Minetest logo

IRC log for #minetest-dev, 2017-04-08

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:27 paramat joined #minetest-dev
01:48 Miner_48er joined #minetest-dev
01:48 ssieb joined #minetest-dev
02:18 Puka joined #minetest-dev
03:02 sofar #  define fips_cipher_abort(alg) while(0)
03:02 sofar huh
03:02 sofar util/sha256.c:36:0: warning: "fips_cipher_abort" redefined
03:02 nore wtf
03:02 sofar that's new
03:02 sofar util/sha2.h:85:0: note: this is the location of the previous definition
03:02 sofar did that get patched recently?
03:14 thePalindrome Could I get somebody to comment on game#1690
03:14 ShadowBot https://github.com/minetest​/minetest_game/issues/1690 -- Screwdriver can't rotate blocks not "facedir"
03:16 sofar make a PR?
03:16 sofar it seems you have the answer right there
03:16 thePalindrome Alrighty, figured it'd be easier for people to apply as a one liner
03:16 thePalindrome Bit out of touch :P
03:16 sofar we still would have to see a PR first
03:17 Tmanyo joined #minetest-dev
03:18 thePalindrome shh, too sleepy for thought :P
03:20 thePalindrome Oh fun, there's been some changes to screwdriver in git :P
03:23 sofar pretty big changes, yes
03:26 thePalindrome game#1691
03:26 ShadowBot https://github.com/minetest​/minetest_game/issues/1691 -- Allow non-native rotations if supported by targeted node. by thePalindrome
03:26 thePalindrome It still had the bug I noted, and I fixed it :P
03:28 sofar did ya test yer own code?
03:35 thePalindrome Doh, forgot an extra check :P
03:58 sofar you borked indentation style
03:58 sofar also, please squash your commits
03:58 sofar 4 commits for 5 lines of text is... yucky
03:59 thePalindrome Yeah, doing that now
03:59 thePalindrome Just wanted to make sure it passed CI :P
03:59 sofar you're also assigning a value twice to new_param2 now
04:00 sofar and the common case is to do it twice :(
04:00 sofar just do
04:00 thePalindrome The problem is I can't call fn
04:00 sofar `local new_param2`
04:00 sofar if fn then
04:00 sofar new_param2 =
04:00 sofar else
04:00 thePalindrome doi
04:00 sofar new_param2 = node.param2
04:00 sofar should be just as fast for the uncommon case
04:00 sofar and faster for the common case
04:01 thePalindrome Only slightly, since I'm not calling a method, but sure
04:01 sofar local new_param2 = fn and fn(pos, node, mode) or node.param2
04:01 sofar but that's terrible to read
04:02 thePalindrome aye
04:03 sofar do you know how to squash?
04:05 thePalindrome Found some tips on doing so, it's been far too long ;P
04:06 sofar are you using the git console client?
04:06 sofar linux?
04:06 thePalindrome yup
04:06 sofar ok so
04:07 sofar remember this one
04:07 sofar `git rebase -i HEAD~10`
04:07 thePalindrome I used the wrong advice first
04:07 sofar (10 is 10 commits back)
04:07 thePalindrome That's causing issues, since I can't push nor pull
04:07 thePalindrome oh wait, I can force fast forward
04:07 thePalindrome doy
04:07 sofar it'll give you the interactive rebase tool
04:07 sofar which is really simple to work with
04:07 sofar you can squash, fixup, drop and reorder commits in there really easy
04:09 thePalindrome Alrighty, squashed
04:10 sofar then `git push -f yourownremotename`
04:11 sofar try not to push to your own master branch
04:11 thePalindrome I realized I had to force the push, because the first advice suggested soft resetting head
04:11 thePalindrome Aye
04:11 thePalindrome I wasn't expecting more than a few lines, my bad
04:11 sofar if this is your only PR, sure
04:11 sofar but once you do more PRs, you're gonna need branches for each of the PRs
04:11 sofar ok I see 1 commit now
04:11 sofar good
04:11 sofar ya still borken da whitespace
04:12 sofar we use tabs
04:12 sofar edit
04:12 sofar git commit --amend mods/screwdriver/init.lua
04:12 sofar git push -f remotename
04:12 sofar also
04:12 sofar while you're at it
04:12 sofar break up the commit subject and message
04:13 sofar don't do a line over 80 characters for the subject if you can
04:13 sofar "Allow rotation for any node with on_rotate`
04:13 sofar ""
04:13 sofar "this also allows .... etc."
