Minetest logo

IRC log for #minetest-dev, 2013-11-07

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

All times shown according to UTC.

Time Nick Message
00:05 iqualfragile_ joined #minetest-dev
00:07 ShadowNinja All C++ -> Lua calls other than script_run_callbacks and calls to table.insert now have a errfunc. The issue is that a LOT of C++ -> Lua calls are made by script_run_callbacks...
00:08 ShadowNinja If anyone can insert a errfunc there in a way that doesn't segfault or cause a
00:09 ShadowNinja ... Lua PANIC then feel free to tell me about it.
00:11 EvergreenTree joined #minetest-dev
00:12 kahrl maybe the whole script_run_callbacks shebang could be moved to lua
00:14 kahrl basically like it's done in minetest.on_craft
00:15 kahrl that would reduce the number of C -> Lua calls too and might be faster
00:23 ShadowNinja Yes, but that means more changes. :-| Although I could probably do it, if I can understand the function to write it in Lua and get the enum copied to Lua or something like that.
00:31 ShadowNinja kahrl: Do you have the ability to add a errfinc to script_run_callbacks? Tell me where on the stack to insert the function and when.
00:32 kahrl sorry not now
00:34 ShadowNinja Do you know who can? This is the only function left...
00:34 kahrl have you tried pushing it from the caller and passing the errfunc index to script_run_callbacks?
00:39 ShadowNinja Nope. But script_run_callbacks should be able to handle this by itself.
00:39 * ShadowNinja tries again.
00:48 ShadowNinja :-) Finally!
00:49 VanessaE ShadowNinja: your code works in singleplayer btw:  I was able to get a backtrace of the wield3d bug.
00:50 VanessaE https://forum.minetest.net/viewtopic.php?pid=117301#p117301  in case you're curious.
00:52 ShadowNinja Great, I should have the last corner for untraceable bugs to hide removed in a minute. (The function should really be in a slightly different position on the stack)
00:53 VanessaE do note btw:  the last line of that backtrace probably isn't being printed properly?
00:57 ShadowNinja Oh, I didn't know that was literal, but it is probably just trimming the long path since it isn't really usefull.
01:00 ShadowNinja VanessaE: Branch updated. Enable all your crashy mods and see if you get tracebacks now.
01:00 VanessaE ok
01:04 VanessaE ok, built and installed.
01:04 VanessaE let's see what breaks...and HOW :D
01:05 VanessaE (wield3d error has been reported to stu already)
01:05 VanessaE ShadowNinja: nope.avi
01:05 VanessaE still not getting a useful backtrace from the vector lib crash in the quarry/nuke
01:06 VanessaE (I'm doing this the smart way and testing in singleplayer with a copy of the survival map)
01:09 ShadowNinja That's... Impossible. The only C++ -> Lua calls without a errfunc are calls to table.insert or tostring.
01:09 ShadowNinja Unless there is a LUA_ERRERR...
01:09 VanessaE commit 63c12e21ee6fab7c3ec438add7810e04beff0f46 (and everything prior) is where I stand with the errfunc patches
01:11 zat ShadowNinja: u there?
01:12 ShadowNinja zat: Yes.
01:13 zat I fixed it all ShadowNinja, now it works perfectly.
01:13 ShadowNinja Alright, I'l test in a minute.
01:14 zat ok, I just pushed it
01:14 hmmmm so I just started using SVN for stuff at work
01:15 hmmmm horrible... horrible...
01:15 hmmmm you can't even amend commits
01:16 zat maybe that is one reason why Linux hated SVN so much
01:16 hmmmm now what's this stuff about script_run_callbacks?  is this related to sapier's new thing?
01:18 ShadowNinja hmmmm: No, I'm fixing how Minetest alwas calls lua_pcall with a NULL errfunc.
01:18 hmmmm oh
01:18 hmmmm how critical would you say that is?
01:19 hmmmm I am frankly sick and tired of seeing the sequence of digits "0.4.7"
01:19 ShadowNinja Well it causes a number of untraceable crashes, eg, the vector crashes.
