Time Nick Message 00:15 MichaelRpdx http://pastebin.com/wNxiQ2RM 00:20 sapier interesting ... let me guess lock is held by 11988? 00:24 MichaelRpdx and yeah, but what is 11988? 00:25 MichaelRpdx we're back to ??? 00:25 MichaelRpdx I didn't see that earlier 00:25 sapier1 11988 is the messed up thread 00:25 MichaelRpdx yes, saw that. 00:26 sapier1 sorry but I need to get some sleep that information seems to be helpfull I'll check tomorrow 00:26 MichaelRpdx Sleeep is good. 16:09 Megaf Folks, I'm looking at the code trying to find an existing way to implement control via arrow keys instead of mouse, but I can't find anything very clear, can anyone help? 17:11 CiaranG Megaf: Try this: https://github.com/CiaranG/minetest/commit/dbb97e12a9a6b92bccdb661cdab772868d01af7c 17:13 CiaranG (in other words, there is no existing way, but if you build with that it will do it) 17:13 Megaf CiaranG: That's really cool 17:13 Megaf So, that's is not part of minetest yet 17:13 CiaranG No 17:14 Megaf CiaranG: did you request a pull? 17:14 CiaranG No. There is a problem that I would want to resolve before I thought about that. 17:15 CiaranG If you map the left/right mouse click to something else (i.e. a key) you can't then map it back to the mouse button (because the gui captures the mouse clicks) 17:15 CiaranG (of course, you could just change the config file to map it back) 17:15 CiaranG But it's good enough for what I needed it for, so I haven't tried to fix that yet. It's probably very easy. 17:16 Megaf CiaranG: did you see my topic? 17:16 CiaranG No? 17:16 Megaf https://forum.minetest.net/viewtopic.php?id=8552 17:17 CiaranG Ha 17:17 Megaf so, it would be great if you help us 17:18 CiaranG Well, I already did, see above ;) 17:20 CiaranG All you have to do is convince someone to merge it I guess 17:22 CiaranG Or testing it would help, if you're able to build? 17:30 Megaf I am, I have a fork too 17:30 Megaf Im going to used that on my fork and see how it works 17:58 Megaf Ok, thank you CiaranG 19:05 sapier MichaelRpdx: are you there? 19:16 celeron55_ CiaranG!? 19:16 celeron55_ the first freaking contributor to the project of all time is here 19:17 sapier yes he's already here for some time ;-) 19:18 sapier celeron55 do you have a opinion to making player client version available through modapi? 19:19 celeron55_ sapier: for what purpose? 19:21 sapier well it's primary for debugging purposes but could be usefull for statistics and player information about best version to use with a server too. Right now it's disabled for non debug builds in my pull request 19:22 sapier I use this feature to create a benchmarking mod for various network protocols. Of course that's nothing to be used on a real server 19:24 sapier as ShadowNinja already intended to abuse it for player blocking I have mixed feelings about making it generaly availabe 19:27 celeron55_ player blocking? 19:27 celeron55_ what for? the engine already does that and it's its responsibility anyway 19:28 sapier well I understood something like block clients beeing not official minetest client ... I don't like ideas like that 19:28 celeron55_ what 19:28 celeron55_ that's ridiculous 19:28 sapier true that's why I already limited it to debug version only 19:30 celeron55_ anyway i don't like that separated major/minor/whatever thing; and also it should contain the name of the client 19:31 sapier you have a mixed mode version in version string 19:31 celeron55_ for statistics purposes a string like "Minetest 0.4.9-git" would be completely fine and probably preferred 19:31 sapier I need the version numbers in core 19:32 celeron55_ what do you do with them? 19:32 sapier right now we have no way to workaround the different client behaviour because of someone didn't realize he changed behaviour in a way that'd have required bumping protocol version 19:32 sapier basicaly any client as of 0.4.7 does behave slightly different 19:33 sapier I can't proove it but I have reason to believe some tablet clients behave different too 19:33 celeron55_ clients having subtle changes without version changes has always been the case 19:33 celeron55_ why is it a problem now? 19:34 celeron55_ do you realize you break all what was beautiful about minetest versioning with that thing 19:35 sapier I do see the risk but on the other hand those changes we didn't realize by time to bump protocol version stop us from beeing compatible 19:35 celeron55_ break compatibility then 19:36 celeron55_ and change things so that both of these apply: 1) we don't need to expose client version for any functional purpose, 2) such breakage is 100 times less likely to occur in the future 19:36 sapier that's easy to say but usually results in "you can't change that thing because of noone can connect after it" ... for things you have to break to fix even worse things ... resulting in blocking dev 19:37 celeron55_ by making that change you just make everything super obscure and nobody can anymore know what works and what doesn't, and also you effectively lock out non-minetest clients due to arbitrary version checks 19:38 celeron55_ i don't approve any of this 19:38 sapier well it's already very obscure 19:38 celeron55_ yeah, so let's make it more obscure 19:38 celeron55_ great idea 19:39 sapier I don't think it's more obsucre to know what's gonna happen, right now it may or may not work and you neigter can predict nor workaround minor issues 19:39 celeron55_ i'll say you are only allowed to add as many lines as you remove for this 19:40 sapier ? 