04:14 thePalindrome Had to look up how to do that
04:14 sofar 12 years of working with git
04:14 sofar you can ask me
04:15 sofar I used to train people how to use git
04:16 thePalindrome From all my editors the whitespace is the same, but github disagrees
04:16 * thePalindrome isn't used to dealing with conflicting whitespace
04:16 sofar the whitespace looks the same
04:16 sofar because you have tabs set to 4 spaces
04:16 sofar but you need to use tab characters, not spaces
04:17 sofar IOW, you may have to hire a few strong guys to talk to your editor
04:18 thePalindrome Honestly, it's more my fault for not understanding the configuration :P
04:21 thePalindrome Alrighty, that should be the last one
04:27 octacian_ joined #minetest-dev
04:29 VanessaE HEAD is bork3d again
04:30 VanessaE compile failure, that is..  http://pastebin.ubuntu.com/24338477/
04:31 calculon joined #minetest-dev
04:33 thePalindrome Phew, for a minute I thought I did that :P
04:33 VanessaE trying to find the last good commit.
04:35 VanessaE ok, it's an incompatibility between luajit (pulled to HEAD 5 mins ago) and minetest.  doesn't build at 94358a70 either (and I know that commit is good)
04:35 thePalindrome Interesting
04:39 VanessaE Minetest HEAD (ff4fef57) builds okay when I roll LuaJIT back to f50bf758 (where I had it this morning).
04:40 * VanessaE sighs and bisects
04:42 thePalindrome Isn't that fun :D
04:42 VanessaE at least I have a script to automate the build.
04:46 VanessaE dc320ca7 [...] is the first bad commit.
04:46 VanessaE Remove old Lua 5.0 compatibility defines.
04:50 VanessaE confirmed clean build with luajit at commit c29afcb1
04:54 VanessaE so, is this luajit's fault, or mt's fault?
04:55 MrIbby joined #minetest-dev
04:56 VanessaE oh G*D how I hate the method the luajit guys use to browse a repo.
04:58 VanessaE http://repo.or.cz/luajit-2.0.git/blobdiff/c29af​cb1a3f633fd970cf893d261d7f53728bc69..dc320ca70f​2c5bb3977b82853bcee6dad2523d01:/src/lauxlib.h  note the second to last line in the diff.
04:58 VanessaE oh, it's on github too.  good.
05:44 Hunterz joined #minetest-dev
06:15 nerzhul joined #minetest-dev
06:18 nerzhul merging trivial #5540
06:18 ShadowBot https://github.com/minetest/minetest/issues/5540 -- [CSM] Add event on_connect player API lua by Dumbeldor
06:24 VanessaE nerzhul: it ain't CSM related, but you may wish to look at the logs (ending an hour an a half ago)
06:26 nerzhul i'm eating my breakfast with my wife will looking for this later
06:26 VanessaE ok.
06:26 VanessaE bon apetit :)
06:26 nerzhul ty
06:39 nerzhul VanessaE, what distro do you use ?
06:39 nerzhul and what are you non standard cmake compilation options ?
06:39 VanessaE Debian sid.
06:40 VanessaE and this is out I build, http://pastebin.ubuntu.com/24338461/
06:43 VanessaE is how*
06:44 VanessaE or wait, the server is debian sta[b]le.  sid on my home PC.
06:45 nerzhul then problem is using a custom LuaJIT build (non packaged i mean), right ?
06:46 VanessaE correct
06:46 VanessaE specifically, something they did in LuaJIT broke Minetest
06:47 VanessaE looks like MT is relying on a Lua 5.0 symbol
06:47 VanessaE (or luajit devs screwed up)
06:47 nerzhul oh it's a LuaJIT problem not a MT problem ? for the symbol if i remember we can use lua 5.3 if needed
06:47 nerzhul then for symbols... strange
06:47 VanessaE It's a "I don't know which project is at fault" problem :)
06:48 VanessaE i.e. is MT at fault for using an outdated symbol?  or is luajit at fault because they declared a symbol outdated that really isn't?