01:19 hmmmm is that a new problem or was it there?
01:20 hmmmm (i really haven't been following things, sorry)
01:20 ShadowNinja This is a bug fix though, so it can go in during a feature freeze. It has existed for as long as Minetest has had a Lua API AFAIK.
01:20 kahrl since mod authors may want to test their mod with the stable version as well as git it would be nice if 0.4.8 had proper backtraces
01:20 hmmmm excellent
01:20 hmmmm that's a bug fix
01:21 hmmmm I say we have a feature freeze *NOW*
01:21 hmmmm we release 0.4.8 next wednesday
01:21 hmmmm everybody agree?
01:21 hmmmm well I don't really care if anybody agrees, I'm going to do it anyway because it's simply not moving along because nobody takes initiative
01:21 kahrl I don't have time soon anyway, so whatever you guys think is best
01:23 hmmmm and as for the super-duper configurable world create dialog in the main menu, I'd rather wait for better controls to be made
01:23 VanessaE no, don't freeze.
01:23 VanessaE not yet.
01:23 hmmmm I want to extend kahrl's table to be editable
01:23 hmmmm vanessae, any reason why *not*?
01:24 VanessaE there are a couple of features that imho should also go in before the freeze, like my 6d patch.
01:24 hmmmm sure, let's see that
01:26 VanessaE #959
01:27 hmmmm and it works?
01:27 hmmmm what do other people say?  have they tested it?  was it already in use in one of your mods?
01:28 hmmmm also, wasn't facedir prediction already added by kahrl a while back?
01:28 hmmmm it was one of his first patches he committed after the break
01:28 VanessaE it's been in use in homedecor for some weeks, in moreblocks for some months, and also recently added to one of Dan Duncombe's mods (I forget which)
01:28 ShadowNinja It seems good to me, although I haven't tested it
01:29 VanessaE idk if kahrl had a similar one or not.
01:32 kahrl that was something else
01:32 kahrl mine was client side
01:33 hmmmm so why does server side need prediction again?
01:34 kahrl I think that's supposed to mean "predict what the user meant"
01:35 hmmmm when is the actual determination on facedir made?
01:37 VanessaE hmmmm: what do you mean?
01:38 VanessaE oh, lines 231-233
01:38 VanessaE it's a simplistic algorithm but it works for 99% of people's use cases I'm sure.
01:38 hmmmm kahrl's thing predicts how the item is placed on the client side, your patch "predicts" how the item is placed on the server side, but it actually does add the node
01:39 hmmmm is your patch intended for mod-placed nodes based on the player's pitch and so on?
01:39 VanessaE yes basically.
01:39 hmmmm that is, facedir'd nodes are only placed as facedir when the player places them himself?
01:40 VanessaE for example, if you're holding a moreblocks slab or a homedecor light panel, it places it either upside down o nthe ceiling, or on the floor, or on the wall sideways depending on your pitch, facedir, and whether you're pointing at a wall
01:40 hmmmm but doesn't that already happen?
01:40 VanessaE nope.
01:40 hmmmm it doesn't??
01:40 hmmmm what
01:40 VanessaE well it does, because I added this code to moreblocks and homedecor ages ago
01:40 hmmmm hold on, let me try that out
01:40 hmmmm I mean with regular nodes, like stairs
01:41 hmmmm I could've sworn you were able to place it pre-rotated based on how you place it
01:41 VanessaE oh regular stairs nodes use some internal non-generic function
01:41 hmmmm aha
01:41 hmmmm that's not good, where is that?
01:41 VanessaE this is intended to replace that and extend it to handle pretty much all use cases.
01:41 VanessaE um....
01:41 VanessaE inside the stairs mod, I'm pretty sure
01:41 * VanessaE checks
01:42 VanessaE yeah
01:42 hmmmm ahh
01:42 hmmmm okay, your patch has an extremely valid use case
01:42 VanessaE stairs/init.lua line 24
01:42 hmmmm and you said it handles the upside-down placement as well?
