Minetest logo

IRC log for #minetest-dev, 2015-10-24

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

All times shown according to UTC.

Time Nick Message
00:56 crazyR joined #minetest-dev
01:05 est31 joined #minetest-dev
01:06 est31 cheapie, can you test around a bit and perhaps make a pr?
01:06 est31 it will slow down entering stuff into the db
01:06 est31 so sort of a trade-off here
01:11 est31 hrmm
01:11 est31 is rollback enabled by default?
01:12 est31 perhaps this could be a reason to change the default to disabled
01:21 cib0 joined #minetest-dev
02:50 cib0 joined #minetest-dev
03:15 leat joined #minetest-dev
03:16 crazyR joined #minetest-dev
03:24 leat joined #minetest-dev
03:50 hmmmm est, https://github.com/minetest/minetest/pull/3281 ??
04:59 cib0 joined #minetest-dev
05:08 Miner_48er joined #minetest-dev
05:39 est31 joined #minetest-dev
05:39 est31 hmmmm, seen my comments on your PR?
05:45 est31 also I don't see why its neccessary to remove the cleanup method
05:45 est31 and integrate it into the places it gets invoked
05:47 est31 but except for my comments the PR looks good
05:49 est31 (I'm ok with removing the cleanup method, thats more of a nitpick)
06:01 hmmmm the cleanup method was wrong
06:01 hmmmm resources need to be freed after the first join()
06:05 est31 okay then
06:14 est31 well, then the only remaining point is about that m_joinable variable
06:15 est31 in the current way it is used, it doesn't seem to be needed
06:15 leat joined #minetest-dev
06:15 est31 the mutex already takes care of two threads trying to join simultaneously
06:15 est31 and m_running takes care of join being called multiple times
06:16 est31 so what you do, to remove the m_running check, and add a m_joinable check isn't neccessary
06:16 est31 and makes the class more complicated than needed
06:16 est31 we can have m_joinable once we have compare and swap
06:17 est31 then we can remove the mutex as well
06:17 est31 but until then I dont think it should be there.
06:19 est31 if that's fixed, I think we can merge the PR. If it still doesn't make it build for Krock he can make a PR.
06:20 est31 and the doc comment is only a nitpick
06:24 hmmmm again: m_joinable is NOT for that purpose
06:25 est31 what is "that" purpose?
06:25 hmmmm m_running == false does NOT imply m_joinable == false
06:25 hmmmm did you read the unit tests?
06:25 hmmmm the thread can exit on its own
06:25 est31 no
06:25 hmmmm m_running == false but m_joinable == true
06:25 est31 ah
06:25 hmmmm you still need to join that thread in order to free the resources
06:25 hmmmm in addition...
06:25 hmmmm m_running == false does not mean the thread is not still running
06:26 est31 why do you need to join a thread in order to free its resources?
06:26 hmmmm it means that it's not in its main run loop any longer, and that it may be starting up or shutting down
06:26 hmmmm because that's the way pthread works
06:26 est31 why cant you simply you know do delete m_thread
06:26 est31 or so
06:27 est31 well, then I approve
06:27 est31 +1
06:28 Hunterz joined #minetest-dev
06:33 est31 After you've merged it can you have a look at #3286 and https://github.com/est31/minet​est/tree/rollback_proper_flush ?
06:33 ShadowBot https://github.com/minetest/minetest/issues/3286 -- Improve rollback database indexing by cheapie
06:33 hmmmm hmm
06:33 hmmmm rollback is only sqlite so this seems okay
06:34 hmmmm wow great :-)  simple yet highly effective optimization
06:34 hmmmm 3286 has a +1 for me
06:35 hmmmm 86144d4 also looks very reasonable to me, +1 too
06:36 est31 okay pushing both
06:36 hmmmm make sure you check to see if your commit fixes any issues filed
06:36 hmmmm this sounds like a combo bugfix and enhancement
06:37 est31 for 3286 there is  #1613
06:37 ShadowBot https://github.com/minetest/minetest/issues/1613 -- /rollback and /rollback_check are horribly slow
06:38 est31 but I won't close it just jet
06:38 est31 yet*
06:38 hmmmm that sounds reasonable, but iqualfragile seems to be inactive for a while now
06:38 hmmmm if I were you I'd close it with a "feel free to reopen if this is still an issue"
06:39 est31 okay as well
06:39 hmmmm we don't want to forget that 1613 has been addressed
06:39 hmmmm I think the main problem is that RollbackManager logs too much stuff to be honest, that's by design
06:40 est31 otherwise I can't find an issue for my commit: https://github.com/minetest/minetest/issues?utf​8=%E2%9C%93&q=is%3Aissue+is%3Aopen+rollback
06:40 est31 yeh
06:40 est31 yeah*
06:40 hmmmm it might be worthwhile to have a feature where server admins can select specific things to roll back
06:40 hmmmm ?
06:40 est31 ?
06:41 hmmmm I don't know just thinking out loud
06:41 est31 we could perhaps add a rollbackthread, and isolate the computation into queues
06:41 hmmmm ehhh
06:41 hmmmm the problem is the volume
06:41 est31 then the server doesn't stop when sb does rollback computation
06:41 est31 and checks
06:42 hmmmm sb?