06:48 nerzhul then they break compat with MT, it's not a good thing, as you said it's their fault
06:48 nerzhul luaL_reg is standard
06:49 nerzhul https://www.lua.org/manual/5.3/manual.html just search luaL_reg here
06:49 Krock joined #minetest-dev
06:49 Krock joined #minetest-dev
06:49 VanessaE ok
06:49 VanessaE I'll post a line note over there.
06:55 VanessaE wait a second..
06:56 VanessaE nope, it's Minetest's fault
06:56 VanessaE both luaL_reg and luaL_Reg (cap/lower 'r') are being used in MT
06:57 VanessaE src/script/lua_api/<several files> match luaL_reg (lower 'r')
06:58 VanessaE (the lua 5.3 manual you pointed me to specifies luaL_Reg, capital R)
06:59 VanessaE (as does the 5.1 manual)
07:03 nerzhul oh it's very subtile
07:04 VanessaE (whoever decided function names, symbols, etc ought to be case-sensitive should be shot :P )
07:04 nerzhul #define luaL_regluaL_Reg
07:04 nerzhul on my PC
07:05 nerzhul same thing in our Lua 5.1 bundle
07:05 nerzhul then the lowercase seems deprecated
07:05 VanessaE yes
07:05 nerzhul i will provide a patch for this then
07:06 VanessaE ok.
07:12 nerzhul okay it's just 31 occurs, i sent the trivial PR to CI and merge it after CI pass
07:12 VanessaE ok
07:14 nerzhul VanessaE, can you try https://patch-diff.githubusercontent.co​m/raw/minetest/minetest/pull/5541.patch ?
07:14 VanessaE moment..
07:16 VanessaE <xkcd "Compiling!">
07:18 VanessaE luajit at cced1786 and MT at 01f9d05f plus your fix, compiles clean.
07:18 * VanessaE gives it a quick run
07:20 VanessaE everything seems to be in order.  lgtm
07:22 nerzhul nice, it's a trivial fix, i'm just waiting for travis
07:23 VanessaE ok.
07:24 VanessaE hard to imagine my little potato here would be able to outrun travis.
07:28 nerzhul okay merging it, only vlangrind build is in progress and there is no functionnal change
07:28 VanessaE ok
07:48 DS-minetest joined #minetest-dev
08:42 nerzhul merging #5539
08:42 ShadowBot https://github.com/minetest/minetest/issues/5539 -- Document that write_json will error on unserializable types. by raymoo
08:43 nerzhul note: i pushed a non github file permitting to build minetest on gitlab.com on various distro, this will permit at the end (not now) to have package build pipelines for daily builds
09:14 QwertyDragon joined #minetest-dev
09:28 paramat joined #minetest-dev
09:55 red-001 could someone review #5528 ?
09:55 ShadowBot https://github.com/minetest/minetest/issues/5528 -- Move chat command handling code from c++ to lua. by red-001
10:03 nerzhul red-001, can you rebase please ?
10:04 nerzhul to fix the mac osx build
10:06 red-001 sure
10:08 red-001 done
10:28 kilbith joined #minetest-dev
10:59 proller__ joined #minetest-dev
11:00 paramat #5519 is now tested and ready for review. nore Zeno` and all
11:00 ShadowBot https://github.com/minetest/minetest/issues/5519 -- Sneak: Add option for old move code by paramat
11:04 mahfoud13 joined #minetest-dev
11:14 mogeid joined #minetest-dev
11:16 Zeno` I'm ok with it paramat. I would have preferred the old code to be in a different file but at this point *shrug*
11:16 Fixer joined #minetest-dev
11:19 paramat thanks for looking
11:53 sfan5 ~tell nerzhul what the fuck, you can't just push a pr without anyones approval whatsoever (#5542)
11:53 ShadowBot sfan5: O.K.
