Time Nick Message 14:54 paramat i think #7582 needs to be reviewed, updated and merged asap to avoid conflicts, looks like good work 14:54 ShadowBot https://github.com/minetest/minetest/issues/7582 -- lua_api.txt fixes and format improvements by pauloue 14:54 paramat even if it means a follow up PR is needed to tidy up a few points 14:55 ANAND Agreed 14:56 paramat anyway, i'm reviewing now 14:58 ANAND Minor suggestions can still be included with the PR if done quickly 15:00 ANAND Updated #7557 to remove accidental addition of irrelevant modifications 15:00 ShadowBot https://github.com/minetest/minetest/issues/7557 -- Expose player FOV to Lua API by ClobberXD 15:01 ANAND Apparently some test_config.h was created in src/ 15:07 nerzhul ANAND never run cmake on the source folder 15:07 nerzhul always create a subfolder and do cmake .. 15:07 ANAND Oh ok 15:07 ANAND Thanks 15:09 paramat +1 for 7582, Krock do you agree to merge it as-is if necessary to avoid conflicts? maybe we can delay anything that can cause a conflict 15:10 Krock #7582 15:10 ShadowBot https://github.com/minetest/minetest/issues/7582 -- lua_api.txt fixes and format improvements by pauloue 15:11 ANAND paramat, minor suggestions can still be included with the PR if done quickly 15:11 Krock it's not complete yet https://github.com/minetest/minetest/pull/7582#pullrequestreview-140231954 15:11 ANAND Like these ^ :) 15:11 Krock trivial #7583 - any objections on merging it? 15:11 ShadowBot https://github.com/minetest/minetest/issues/7583 -- Replace apt-get with apt by niansa 15:13 paramat +1 15:14 paramat i know 7582 isn't complete, i'm suggesting merging unfinished is better to causing a conflict :) 15:15 Krock no, that's not the right PR handling policy 15:15 paramat anyway we'll see what happens 15:17 paramat no, that's what makes sense in this situation :) 15:17 Krock what makes sense is to fix the issues and merge a PR when it's done (mergeable) 15:18 Krock !tell lhofhansl Is PR 7544 (readonly db) done? It works, but would you mind removing the WIP tag in case it's mergeable 15:18 ShadowBot Krock: O.K. 15:19 paramat we often split a task into multiple PRs if necessary, it's no big deal 15:20 ANAND Then I guess this goes into a separate PR: https://github.com/minetest/minetest/pull/7582#issuecomment-407791483 15:28 paramat well, hopefully it will be updated soon 15:28 paramat +1 for #7580 Krock 15:28 ShadowBot https://github.com/minetest/minetest/issues/7580 -- Settings: Fix indents, chat_message_*. Update translations by SmallJoker 15:43 lhofhansl Hi all... #7544 is ready to be merged. 15:43 ShadowBot https://github.com/minetest/minetest/issues/7544 -- Allow an optional readonly base database by lhofhansl 15:45 ANAND Hmm... '!tell' doesn't work? 15:45 paramat is approved so go ahead 15:45 ANAND !tell lhofhansl Is PR 7544 (readonly db) done? It works, but would you mind removing the WIP tag in case it's mergeable 15:46 lhofhansl It worked. :) 15:46 lhofhansl Got that message and remove the WIP and Discussion tags. 15:46 lhofhansl Will merge in a few. 15:53 lhofhansl Has anybody used Github's "Squash and merge"? I usually don't trust it and apply the change myself to master. 15:53 paramat yes, most of us use it 15:53 paramat even i do now :) 15:53 lhofhansl OK... I'll try it this time. *Crossing fingers :) 15:54 paramat the commit message may need a little reformatting 15:54 lhofhansl Done. 15:55 lhofhansl Hmm, yeah, should've edit the commit message :) Could amend it, but it's good enough I think 15:56 paramat it's not so bad 15:58 paramat merging #7580 in 5 moins 15:58 ShadowBot https://github.com/minetest/minetest/issues/7580 -- Settings: Fix indents, chat_message_*. Update translations by SmallJoker 15:58 paramat *mins 16:03 Krock please use squash & merge. it's the best way to merge PRs where the single commits don't matter 16:03 Krock and never merge alone. it causes the history to go non-linear 16:05 paramat merging 7580 16:07 paramat done 16:09 lhofhansl Can you tell I'm on vacation? :) ... I get actually get some time to MT patches. 16:14 paramat i can =) 16:15 paramat it's good to see you active 16:35 lhofhansl It's nice and fun to have a little bit of time. 16:35 rubenwardy Agreed 16:44 lhofhansl What's our general Android story? Right now it seems 99% of the work is done, but the last 1% makes it look bad. The loading, the config UI, Sun/Moon/FarFog/Stars not rendered. 16:44 lhofhansl I guess just not enough folks interested in it...? 16:45 paramat seems like much less than 99% is done, see the issues :) 16:46 Krock the opengl issues are right now the most important 16:46 paramat android was difficult to build until recently, many of us don't own a device or see mobile as low priority 16:46 Krock controls are pretty good now, other things only need tweaking 16:47 Krock > many of us don't own a device 16:47 paramat but of course, bad apps on servers are high priority to servers 16:47 rubenwardy Compiling puts me off 16:47 rubenwardy Last few times I just gave up 16:47 paramat it was mostly the controls putting players off 16:47 Krock almost everyone in europe owns a device which can install apps 16:47 lhofhansl I think it looks pretty goo now (rubenwardy sent a screen shot of the new controls on an issue), just need to figure out how to render the sun, moon, etc, and then it might be playable. 16:48 paramat except paramat 16:48 lhofhansl Perf is not good compare to MC, though. 16:48 rubenwardy Except paramat 16:48 paramat lol 16:48 rubenwardy Well, we need someone to be an optimisation wizard 16:48 rubenwardy Would help on desktop too 16:59 Krock Merging #7583 in 5 minutes 16:59 ShadowBot https://github.com/minetest/minetest/issues/7583 -- Replace apt-get with apt by niansa 17:06 Krock merging.. 19:01 Krock Shall I include a protocol bump (+ backwards drop) for #7569 ? The network will surely not be compatible anymore.. but it's 5.0.0-dev. 19:01 ShadowBot https://github.com/minetest/minetest/issues/7569 -- Send custom floats (log2) by SmallJoker 19:01 Krock ^ sfan5 ? 19:01 sfan5 of course 19:08 Krock lovely. Just found out that the entire "BufReader" class in the serialization code is only used in the unittest 19:09 sofar LOL 19:09 Krock let's see when this stuff was added.. 19:09 nore hmm... is there any reason not to use IEEE754 floats? 19:10 Krock added by 1a5b4b38f3, never used yet. great. 19:11 Krock nore: there are already a few links in the PR. the memory float representation is not defined by the standard 19:13 Krock and the code to re-assemble the entire float to how we want it seems rather slow 19:13 nore well, since most compilers actually use IEEE754, it could just be possible to use the identity function on those compilers 19:14 nore and use fallback code for the other cases 19:16 Krock btw, a float can represent 1E-39 when the binary is compiled with "g++ main.cpp", but not when "-O3 -ffastmath" is specified 19:17 nore what do you mean? 19:19 Krock that the float handling varies depending on the compiler flags. I did not investigate the occupied bytes, but that alone makes me doubt that it's always the same (even on the same compiler) 19:19 nore ah, the fast that 1E-39 is a denormalized number? 19:19 nore n*fact 19:20 nore although, I think this shouldn't affect ffastmath 19:33 Krock Ah I see. #3013 BufReader was actually designed for the network - but never found its way into "the production" 19:33 ShadowBot https://github.com/minetest/minetest/issues/3013 -- Add BufReader and vector-based serialization methods by kwolekr 19:55 p_gimeno IEEE floats are the best idea. 1E-39 is not representable in a normal float, it needs a denormal. Denormals are not universally supported, but when they are not, the value 1e-39 gets turned into 0, which is not a big deal. 19:57 p_gimeno I already suggested to have up to 4 functions to handle 4 different kinds of byte order for machines that support IEEE floats, and a slow fallback function in case the machine doesn't (which is unlikely to be any machine in which Minetest runs) 20:00 p_gimeno I don't have much time to prepare a patch lately, unfortunately 20:55 Krock even with real float representation - the unittests put me off. The float might have the one or other value 20:55 Krock depending into which direction it rounds