06:42 est31 somebody
06:42 hmmmm ah
06:42 hmmmm a rollback thread would be a helpful first step I think
06:42 est31 its not really interwined with other code, all the places it communicates you can queue-ify
06:42 hmmmm because like I said, rollback is primarily I/O bound
06:42 est31 yup
06:42 hmmmm it doesn't make sense to me to block on posting rollback events
06:43 hmmmm or wait
06:43 hmmmm this needs to be looked at first, maybe it only flushes to DB every X number of events which somewhat mitigates this
06:43 est31 yes
06:43 est31 that is precisely what happens
06:43 est31 all 500 events
06:43 hmmmm oh
06:43 hmmmm not sure if it's really big enough to make a new thread for then
06:44 hmmmm adding a new thread is sort of a big deal imho
06:44 est31 and guess what, when the server shuts down, it forgets the last elements in the queue
06:44 est31 thats what my commit is about
06:44 est31 partl
06:44 est31 y
06:47 est31 wtf
06:47 est31 github is weird
06:47 est31 it writes "cheapie committed for est31"
06:50 hmmmm they must have just changed that bit of text today
06:51 hmmmm it doesn't make as much sense as the previous sentence
06:56 Hunterz est31: I tested yours ncurses console, looks like works fine, but after some time (idk hours, days) server runs, but when type any command to console server does not respond to commands
06:58 est31 okay
06:58 est31 can you still type
06:58 Hunterz yes
06:59 est31 so that when you press "a", it shows you "a" on the console?
06:59 Hunterz server is now without console, but when I remember, not show me what I typed
07:00 est31 ok
07:00 Hunterz after press enter
07:00 est31 you sure you haven't pressed ctrl + s inside the console window?
07:01 Hunterz running mt in the screen, when leave pressing ctr +a then s
07:03 Hunterz sorry ctrl +a then d
07:04 est31 hrmm
07:04 est31 the server is still responsive, so you still can log in via normal client?
07:05 Hunterz yes players can play
07:06 Hunterz now srtared server with --terminal I will report you soon when problem come again
07:06 Hunterz started
07:10 Hunterz when every line starts with <1> - this is some kinf of ncurses format ?
07:12 est31 no
07:12 est31 its the log level
07:12 est31 its just not printed nicely :)
07:12 est31 waited for SN's logging PR before touching that
07:12 Hunterz ah ok
07:13 est31 the logging PR has an extra method for it
07:13 hmmmm SN's logging pr is sort of poorly structured
07:14 hmmmm do you notice how there's logToSystem and logToOutputs
07:14 hmmmm ...is the system not an output?!
07:15 hmmmm so there should be a SystemLogBuffer in addition to the LogBuffer/RawLogBuffer and then I guess the ncurses console output gets a new kind of Buffer
07:15 hmmmm NCursesBuffer ??
07:16 est31 ewww
07:16 est31 he removed functionality I have used before :(
07:16 hmmmm yeah I don't know man
07:16 hmmmm SN's "refactorings" don't always seem to be 100% helpful
07:16 hmmmm I have mixed feelings on merging any more of them
07:17 cib0 joined #minetest-dev
07:19 est31 ncurses manages its own buffer
07:32 Calinou joined #minetest-dev
07:33 est31 (at least my pr does)
07:33 est31 ermm branch
07:51 nrzkt joined #minetest-dev
07:53 Krock joined #minetest-dev
08:05 proller joined #minetest-dev
08:09 est31 okay
08:10 est31 I do a PR with the changes I think are neccessary for the ncurses PR
08:23 DFeniks joined #minetest-dev
08:41 proller joined #minetest-dev
09:04 eeew joined #minetest-dev
09:21 Megaf joined #minetest-dev
09:37 ElectronLibre joined #minetest-dev
10:00 rubenwardy joined #minetest-dev
10:10 BlockMen joined #minetest-dev
10:23 est31 joined #minetest-dev
10:29 BlockMen can i push #3283 ?
10:29 ShadowBot https://github.com/minetest/minetest/issues/3283 -- Fix on_rightclick() being called directly after placing node by BlockMen
10:30 est31 have you tested it?
10:30 est31 if yes then yes
10:30 VanessaE return BlockMen.has_tested
10:30 VanessaE :P
10:31 BlockMen "true"
10:31 est31 answer("can i push #3283 ?", "yes")
10:31 ShadowBot https://github.com/minetest/minetest/issues/3283 -- Fix on_rightclick() being called directly after placing node by BlockMen
10:33 VanessaE I can reproduce that bug also, fwiw.
10:33 crazyR joined #minetest-dev
10:34 BlockMen ok, done
10:35 BlockMen any comments on #3209 and #3040?
10:35 ShadowBot https://github.com/minetest/minetest/issues/3209 -- Disable backface culling for models; fixes #2984 by BlockMen
10:35 ShadowBot https://github.com/minetest/minetest/issues/3040 -- Fix jittering sounds on entities (fixes #2974) by BlockMen
10:35 BlockMen is there a specific reason its not merged yet?