11:55 xunto joined #minetest-dev
12:06 Krock joined #minetest-dev
12:06 Krock joined #minetest-dev
12:07 YuGiOhJCJ joined #minetest-dev
12:10 red-001 #5544
12:10 ShadowBot https://github.com/minetest/minetest/issues/5544 -- [CSM] add `on_item_use` by red-001
12:22 Zeno` #5542
12:22 ShadowBot https://github.com/minetest/minetest/issues/5542 -- Prepare for Linux daily build packages (using Gitlab.com CI) by nerzhul
12:43 kilbith I'm rather scared of the stability since nrzkt is so pushy these days
12:44 sfan5 celeron55: can i get a clarification on the rule about 2 coredevs agreeing on a pr
12:44 sfan5 can these 2 agreeing coredevs include the person who proposed the pr?
12:46 Krock ^ should only be allowed for trivial changes IMO
12:47 sfan5 trivial changes can be pushed without any approval if #-dev is notified prior
12:48 Krock oh good then.
12:49 Krock sfan5, "hm why does this code work without this then?" what exactly do you mean with "this"?
12:50 Krock surely it works without the manual collision entry but then it would allow players to evade being damaged
12:50 Krock since Y speed can be set to 0 here
12:51 sfan5 in the original code i don't see m_speed
12:52 proller__ joined #minetest-dev
12:52 sfan5 set*
12:52 sfan5 which also isn't needed since the cbox of the player is on the node
12:52 Krock m_speed.Y = MYMAX(m_speed.Y, 0);
12:52 sfan5 (and thus he won't fall down)
12:52 Krock // legacy behaviour that sometimes causes some weird slow sinking
12:52 sfan5 setting m_speed is only needed if you want to replicate the old behaviour with the bigger sneak_max
12:53 Krock yes, and that's why it is there
12:53 sfan5 in your code it isn't inside an if() for the sneak_glitch override, is it?
12:54 blaze joined #minetest-dev
12:54 Krock ah right, also used for the new sneak collision boxes for flying in the air
12:54 sfan5 oh
12:54 sfan5 it can be useful to set m_speed to 0 in both cases (and in your case it's actually required)
12:55 sfan5 but to avoid any damage evasion i specifically avoided setting m_speed in the code
12:55 Krock yes I know. That's now solved by adding a seperate collision
12:55 Krock *  "solved"
12:55 Krock brb. eating
12:59 celeron55 i have always thought it is irrelevant who makes a PR
12:59 celeron55 and some have always thought otherwise
13:00 celeron55 i don't see a reason to change my viewpoint and they don't see a reason to change theirs :P
13:02 DS-minetest joined #minetest-dev
13:02 rubenwardy joined #minetest-dev
13:05 Krock re
13:12 sfan5 it's just that in the past most(?) coredevs have handled this rule as "2 other coredevs need to agree on the pr"
13:14 paramat the way we have been working for 1-2 years is that one approval can come from the author if they're a core dev
13:14 Krock okay, moved some code and made sneak ladders unreasonable fast again in the recent commit. Not quite happy about this change
13:15 rubenwardy I'd say it depends on the PR complexity
13:15 rubenwardy If it's a refactor, then 2 others
13:16 rubenwardy if it's a small/medium modification in an area which isn't used much, and you've tested well, then 1 other is fine
13:16 zorman2000 joined #minetest-dev
13:16 rubenwardy however, I tend to err on caution and wait for 2 others
13:17 zorman2000 Hi all, sorry to bother you, I have an issue with minetest.find_nodes_in_area(). Is this the right channel to ask?
13:18 rubenwardy #minetest would be better
13:18 zorman2000 Ok, thanks rubenwardy
13:18 paramat however, confusingly, a core-dev author's own 'approval' is not counted in the 'one/two approvals' labels. so a PR by a core dev with 'one approval' actually has the necessary 2
13:18 sfan5 Krock: i'm also not happy with supporting sneak ladders generally but it is what it takes to get people to stop complaining
13:18 Krock but it's quite hard to decide what changes would require two other approvals
13:19 Krock sfan5, from what I've seen there's a pull to disable it by default.. I hoped to find a midway solution by slowing it down
13:24 proller__ joined #minetest-dev
13:24 paramat yes i will make a PR before release to make the ladder option in new code disabled by default
13:27 Zeno` 1 approval is ok if another dev made the PR
13:27 Zeno` any other way and it's a bit silly.
