathroom and permanently open
18:22 erlehmann        Jordach how is your quake inspired minetest shooter going?
18:22 MTDiscord        <Jordach> does using a different networking system make any sense
18:22 erlehmann        no idea, have you tried using SCTP?
18:23 MTDiscord        <Jordach> i'm writing something with simple sockets
18:23 erlehmann        if you don't know, check this out https://en.wikipedia.org/wiki/Stream_Control_Transmission_Protocol
18:24 erlehmann        have you by now put an indestructible wall between your map content and the ignore nodes?
18:24 MTDiscord        <Jordach> i have other things in my development pipeline
18:24 erlehmann        ok, well, think of me when you get strange interactions with ignore nodes!
18:24 erlehmann        they are so very funny
18:25 MTDiscord        <Jordach> i'll give you a hint
18:25 MTDiscord        <Jordach> a way to load mapblocks with regular SAO would solve it as ignore is treated as collidable
18:25 erlehmann        uh what
18:25 erlehmann        like a mapblock load anchor?
18:25 MTDiscord        <Jordach> basically any entity would then load the world like players
18:26 erlehmann        that does sound like a performance nightmare
18:26 erlehmann        but your maps are p small
18:26 MTDiscord        <Jordach> the server loads a bigger area than you think
18:26 erlehmann        as long as it does not become a default in minetet proper …
18:27 erlehmann        that is always true for the server
18:27 erlehmann        the aforementioned ignore blocks are a symptom of that
19:46 independent56    My server migration is complete. Wordpress is broken, but other then that, everything is fine. You probably wouldn't even notice any difference apart from speed.
19:51 independent56    Ha, who needs to migrate to mediawiki when you can just install the monobook tempalte XD
22:37 Alias            A couple of minutes ago, my server crashed and sent me error messages like so:
22:37 Alias            StaticObjectList::remove(): id=39444 not found
22:37 Alias            ServerEnv: Failed to emerge block (29242,29793,0) when deleting static data of object from it. id=39442
22:37 Alias            Server is on 5.4.1 and postgresql 11
22:38 Alias            Should I worry? take any action? Or shrug it off?
22:39 sfan5            what you are quoting is not a fatal error, if it crashed it should be because of something else
22:44 Alias            Those and similar were the last messages in the log, no other error before or after at that time
22:45 Alias            https://pastebin.com/kFW4hUWW
22:46 Alias            Those were the exact lines, before that came around 15 logouts of players
22:46 Alias            log level is action
22:46 sfan5            nothing before the logout of all players?
22:48 Alias            Nothing suspicious. People placed blocks and sent chat messages
22:48 Alias            Noone was in the area mentioned, as far as I can tell
22:50 sfan5            well the server suddenly existing must have come from somewhere, it wasn't this
22:50 sfan5            exiting*
22:53 Alias            I grepped for ERROR in the log, last error was half an hour ago
22:54 Alias            I seem to have "suspiciously large amount of objects detected" often
23:20 AliasAlreadyTake Is there anything specific I should look for?
23:21 AliasAlreadyTake I went several minutes into the past from the point of crash, but there's nothing but placing actions, logins, logouts, a couple of mobs attacking
23:24 erlehmann        AliasAlreadyTake it is possible for minetest to unceremoniously crash without leaving a trace
23:25 erlehmann        in that case whatever came before is not necessarily the reason
23:25 erlehmann        even *most likely* not the reason
23:25 AliasAlreadyTake Nonono! I have not allowed minetest to do such a thing! :D
23:28 AliasAlreadyTake However, if those error messages are not responsible for the crash, but a consequence, then could anyone take a guess what *might* have lead to it?
23:38 erlehmann        AliasAlreadyTake, have you checked that the OOM killer has not gone on a murderous rampage?
23:39 erlehmann        > Failed to emerge block (29242,29793,0)
23:39 erlehmann        looks edgy
23:40 AliasAlreadyTake How would I collect evidence on the OOM killer thingy to put it in jail?
23:44 AliasAlreadyTake Not sure why it would want to emerge this block or the other one (0,704,-23656) for that matter. No player near, noone has any business there
23:47 erlehmann        nether portals or shit?
23:47 erlehmann        i'd check all the mods for emerge calls prob