10:36 est31 no
10:36 est31 the backface culling has the issues pilzadam pointed out I guess
10:37 est31 most times you are not inside other players
10:37 est31 but you want fps to be high :9
10:37 est31 s/9/\)/
10:37 BlockMen by theory there is an impact, i was not able to notice any.
10:38 BlockMen second it influences all types of objects, mobs aswell
10:38 est31 and dropped items
10:39 BlockMen third we want be compatible to mc skins, so this is needed for
10:39 BlockMen yes
10:39 est31 there its only needed for players
10:39 BlockMen mobs aswell, e.g. ghosts from creatures look much better
10:40 est31 what about making it a flag as you proposed?
10:40 est31 but default culling on?
10:40 est31 (and culling off for players)
10:40 BlockMen fine for me
10:43 est31 and I agree with pilzadam its better to have whitespace fixes in a different commit
10:47 BlockMen then we should add this to rules, as stated there aswell ;)
10:48 est31 seems reasonable
10:53 BlockMen ok, will add it to dev wiki then and wont mix it next time
10:54 BlockMen can i merge #3040 otherwise?
10:54 ShadowBot https://github.com/minetest/minetest/issues/3040 -- Fix jittering sounds on entities (fixes #2974) by BlockMen
10:57 crazyR joined #minetest-dev
11:00 est31 what have you changed
11:00 BlockMen basically this: https://github.com/minetest/minetest/pull/3040/​files#diff-efb51cde4bb3c805a95dc31b2b1781d0R331
11:01 BlockMen it checks in collisoncheck whether you stand on object
11:01 BlockMen and stops playing the node sounds when standing on object instead of ground
11:01 est31 well, it might not be really proper, but at least fixes that bug
11:01 est31 later it can be found out how to fix the root case
11:02 est31 but better to split the code cleanup out
11:02 BlockMen really?
11:02 est31 it makes the commit much easier to read
11:04 BlockMen im hoestly not willing to fiddle around with it again
11:05 BlockMen not sure where to add this "dont mix" statement. here? http://dev.minetest.net/Git
11:05 Krock joined #minetest-dev
11:05 BlockMen or here http://dev.minetest.net/Git_Guidelin​es#Upstream_pull_requests_and_issues
11:05 est31 the first one is more of a help site
11:06 est31 the second more a rules page
11:06 est31 id take the second one
11:08 BlockMen added
11:12 CraigyDavi joined #minetest-dev
11:14 Darcidride joined #minetest-dev
11:25 Lunatrius` joined #minetest-dev
11:38 BlockMen bbl
11:51 Amaz joined #minetest-dev
11:58 crazyR joined #minetest-dev
12:15 DFeniks joined #minetest-dev
12:25 harrison can we talk about raytracing and voxels and stuff? please?
12:25 harrison after all it is the future of realtime graphics -- it just has to be!
13:07 Warr1024 joined #minetest-dev
13:38 younishd joined #minetest-dev
13:51 PilzAdam joined #minetest-dev
14:56 est31 joined #minetest-dev
15:07 zat joined #minetest-dev
15:22 VanessaE joined #minetest-dev
15:26 Player2 joined #minetest-dev
15:51 zat joined #minetest-dev
16:01 crazyR joined #minetest-dev
16:03 est31 joined #minetest-dev
16:09 est31 awww
16:09 est31 a subtle bug!
16:10 crazyR joined #minetest-dev
16:12 VanessaE a bug?  say it ain't so!
16:12 VanessaE :P
16:13 Calinou it's notabug.org, it's a feature
16:13 Calinou :)
16:13 est31 well until now rebasing my terminal was pretty easy
16:13 est31 basically rebase
16:13 est31 but thanks to threading AND logging changes...
16:13 est31 well I've asked for them
16:13 est31 die geister die ich rief...
16:14 est31 (to be merged)
16:18 Hunterz terminal console going to upstream ?
16:19 hmmmm joined #minetest-dev
16:23 crazyR joined #minetest-dev
16:23 est31 joined #minetest-dev
16:37 Gael-de-Sailly joined #minetest-dev
16:44 crazyR joined #minetest-dev
16:59 paramat joined #minetest-dev
17:03 paramat PilzAdam perhaps #3271 should be merged soon?
17:03 ShadowBot https://github.com/minetest/minetest/issues/3271 -- Lua settings improvements by PilzAdam
17:04 PilzAdam yes, I'm waiting for someone to tell me to do it
17:04 nrzkt PilzAdam, please squash
17:04 BlockMen joined #minetest-dev
17:06 VanessaE that PR is gonna be a bitch to squash properly.....
17:08 Krock hmmmm, what happened to 3281? Can I strikethrough that compiling request from my todo list?