13:27 paramat there's also #5527 which makes 2-node sneak jump work but also manages to make simple sneak ladder work
13:27 ShadowBot https://github.com/minetest/minetest/issues/5527 -- Sneak glitch: Detect ledge for 2-node climb-up by paramat
13:28 Zeno` there are also the "subsystem maintainer" rules
13:28 rubenwardy deprecated
13:28 Zeno` I don't know why people want to forget these things
13:28 rubenwardy http://dev.minetest.net/Organisation
13:28 rubenwardy "2015-11-02: The subsystem maintainer method was officially taken out of use."
13:28 Zeno` yeah that's fine
13:29 Zeno` but the core dev PR only needing 1 other approval needs to stay
13:29 Zeno` it's always been that way
13:29 Zeno` and it's always worked well
13:30 paramat yes i would be concerned about the speed of dev if 2 'others' were required
13:46 red-001 could someone review #5492?
13:46 ShadowBot https://github.com/minetest/minetest/issues/5492 -- [CSM] Add local particles and particlespawners. by red-001
13:47 MrIbby joined #minetest-dev
13:51 celeron55 you can always err on caution, nothing forces you to merge at 2 approvals
13:56 rubenwardy left #minetest-dev
13:56 rubenwardy joined #minetest-dev
13:56 Krock red-001, there are still various code style errors, starting with "ParticleManager::stepSpawnersLocal"
13:58 Krock seems like my previous comment was marked as outdated
14:05 Icedream joined #minetest-dev
14:11 behalebabo joined #minetest-dev
14:12 red-001 Krock, whats the code style issue with it?
14:13 Krock http://dev.minetest.net/Code_style_guidelines
14:13 Krock for (<condition>) {
14:13 Krock } else {
14:13 Krock ^ in about that
14:13 red-001 the braces?
14:13 Krock braces, spaces and new lines
14:15 red-001 oh https://github.com/minetest/minetest/pull/5492/c​ommits/cc1fcad6409f6812af9ae77b512fa5f4c8f9e9ca must have been the reason it got marked as outdated
14:20 Krock maybe, yeah.
14:27 proller__ joined #minetest-dev
14:30 red-001 Krock, are there any other code sytle issues?
14:30 red-001 style*
14:33 Krock why do you always increase the particle iterator manually inside the loop?
14:33 Krock example: https://github.com/minetest/minetest/pull/5492/​files#diff-e645b6869505fb7c2f21eed5098a56e4R501
14:37 red-001 copied existing code
14:39 Krock bad code :<
14:51 psedlak joined #minetest-dev
14:52 Megaf joined #minetest-dev
15:19 Megaf joined #minetest-dev
16:03 proller__ joined #minetest-dev
16:35 proller__ joined #minetest-dev
16:39 Krock rubenwardy, what kind of "problems" does it cause? (Topic: Store lagecy metadata pull)
16:40 rubenwardy see code snippet
16:41 rubenwardy https://github.com/minetest/minetes​t/pull/5531#issuecomment-292666198
16:45 numzero joined #minetest-dev
16:52 Krock well, one of the values will always be overwritten.
16:53 Krock as much I can see the easiest "solution" for your snippet would be changing the behaviour of "c_content.cpp"'s "read_item" function
16:53 Krock where the metadata is deserialized AFTER reading the old metadata
16:54 Krock which could/does overwrite the legacy metadata
16:55 Puka_ joined #minetest-dev
16:55 Krock this should have the same effect as your entire PR
16:56 Gael-de-Sailly joined #minetest-dev
16:58 Krock (  ^ rubenwardy  )
16:59 Puka joined #minetest-dev
17:02 numzero finally I understood the problem
17:03 numzero it’s that a single value is made available in 2 places by ItemStack:to_table()
17:03 numzero and ItemStack(table) have to decide which one to use, if they differ
17:03 numzero correct?
17:06 Krock yep, that's the entire trouble
17:06 rubenwardy yeah
17:08 rubenwardy Krock's solution is option 1 here: https://github.com/minetest/minetes​t/pull/5531#issuecomment-292668530
17:09 * numzero thinks
17:11 numzero well, in theory it’s possible to expose the whole ItemStack, including its meta, as a flat table
17:12 numzero so that there would be 3 special fields—"name", "count" and "wear"—while others would form the meta
17:13 numzero sounds a bit weird, but...