19:40 sapier what do you mean? 19:41 celeron55_ i mean i don't like making it more compicated 19:41 celeron55_ +l 19:43 sapier well if you don't want the version number I'm not gonna add it but I recommend bumping min protocol version to 23 too as old clients are incompatible and there's no way to fix the basic issues without breaking compatibility 19:43 MichaelRpdx sapier here now 19:45 sapier working around issues by protocol version which might have been changed on breaking behaviour or not is at best a ugly hack not a solution 19:45 celeron55_ i think it is a right time for breaking compatibility and making a clean solution for the current time and the future 19:46 sapier actully I thought we wanted to do this in final decision about network protocol 19:46 celeron55_ what's the news on that then? 19:47 sapier right now we >can< handle old clients in almost 99% of cases by abusing protocol version to do compatibility ... but that's as ugly as using client version 19:48 sapier nothing new because everyone lost interest and I'm still tracking down some strange bugs ... some jthread relate some in there forever 19:49 celeron55_ i think this is going to a really bad direction 19:49 celeron55_ we should be making things more simple and robust, not more complicated and unstable 19:49 Megaf indeed 19:49 sapier well removing the big logs is prone to make it unstable but it's necessary 19:49 sapier locks 19:50 sapier and right now I'm the only one really looking for those issues ... most of people do android development only 19:52 sapier ok just to make sure, is it commonly accepted to bump min protocol version? 19:55 sapier guys I can't do this without clear position of at least some core devs 19:56 Megaf sapier: Can't you create o poll or something like it? 19:56 celeron55_ no it isn't accepted, without being sure that the chance is then taken to improve everything made possible by it 19:56 celeron55_ but if that is done, then it is absolutely accepted 19:57 sapier "improve everything made possible" ... it'd be more easy to rewrite it in total then to guarantee this "I want everything and now" demand 19:57 celeron55_ (and if it isn't, then that's just because of ignorance) 19:58 sapier especally as there's noone to review any changes 19:58 celeron55_ well i think everyone can agree that we don't have the resources to maintain complicated things or to cause any more complication to anything, and neither do we have resources for breaking compatibility and improving everything 19:59 celeron55_ so maybe the only real alternative then is to break compatibility and hope that it frees developer time for productive things 19:59 celeron55_ in the long run 19:59 sapier ok so why didn't you stop adding a insane complicated thing like texture downloading and https? 19:59 sapier -s 19:59 sapier both requires client/server handshake as server needs to know about client beeing ready to receive further information 19:59 celeron55_ me? i've done and looked at absolutely nothing in the past month 20:00 celeron55_ (on the other hand, i've made great progress in an another project of mine) 20:00 sapier exactly but now as I try to fix the issues added in this time you seem to expect me to fix everything bad at once ... at least that's what I understand 20:01 sapier everytime I fix a issue I stumble into another one beeing in there for quite some time but beeing hidden of another one 20:02 CiaranG celeron55_: yes, me. hey :) 20:02 sapier but that's something different. what I'm interested atm is shall I remove compatibility handling or not 20:03 sapier it uses protocol version same way it'd use a version number so it's basicaly no difference 20:03 celeron55_ whatever answer i give, it's a bad answer 20:03 celeron55_ maybe everything is going to end horribly and we should stop before it's too late 20:03 celeron55_ it's done now 20:04 sapier well none is as bad as not deciding at all that's just "I don't wanna decide because of I want to be able to blame you for it later" 20:04 celeron55_ if the compatibility is removed, then we blame the ones that made compatibility too hard, right? 8) 20:05 celeron55_ (which is me) 20:05 sapier actually the last months I have the feeling I'm made reponsible for everything beeing wrong in mt ;-P 20:07 sapier I'm gonna open up a poll in forum ... I hope core devs will respond 20:07 celeron55_ at least you expose every possible problem with whatever you do and then ice it up with a big chunk of complicated code that mostly makes it worse 20:07 celeron55_ or at least that's how it looks like 20:08 sapier well adding client/server init handling can't be done with no code ... there was none before 20:08 sapier ok it was there by using a bunch of bool variables to form "some sort of client state" 20:08 celeron55_ wasn't the previous "no code" proof that it can? 8) 20:09 sapier well that's why it didn't work e.g. uninitialized players in world, lost player models on player object sent prior model transmission, chat messages not sent to those who should receive it 20:09 sapier basicaly it's been your idea to fix the player standing in world 20:14 celeron55_ should we start a 0.5.0 branch that aims for more stability and simplicity with no compatibility, and kind of opportunistically imporve things on it until there doesn't seem to be more to do (which is when it's released) 20:14 celeron55_ and give that enough time 20:14 sapier well I don't believe that's gonna happen anytime soon there's way to much to do 20:14 sapier e.g. splitting env lock 20:15 celeron55_ what's not going to happen? 20:15 sapier and community is already split I don't believe splitting it even more will improve situation 20:15 sapier "nutil there doesn't seem to be more" 20:15 celeron55_ that's not a split; it's a way larger changes are usually developedd 20:15 celeron55_ -d 20:16 celeron55_ well ffs i can decide when it's done if that's the problem 20:16 celeron55_ seems like the smallest of problems 20:20 celeron55_ would *you* be fine with that branch? 20:20 celeron55_ (there's nobody else here anyway) 20:20 sapier well I don't know what to do there 20:20 sapier https://forum.minetest.net/edit.php?id=129782 20:21 * VanessaE peeks in 20:21 sapier imho this feels like a second freeminder branch ... I did compare it to current master yesterday ... it's argonizing how much we have differed, both have fixes in there the other one doesn't have (not counting controversial things) 20:22 VanessaE sapier: missing player models and/or player nametags 20:24 VanessaE if it was up to me, and I was the one making this decision, I would side with c55 here - start a new branch, break compatibility, but rewrite the entire network protocol from scratch if that's what it takes to "get it right", make that the official code when the time is right. 20:26 sapier if I'd believe it was only about network protocol I'd do it but there'll be package format, threading data format ... all of those issues which regularly pop up 20:26 sapier And I can't do all of them on my own 20:31 sapier well right now we have a draw in that poll ... but doesn't seem to draw a lot of interest from core devs 20:35 sapier sometimes I understand why proller and xyz have left 20:47 celeron55_ sapier: you linked to the edit page of your post; this is the actual post: https://forum.minetest.net/viewtopic.php?id=8571 20:48 sapier thanks, fixed it in #minetest forgot to fix it here ... sorry 20:50 VanessaE sapier: regarding your current branch, where do we stand? is that going to be pushed upstream then? 20:50 VanessaE (the one I've had under test) 20:50 sapier no not without a decision 20:50 VanessaE ok 20:51 VanessaE do you have that link handy to that compatibility table you had worked out? 20:53 sapier https://gist.github.com/sapier/8907170 20:58 EvergreenTree Idea that I'm willing to do: Give the furnace fire an animated texture? 20:58 EvergreenTree instead of that static flame image 20:59 EvergreenTree Anyone think it should be done? 20:59 VanessaE https://github.com/VanessaE/animated_furnace' 20:59 VanessaE https://github.com/VanessaE/animated_furnace 21:00 VanessaE Dan already did it (that's just my copy if it) 21:01 EvergreenTree ah 21:01 EvergreenTree Well, I can integrate and merge it then. 21:02 VanessaE but that's a minetest_game issue, not a core issue like what sapier and c55 are talking about 21:04 EvergreenTree Well, sorry if I'm breaking the discussion 21:05 VanessaE noit's okay 21:05 VanessaE it's a good idea 21:06 sapier discussion fell a sleep some minutes ago so don't bother EvergreenTree 21:06 VanessaE textures and code in that mod are suitable for minetest_game, so go for it :) 21:06 VanessaE s/don't bother/don't worry about it/ 21:06 EvergreenTree Okay 21:06 EvergreenTree Thanks 21:07 EvergreenTree I might make my own texture for it 21:08 VanessaE ok 21:09 EvergreenTree I'm guessing he used the default animated fire texture for it 21:14 VanessaE I'm not sure how he made that texture actuall 21:14 VanessaE +t 21:14 VanessaE -t+y 21:26 sapier https://github.com/minetest/minetest/pull/1142 this one seems to get more and more important, MichaelRpdx's problem seems to be a luajit issue too ... this time it's not in mainmenu but in generic server startup 21:40 sapier well if noone complains I'm gonna push 1142 in about an hour 21:54 VanessaE sapier: it seems that "out with the old, in with the new" is the general consensus on your poll so far. 22:11 VanessaE bbl 22:31 sapier ok pushing #1142 now