17:09 hmmmm yeah, the threading compilation fix was on my list of things TODO with 3281
17:09 hmmmm I suppose you could rebase your PR to fix the gettext compilation too, but it's probably not worth it because I'm going to refactor that piece anyway
17:10 hmmmm like, that really should not be a platform-conditional function definition
17:10 hmmmm it hides problems exactly like what you just fixed
17:10 Krock Oh well, if you're going to change code there it's nonsense to rebase my pull request
17:10 hmmmm sorry :(
17:13 PilzAdam nrzkt, done
17:13 PilzAdam VanessaE, why? it worked fine for me
17:13 hmmmm yeah, so what is 3271 waiting on exactly?
17:13 VanessaE PilzAdam: well I wasn't counting on you squashing it into just one commit :)
17:13 hmmmm I'm not 100% sure I understand what that PR does so I'm withholding judgement
17:14 PilzAdam hmmmm, it does good things ;-)
17:14 VanessaE (to me, "proper" in this context would be... well basically split and merge them up according to the 9 items in your "Changes")
17:14 hmmmm Yeah, you know what, I really don't think this is something that should be squashed
17:14 VanessaE too late :P
17:14 hmmmm PilzAdam, are you able to revert back to the old history or is it too late?
17:15 PilzAdam I'm not going to sqash this into several commits
17:15 PilzAdam if someone is willing to invest some hours on this then feel free to
17:15 hmmmm I didn't see the actual commit log for this branch
17:15 hmmmm was it like...
17:15 VanessaE translation: "well it would indeed be a bitch to do if I did it like that"  ;)
17:15 hmmmm "Improve lua settings"
17:16 hmmmm "Fix this tab on this line"
17:16 hmmmm "Rename this variable"
17:16 hmmmm etc..
17:16 hmmmm or was it
17:16 hmmmm "Do one actual thing here"
17:16 hmmmm "Implement other piece of functionality here"
17:16 PilzAdam hmmmm, a bit of both
17:16 hmmmm the former of the two situations is where a PR's commit log should be squashed, the second not so much
17:17 hmmmm I really wish you people would start making more regular commits and not jamming it all into one huge opaque package
17:17 PilzAdam I would prefer that too, but other devs seem to prefer this
17:17 hmmmm listen to what makes more sense
17:18 Krock hmmmm, rebased the pull - as a fallback fix
17:19 hmmmm if people keep doing this, we're inevitably going to merge a big "refactor" PR that'll essentially be the minetest equivalent of the TTIP
17:19 PilzAdam so, paramat, what about merging #3271 now?
17:19 ShadowBot https://github.com/minetest/minetest/issues/3271 -- Lua settings improvements by PilzAdam
17:19 est31 joined #minetest-dev
17:20 PilzAdam hmmmm, btw, what 3271 does: 1) it expands descriptions on a bunch of settings 2) it adds new types and better input mechanisms 3) it gives mods / games the ability to define their own settingtypes.txt
17:20 PilzAdam that would sum it pretty much up
17:21 est31 my position on this is that at the time of merging, the history of a pr should be clearly separated in topics, and large prs (like the logging or threading pr by SN) should be split, but the history shouldn't reflect how the PR was created, but rather reflect the topics.
17:21 hmmmm well
17:21 est31 and I rather like a single big commit msg than 40 small "update to est's suggestions" "fix crash by previous commit" etc
17:21 hmmmm I guess this can be okay, I don't particularily like 2000 line change commits
17:22 hmmmm no doubt this PR has been brewing for quite some time
17:22 est31 I know however that it is hard to make a faked history so I tell people to squash because they'll do that rather than create a clean sane history
17:22 hmmmm but let it be a warning for next time - this gigantic, impossible-to-review, opaque PR stuff needs to stop
17:23 PilzAdam I personally would prefer a system where people use actual dev branches with proper commits, and we merge them with merge commits
17:23 hmmmm that has its own big problems
17:23 est31 I don't like that.
17:24 hmmmm I would argue the problems with that sort of workflow are worse than the problems of our current setup
17:24 est31 you have tons of "merge this merge that"
17:24 est31 a faked history is something more "higher" of a good as it allows you to see what a commit actually does
17:24 hmmmm and not only that, but commits show in the history in order of their timestamp, not when they were actually added
17:24 est31 and not just an entry in a changelog
17:26 est31 the project is to small both in code base as well as team size for such a change to be neccessary imo
17:27 hmmmm even with large teams/codebases I'd prefer this setup...
17:27 hmmmm otherwise it's just a total mess
17:27 est31 tbh I dont really have experiences with large codebases, so my opinion isnt very funded there
17:29 hmmmm init_gettext is a freaking wreck
17:29 hmmmm who wrote this initially
17:30 paramat PilzAdam the parts of 3271 i understand look good +1
17:30 hmmmm I recommend somebody refactor this to not have so many damn conditional compilation statements - this is nuts
17:32 hmmmm https://github.com/kwolekr/minetest/commit​/5b262785b2312e49c2faf222b0389b89ea1c79b0 PTAL
17:32 est31 I've looked at large parts of 3271 it seems good now
17:33 hmmmm note to everybody - stop doing the kind of crap I just removed in this commit - it is not helpful
17:34 est31 looks good
17:35 est31 PTAL #3290
17:35 ShadowBot https://github.com/minetest/minetest/issues/3290 -- Small logging refactor and additional options by est31
17:36 hmmmm paramat:  I gave you the thumbs up for 3284 btw
17:36 hmmmm did you see that?