17:14 rubenwardy Just realised that stack:to_table() != stack:get_meta():to_table(), therefore you could modify stack:to_table() to not export the "" in the meta part, but just as `metadata`
17:15 numzero possible too
17:18 numzero actually that seems to be the shortest solution
17:21 kilbith I hope that PR has-been triple-checked before merging
17:23 rubenwardy #5546
17:23 ShadowBot https://github.com/minetest/minetest/issues/5546 -- Fix inability to change metadata using stack:to_table() by rubenwardy
17:23 rubenwardy compiling and testing now
17:25 numzero #5547
17:25 ShadowBot https://github.com/minetest/minetest/issues/5547 -- Item metadata compatibility by numberZero
17:26 numzero my is smaller :p
17:26 rubenwardy ew continue :)
17:27 Krock the pull recycling activists start crying now
17:29 numzero I hope all these PRs will be closed very soon
17:30 rubenwardy I don't care too much, tbh, closed mine
17:30 Krock sure, can _close_ them anytime. But they stay open to be reviewed and in the best case, merged.
17:30 rubenwardy Please may you use an imperative commit message? Ie: Fix ...
17:31 rubenwardy You'll notice that all commits apart from Nerzhul's do this https://github.com/minetes​t/minetest/commits/master
17:31 rubenwardy :)
17:32 Krock also a slightly more informative description wouldn't be too bad.
17:32 rubenwardy yeah, that too
17:36 numzero copied your description
17:53 Krock merging #5528 in 10 minutes
17:53 ShadowBot https://github.com/minetest/minetest/issues/5528 -- Move chat command handling code from c++ to lua. by red-001
18:16 Grandolf joined #minetest-dev
18:28 Krock red-001, your most recent, change reverting commit makes it worse again >> PANIC <<
18:28 red-001 I was getting segfaults due to the change
18:31 red-001 I will try and fix the segfault and get rid of the messy code tomorrow
18:31 Krock ok
18:50 YuGiOhJCJ joined #minetest-dev
18:55 Remy joined #minetest-dev
19:01 Remy_ joined #minetest-dev
19:02 Remy_ Hi ! I'm running some benchmark on sqlite3 storage. Have you got some protips ?  I want to modify some PRAGMA setting, but it seems that it's static in the code (https://github.com/minetest/minetest/b​lob/b88595050f3af5ccac06aac331ead4ebdc​b9deb9/src/database-sqlite3.cpp#L169) ... It will be posible to config that parameters in the futur ?
19:17 froike_ joined #minetest-dev
19:21 Krock Remy_, the closest pull request in that direction is https://github.com/minetest/minetest/pull/4420
19:22 Remy_ Thanks
19:22 Krock !next
19:22 ShadowBot Another satisfied customer. Next!
19:46 proller__ joined #minetest-dev
20:12 kilbith_ joined #minetest-dev
20:35 numzero joined #minetest-dev
20:50 proller__ joined #minetest-dev
20:52 red-001 is there anyway to make sure gettext knows about strings in lua files?
20:52 red-001 I see there is a script for config files but is there one for lua files?
20:56 red-001 nvm
21:18 numzero joined #minetest-dev
21:34 Megaf_ joined #minetest-dev
21:40 red-001 why is oldcoder trying to claim the trademark?
21:41 nore yeah, I'd like to know more about that as well
21:42 red-001 also I though prior use was ment to stop that
21:42 red-001 meant*
21:42 celeron55 it should stop it, but you can always lie and make a mess i guess
21:43 red-001 well if oldcoder wants to waste everyones time and money he is welcome to
21:44 celeron55 if someone in the US would like to spend their time making a trademark claim, /msg me; i don't expect anyone to want to, though
21:46 GreenDimond joined #minetest-dev
21:51 numzero joined #minetest-dev
21:52 lisac joined #minetest-dev
22:03 thePalindrome Knee-jerk reactions here: but thoughts on allowing SimpleSoundSpec to specify a pitch?
22:03 thePalindrome OpenAL supports it, and in my head older clients and servers would just ignore that field
22:16 lisac joined #minetest-dev
23:03 red-005 left #minetest-dev
23:08 MrIbby joined #minetest-dev
23:30 proller__ joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext