Time  Nick      Message
12:21 _Megaf    good morning gentleman
12:47 nrzkt     sfan5: why don't you test LZ4 ?
12:48 sfan5     i have no time right now
12:49 sfan5     i can do lz4 later
12:49 nrzkt     okay, lz4 wrapper and compilation isquite easy, two functions :)
12:50 nrzkt     1 compress, 1 decompress, 4 files to add to cmake lib
12:50 nrzkt     lz4 doesn't take any cpu on my server it's impressive :)
12:50 nrzkt     for your compressions tests, did you want to change compression for sent blocks to client onr only for storage?
12:51 Calinou   sent blocks should definitely be compressed
12:51 sfan5     they are
12:52 sfan5     blocks sent to clients do not contain any additional compressed data
12:52 sfan5     so it doesn't matter
12:52 _Megaf    [13:47:53] <nrzkt> sfan5: why don't you test LZ4 ?
12:52 _Megaf    ^
12:52 _Megaf    +1
12:52 Krock     <sfan5>i have no time right now
12:52 nrzkt     lol
12:53 nrzkt     Calinou: they are already compressed, my question is to change the compressed mode for sending blocks
12:53 _Megaf    wait, blocks are decompressed client side?
12:53 nrzkt     server load block decompress, sent it compressed to client, client decompress yes
12:53 nrzkt     it's normal
12:53 _Megaf    maybe it would be interesting to have an option for server and client about that.
12:53 sfan5     nrzkt: my changes obviously make it use brotli for server-client transfer too
12:54 nrzkt     no _Megaf because uncompressed mapblocks take many MB memory, BW is not cheap for everybody
12:54 nrzkt     sfan5: super, i think we should do this
12:56 sfan5     now i just need to find out why it wants to send brotli blocks to clients who don't support then (normal 0.4.14-dev)
12:56 sfan5     bbl
12:58 sfan5     (i probably need to skip ser ver 26 entirely)
12:58 _Megaf    so it would be very interesting if indeed we could use brotli or lz4 for eveyrhting, compress the store map and transfer
14:58 sfan5     _Megaf: transfer already is compressed
15:10 _Megaf    sfan5: I understand, what I mean is compressing everything that is compressed with the same lib,
17:47 Hijiri    anyone want to seal the deal with another review of #4409?
17:47 ShadowBot https://github.com/minetest/minetest/issues/4409 -- Attached Particle Spawners by raymoo
19:08 paramat   since i'm about to merge an update to lua_api.txt perhaps i'll also finally document 'nodeupdate(pos) and nodeupdate_single(pos)'
20:00 Megaf     paramat: Hello
20:00 Megaf     paramat: if you wish, you can delete your comments about the code review
20:00 Megaf     I will delete my replies to your posts
20:02 Megaf     paramat: https://github.com/minetest/minetest_game/issues/1311#issuecomment-252115384
20:04 paramat   yeah i will
20:04 sfan5     lz4's api is absolutely horrible
20:05 sfan5     they expect you to keep the last 64KB of something(?) in memory
20:05 Megaf     Oo
20:05 sfan5     it says "blocks"
20:05 sfan5     i have no idea what that is
20:05 sfan5     also it doesn't say what this function returns
20:06 sfan5     well i can't code it like this
21:44 paramat   perhaps i'll just document 'nodeupdate_single(pos)' for now, because of #4594
21:44 ShadowBot https://github.com/minetest/minetest/issues/4594 -- Interesting bug with nodeupdate(pos)
22:13 paramat   updated #4589 still ok to merge? added nodeupdate_single(pos) docs
22:13 ShadowBot https://github.com/minetest/minetest/issues/4589 -- Lua_api.txt: Add biome and nodeupdate documentation by paramat
22:57 paramat   +1 for #4330 i'll merge this too soon
22:57 ShadowBot https://github.com/minetest/minetest/issues/4330 -- Textures: Ignore unknown node in override.txt by SmallJoker
23:04 paramat   will merge 4330 4589 in a moment
23:06 T4im      Hm, just a thought; shouldn't this be moved to minetest.nodeupdate_single or perhaps falling.nodeupdate_single?
23:07 T4im      if this becomes public api i mean
23:10 T4im      and, can of course be done in an extra PR, i guess, so nvm
23:11 paramat   indeed the issue discusses this
23:12 paramat   i won't close the issue but will change it's title
23:12 paramat   merging
23:16 paramat   complete
23:22 paramat   hi hmmmm i'm wondering if you are willing to complete #4530 before release? gregorycu is taking a break from / has left MT dev
23:22 ShadowBot https://github.com/minetest/minetest/issues/4530 -- Fix superflous shader setting updates by gregorycu
23:53 PR4530    works nice already in current state for almost 3 weeks of testing (calendar, playing each day)
23:54 paramat   heh the PR speaks