17:36 paramat yeah thanks
17:38 Amaz joined #minetest-dev
17:38 PilzAdam paramat, est31, so merge now?
17:38 PilzAdam it's already rebased on master
17:38 paramat sure
17:38 est31 ok
17:40 PilzAdam done
17:41 paramat PilzAdam sfan5 game#713 for whenever you have time to consider
17:41 ShadowBot https://github.com/minetes​t/minetest_game/issues/713 -- Default/functions: ABM for moss growth on cobble near water by paramat
17:41 VanessaE paramat: ^^ you never addressed my comments about that
17:42 PilzAdam paramat, I like what it does, I like the code, but I haven't tested it
17:42 PilzAdam so if it's tested, then +1
17:42 paramat okay, yes it's tested
17:43 paramat VanessaE you would like checks for air and water?
17:43 PilzAdam sfan5, game#709 can now be merged
17:43 ShadowBot https://github.com/minetes​t/minetest_game/issues/709 -- Add settingtypes.txt by PilzAdam
17:44 VanessaE paramat: not necessarily - rather you should add something that will allow me to rewrite gloopblocks to totally eliminate its mossy feature entirely, i.e. register_mossy("stone", "gloopblocks:stone_mossy") or something similar
17:45 VanessaE (though having the target node being bound by both air and water would be a good idea.  I've seen what happens when it isn't :) )
17:54 est31 okay hmmmm updated my pr
17:54 hmmmm LGTM
17:56 PilzAdam Krock, can you rebase #3267 ?
17:56 ShadowBot https://github.com/minetest/minetest/issues/3267 -- Standardize the menu button order and sizes by SmallJoker
17:56 Krock <Krock>rebase time!
17:57 Krock trying to find optimal positions
17:57 Hunterz with #3290 ncurses console get timestamp right?
17:57 ShadowBot https://github.com/minetest/minetest/issues/3290 -- Small logging refactor and additional options by est31
17:58 est31 yes, its planned
17:58 Hunterz nice
17:59 paramat hm the 2nd abm in that commit could cause some 'buried' mossycobble, but the fix needs 1 or 2 extra scans, it might not be worth it. we could perhaps consider moss has penetrated into a thick cobble wall through cracks
17:59 paramat moss underwater seems okay
18:00 VanessaE moss deep underwater isn't
18:00 VanessaE in practice, it doesn't look good
18:02 Krock PilzAdam, done. Check it out.,
18:04 PilzAdam est31, you kinda handle translations currently, right?
18:04 est31 yup
18:04 PilzAdam can you run updatepo.sh soon-ish?
18:05 PilzAdam all the new setting comments need to be translated
18:05 est31 don't we want to fix typos in settingstypes.txt etc first?
18:05 Krock ^
18:05 est31 otherwise translations will get the string with the typo
18:05 est31 then its translated
18:05 est31 then the typo gets fixed
18:06 est31 then the translator has to look at it again
18:06 est31 at least to remove the fuzzy flag
18:07 PilzAdam true, but AFAIK there no typos left
18:07 PilzAdam but I guess / hope many people will make PRs to improve comments
18:08 PilzAdam Krock, I would expect the buttons to be aligned right, not in the center
18:08 PilzAdam est31, I guess translating the 80% of the comments that will stay the same is worth it, though
18:08 Krock arabic languages might disagree
18:09 PilzAdam do you want to run the positions through gettext?
18:09 est31 okay found another reason to stall it lol: can you have a look at #3230 so that it can be merged before updatepo runs ?
18:09 ShadowBot https://github.com/minetest/minetest/issues/3230 -- Better gettext support for protocol version mismatch messages by est31
18:10 PilzAdam est31, that code is absolutely unreadable
18:11 paramat central buttons look better to me
18:12 PilzAdam est31, can you split that up into seperate ifs?
18:13 PilzAdam so don't use "string" .. condition and "string"
18:13 PilzAdam rather use: var = "string"; if condition then var = var .. "string"
18:14 BlockMen #3267 why does it have to be the same location everywhere? In simple dialogs centered looks good but in the config dialog it looks displaced IMO
18:14 ShadowBot https://github.com/minetest/minetest/issues/3267 -- Standardize the menu button order and sizes by SmallJoker
18:16 PilzAdam BlockMen, true
18:16 Krock :<
18:17 est31 PilzAdam, updated
18:17 BlockMen est31, #3209 good now?
18:17 ShadowBot https://github.com/minetest/minetest/issues/3209 -- Add option to disable backface culling for models by BlockMen
18:17 PilzAdam Krock, the "start" buttons in singleplayer, client and server tab are right aligned
18:17 PilzAdam you should be consistent with that
18:18 PilzAdam est31, it's good now +1
18:18 Krock PilzAdam, that's logical because it's related to the selection box above
18:19 PilzAdam est31, you can merge that now
18:19 est31 okay
18:19 est31 thx
18:20 crazyR joined #minetest-dev
18:21 est31 okay found at least one typo in settingstypes.txt "Smooths"
18:22 est31 right version is "Smoothes"
18:22 est31 at least according to what I think
18:22 * est31 is native no speaker
18:22 Krock ^ is native writer
18:24 PilzAdam est31, rubenwardy had some remarks on some things, too
18:24 est31 should I run the updatepo, or should we wait till rubenwardy's requests are met?
18:25 PilzAdam I would run it now
18:25 PilzAdam (I want to test the weblate interface ;-))
18:26 est31 okidoki
18:28 est31 note that it takes the email address you use for sign up for the commits
18:28 est31 so if thats private, change it :)
18:29 PilzAdam I signed in via github
18:30 est31 then its only an issue if you have turned on githubs email hiding feature
18:30 est31 which it ignores as well :)
18:31 est31 so there it is, pushed
18:32 PilzAdam will weblate update automatically?
18:32 est31 yes
18:32 est31 idk whether there is a button for it
18:32 est31 ah "Pull"
18:32 est31 yup its in there
18:33 est31 loool
18:33 est31 the score just decreased for all strings
18:34 est31 for german, from ~100% to ~30%
18:34 PilzAdam that's why I asked to update the languages so early
18:35 est31 hrmm multiline msgids
18:36 est31 will have to make my vandalism checker tool ready for that
18:36 * est31 has to do so much things
18:36 est31 in so little time... :/
18:36 crazyR joined #minetest-dev
18:37 paramat left #minetest-dev
18:45 est31 PTAL https://github.com/est31/minetest/commit/​719f0fc56af81e171643e219dae54cbada623c6d
18:45 est31 I remove lotsa trailing whitespaces
18:45 est31 and fix that bug pointed out by @arpruss
18:45 est31 arpruss
18:47 est31 if you want, I can do the whitespace removing in a separate commit
18:48 est31 otherwise you can pass an argument to git diff to ignore whitespaces
18:48 est31 I think it was --whitespace=ignore
18:48 est31 okay better msg: https://github.com/est31/minetest/commit/​49bda7f98d2ca1d3553460d27c7308856969fb44
18:49 est31 if nobody objects I'll push it in 10 minutes
18:50 BlockMen est31, +1
19:31 PilzAdam est31, these are a lot of strings; I already work an hour on this and I'm only ~ 25% done
19:32 est31 wow
19:36 PilzAdam I have no idea how weblate sorts the strings
19:37 est31 me neither
19:42 PilzAdam I'm done for now
19:42 est31 weblate will automatically commit your changes
19:42 est31 and once a month or if there is another specific reason I look at weblate's repo and pull all the commits over
19:42 est31 including hand-reviewing them
19:43 est31 review mostly done bc vandalism
19:43 rubenwardy why is french translation locked?
19:43 est31 bc sb else translates right now
19:43 est31 by kilbith it seems
19:44 est31 (you have to hover over the lock icon)
19:44 rubenwardy I didn't know who Jean Patrick GI was
19:44 rubenwardy oh
19:44 rubenwardy I didn't know who Jean Patrick G was
19:48 PilzAdam I posted the translation stuff on the forums
19:57 PilzAdam rubenwardy, https://github.com/minetest/minetest​_game/pull/709#discussion_r42937687 I don't get it
19:59 rubenwardy If you're trying to configure a server using the settings, it'd say "true" when it actually becomes "false".
19:59 PilzAdam you would only see "true" if it's not set
19:59 rubenwardy https://github.com/minetest/minetest_​game/blob/master/mods/tnt/init.lua#L5
19:59 PilzAdam every other case is correct
19:59 Icedream joined #minetest-dev
20:00 nrzkt rubenwardy, it's kilbith yes
20:00 PilzAdam this is because I can't set a different default value based on arbitrary conditions
20:00 rubenwardy if it's not set, then you would see "true", but when you create a server it's "false.
20:00 hmmmm [02:47 PM] <est31> if you want, I can do the whitespace removing in a separate commit
20:00 hmmmm [02:48 PM] <est31> otherwise you can pass an argument to git diff to ignore whitespaces
20:00 hmmmm don't do this
20:00 rubenwardy It's a small issue though, not worth fixing
20:00 hmmmm we made a decision a long time ago, it's so much easier to just leave the whitespace fixes in with your commit
20:00 hmmmm just make sure you don't introduce any more, and the current whitespace errors are in the gradual process of being fixed
20:01 PilzAdam est31, what happens when one creates a new language in the weblate interface? is everything correctly set up?
20:01 hmmmm at some point I'm going to modify cpplint.py to suit our own style, which catches a lot of things
20:01 Gael-de-Sailly joined #minetest-dev
20:01 hmmmm and we'll make it a requirement to run this on your commit before merging anything
20:01 hmmmm hopefully this will catch a lot of the low-level dumb errors that we shouldn't be wasting human time on
20:02 est31 PilzAdam, I'm not sure, it somehow worked for some languages
20:02 est31 eg lobjan was added that way
20:02 PilzAdam est31, btw, remember to add new languages to the "language" enum in settingtypes.txt from now on
20:03 est31 hmmmm, you might want to have a look at #3142 then
20:03 ShadowBot https://github.com/minetest/minetest/issues/3142 -- Let travis check for git format mistakes by est31
20:04 hmmmm I suppose that's a good start, but it doesn't have very sharp teeth
20:04 est31 PilzAdam, can't that be automated?
20:04 est31 e.g. by having a lua callback get_language_list
20:04 PilzAdam that would require hacks when parsing settingtypes.txt
20:04 est31 and extra type for it
20:04 PilzAdam the current design is that everything is statically defined in the file
20:05 paramat joined #minetest-dev
20:05 PilzAdam I already thought about adding some "hack_" types to solve different issues, but I don't really like it
20:05 PilzAdam remember that this format is now part of our mod API
20:06 est31 ah yeah
20:07 est31 what about types that only we can use in the engine's settingtypes.txt?
20:07 PilzAdam I mean, it can be done and I won't object it; but I just have a not so good feeling about it
20:07 est31 I just know from experience that people won't update
20:07 rubenwardy how about a "mods" folder, where enable security is a child, as a place for mods to put their settings?
20:07 est31 I'll try to do it, but you know, on the long term...
20:08 est31 there is one
20:08 rubenwardy consider forcing mods to put all their settings in Server > Mods > Modname
20:08 paramat i'll merge #3284 soon, after checks
20:08 ShadowBot https://github.com/minetest/minetest/issues/3284 -- Mgfractal: Independent iterations and scale parameters by paramat
20:10 est31 hmmmm, so, can it be merged? With all the shit commits removed ofc
20:11 PilzAdam rubenwardy, currently the settingtypes.txt of mods and games is just appended to the list in the builtin settingtypes.txt
20:11 PilzAdam putting them in categories of the builtin stuff would require a certain format of it
20:11 rubenwardy I understand that
20:11 rubenwardy you can't add settings to an existing group?
20:11 rubenwardy only append?
20:11 est31 btw there is a problem
20:12 rubenwardy I guess that's fine :/
20:12 PilzAdam rubenwardy, the setting is a list, internally, not a tree
20:12 rubenwardy I see
20:12 PilzAdam that is easier when doing the formspec transformation
20:12 est31 a mod could add an option "Improve speed 10x" which actually points to enable mod security
20:12 est31 or disable
20:12 est31 whatever
20:12 est31 thats not good
20:13 PilzAdam est31, it can't add itself to the trusted_mods, though
20:13 rubenwardy but could turn it off?
20:13 PilzAdam we could add checks to forbid double-adding of settings
20:13 rubenwardy or adding of secure.*
20:13 est31 yea both seem reasonable
20:14 * PilzAdam starts coding
20:14 rubenwardy double adding may be useful if a setting is relevant in two sections. However, that isn't good design
20:15 est31 checking secure.* is problematic if its ever changed in the engine
20:15 rubenwardy why would it be changed?
20:15 est31 and forgotten to be changed in lua
20:15 est31 dunno, perhaps sb wants to shorten it to sec. ?
20:15 est31 or security. ?
20:16 est31 or some other reason
20:16 rubenwardy It's not just useful for security though
20:16 PilzAdam I would expect people to grep to all source files in this case
20:16 PilzAdam s/to/through/
20:16 rubenwardy also useful for URL whitelists, etc
20:16 rubenwardy exactly
20:16 est31 my command is usually grep -rIn "text" src
20:17 est31 I never scan . bc it includes the build directory
20:17 est31 and build includes lotsa stuff
20:17 paramat now merging 3284
20:17 est31 (not in the repo, but all the android libraries that get downloaded during android build)
20:17 est31 and ofc it doesnt get the symlink
20:19 est31 there is a symlink in build/android/jni/src/  pointing to the src folder
20:19 est31 so grepping in . will lead to the searched text appear twice
20:19 est31 and will be super slow bc it also scans irrlicht and friends
20:20 est31 yeah and mainmenu is in builtin/
20:21 paramat merge complete
20:22 est31 5090 commits
20:22 est31 thats 10 from 5100
20:22 est31 it was like yesterday when we had commit 5k
20:32 PilzAdam https://github.com/PilzAdam/minetest/commi​t/9ee0d376d4e4679be074ed3a27e9700169aecea3
20:33 est31 why is the diff so large?
20:33 PilzAdam :-/ this diff is kinda huge due to some simple restructuring
20:34 PilzAdam basically return directly when encountering an error, instead of continuing in nested ifs
20:39 paramat i have a request for hmmmm or anyone #2977 this is beyond my capability. much needed for floatland realms, alternative mapgens and travel by water column
20:39 ShadowBot https://github.com/minetest/minetest/issues/2977 -- Stop liquids from flowing over ignore. by red-001
20:42 PilzAdam est31, can I merge that soon or should I create a PR?
20:44 est31 looks good
20:44 est31 so good for merge
20:44 est31 if tested
20:45 PilzAdam ok, I'll wait a few minutes to give others a chance to say something, too
20:46 nrzkt #3287 if you want est31
20:46 ShadowBot https://github.com/minetest/minetest/issues/3287 -- Credits: Remove my name by Rui914
20:46 est31 yea seen it, I'm busy right now with rebasing my console
20:47 est31 getting the weird error that verbosestream works
20:47 est31 but actionstream doesn#t
20:47 est31 s/#/'/
20:51 PilzAdam est31, does it fail to compile or isn't it logging stuff?
20:51 est31 I made compiling work
20:51 est31 so that works again
20:52 est31 verbosestream works really fine
20:52 est31 added a log emit to each server step
20:52 est31 now trying to set the log level of actionstream to LL_VERBOSE
20:52 est31 yeey it works
20:53 est31 now replacing it with action again, and applying the hack at another position
20:54 PilzAdam pushing https://github.com/PilzAdam/minetest/commi​t/9ee0d376d4e4679be074ed3a27e9700169aecea3 now
21:04 est31 no
21:04 est31 dont say this is true
21:05 * est31 rebuilding to check with valgrind
21:05 Icedream joined #minetest-dev
21:11 PilzAdam oops
21:11 est31 how can you make valgrind check for uninitialized reads?
21:12 PilzAdam est31, https://github.com/PilzAdam/minetest/commi​t/7d5c7365314c9517369a7a7d6e4fc5d8712b93c7
21:12 PilzAdam memcheck does that default, AFAIK
21:12 PilzAdam +by
21:12 PilzAdam this commit fixes a bug in my last commit
21:13 est31 ok +1
21:13 est31 whats the difference
21:14 PilzAdam settings.current_comment is always nil in this place
21:14 PilzAdam while the local variable current_comments contains the actual comment
21:15 est31 hrmm isnt it "" ?
21:15 est31 https://github.com/PilzAdam/minetest/commi​t/7d5c7365314c9517369a7a7d6e4fc5d8712b93c7​#diff-b61e3e14dbe70423022f4ee065fa3f94L44
21:15 PilzAdam eh, yes
21:15 PilzAdam whatever
21:15 PilzAdam it's late
21:15 est31 yea 4 me too
21:16 est31 if that actionstream bug is fixed I update my branch and hope the best xD
21:16 PilzAdam you shouldn't be allowing me to push upstream after 22:00, btw
21:16 est31 meh, I do mistakes as well
21:17 est31 my specialty is making gramatically correct commit msgs with words swapped
21:17 est31 https://github.com/minetest/minetest/commi​t/49bda7f98d2ca1d3553460d27c7308856969fb44
21:17 est31 bad word is "statically"
21:18 est31 should be "dynamically"
21:19 PilzAdam that's basically the same
21:19 est31 man how can this happen
21:20 est31 I do know that m_silenced_levels is at the index of LL_ACTION set to true
21:20 est31 now I wonder where it is set to true
21:20 est31 but valgrind reports no uninitialized read
21:20 est31 and setting it to false has no effect
21:21 est31 so somehow it is initialized before first read
21:21 est31 and after my failed attempt to change it to false :)
21:21 est31 now the only place where that private member is set in class is setLevelSilenced
21:22 est31 but that method is never called outside testing
21:22 est31 okay next try
21:22 est31 perhaps volatile doesnt reserve memory
21:23 est31 and it shares memory with the map below
21:23 * est31 removes volatile
21:24 Hunterz joined #minetest-dev
21:27 paramat left #minetest-dev
21:51 crazyR joined #minetest-dev
21:52 PilzAdam est31, there is a bug where " are not escaped in the settings translation file
21:53 PilzAdam I will fix this tomorrow; the translation files need to be updated again, then
22:04 est31 okay, anybody here?
22:06 BlockMen wanted leave now
22:06 BlockMen y?
22:06 est31 can you look at https://github.com/est31/minetest/commit/​fae1eb9a1969194c61feb25312fdbb5d397c0d8a
22:07 zat joined #minetest-dev
22:08 est31 took me hours to find the cause.
22:11 BlockMen is the int conversion necessary?
22:11 BlockMen otherwise seems right, +1
22:11 est31 I think so
22:11 est31 oh
22:12 est31 not neccessary after all
22:12 est31 removing it
22:14 BlockMen ok, good. im going bed now. feel free to merge from my side
22:15 est31 fine
22:15 est31 I'll do that as well I think
22:22 domtron joined #minetest-dev
23:07 est31 joined #minetest-dev
23:07 est31 okay, there it is: #3292
23:07 ShadowBot https://github.com/minetest/minetest/issues/3292 -- Add server side ncurses terminal by est31
23:07 est31 still missing a polish, but ok for a first glimpse and an overview of general architecture
23:52 hmmmm lol
23:52 hmmmm est31:  nice catch
23:52 hmmmm I missed this too when reviewing it
23:55 hmmmm [04:10 PM] <est31> hmmmm, so, can it be merged? With all the shit commits removed ofc  -- huh?
23:56 hmmmm oh you mean squashed?
23:56 hmmmm yeah, see that's an example of a PR that needs squashing
23:57 hmmmm lol, https://github.com/est31/minetest/commit/​91e70b0a37c6dac86defdad7db56c97906149ac1
23:57 hmmmm inspired by Gerogerigegege

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