01:42 VanessaE yes
01:42 hmmmm then it's superior to the stairs builtin version
01:42 VanessaE you can also *force* it by setting some flags :)
01:43 hmmmm give me one good reason why we shouldn't switch stairs to use vanessae's function right now
01:43 VanessaE the idea is that the mod calling this function will set the various flags e.g. invert_wall when the user holds, say, shift.
01:43 hmmmm that's even better
01:43 hmmmm i'm sold, that'll get committed right away
01:44 VanessaE the mod calling this function should set infinitestacks to true when in creative, btw.
01:44 VanessaE (I left it external because of Unified Inventory's handling of same)
01:44 hmmmm I have no idea what the whole controversial creative infinite stacks thing is about
01:44 hmmmm a lot of people sure talk about it though
01:45 VanessaE hmmmm: in short, default creative mod makes all stacks infinite (which makes crafting a bitch).  Unified Inventory treats all stacks as if you're in survival mode, even when in creative.  Pilzadam asserts that creative == infinite stacks, period, end of discussion.
01:46 VanessaE I assert that a separate flag should be set somewhere that a mod can refer to (or change).  This code allows for that, but if used by a default mod, you should just pass the creative mode flag directly as infinitestacks, to keep the existing behavior.
01:47 hmmmm mods setting their own flags.... yuck
01:47 hmmmm what if there's a conflict?
01:48 hmmmm what if one says "oh yes, I think I am comandeering this person's minetest, I say there are no infinite stacks"
01:48 hmmmm but then another says the opposite
01:48 hmmmm the former mod breaks
01:48 VanessaE yeah I know, that's the only drawback of such a flag.
01:49 hmmmm pilzadam is right here frankly
01:49 VanessaE but that's why I left it external to the call, so that by default the passed stack is left unchanged.
01:49 hmmmm if anything, the problem that makes crafting with an infinite stack should be fixed
01:49 VanessaE (if you set infinitestacks == to the creative flag)
01:50 VanessaE well anyways, have fun with it :)
01:51 VanessaE I tried to cover every possible use-case ( http://xkcd.com/927/ )
01:51 hmmmm yeah yeah
01:52 VanessaE second feature, which c55 kinda hates, #967
01:53 VanessaE I've been using this one since it came out.  works flawlessly, but I think the sun and moon need to be displayed at a larger angular size (and compensated for by visual padding in the texture files) so that things like a bright, big corona are possible.
01:55 hmmmm i'd rather it get perfected first
01:57 VanessaE perfected?
01:57 VanessaE well I guess
01:57 * VanessaE shrugs.  as RealBadAngel features go, this one just works :)
01:59 hmmmm shoot
01:59 zat ShadowNinja: you need to let it migrate to the sqlite file again, I changed the structure of a table.
01:59 hmmmm it's been too long and i actually forgot how to fetch a branch from someone else's repository
01:59 zat sorry for the inconvenience (time)
01:59 hmmmm i feel stupid :/
01:59 VanessaE hmmmm: git pull <git url> <branchname>
01:59 VanessaE or at least that's what I dp.
01:59 hmmmm no pull, that merges as well
01:59 VanessaE do*
01:59 hmmmm I just want to fetch.
02:00 VanessaE git remote add -f <username> <git url>  ; git cherry-pick <commit hash>
02:00 VanessaE that's the other way that I do it.
02:00 hmmmm I do a cherry pick after fetching the other person's repo
02:00 hmmmm shouldn't have to remote add anything
02:00 VanessaE oh JUSt fetch
02:00 VanessaE duh
02:00 hmmmm yes, I *just* want to fetch
02:00 VanessaE git fetch <username>  :)
02:01 hmmmm username?
02:01 hmmmm you mean url?
02:01 VanessaE no
02:01 VanessaE the name you gave when you did the last git add -f
02:01 VanessaE like I call shadowninja's fork just "ShadowNinja"
02:01 hmmmm oh, I didn't do that ever
02:02 hmmmm fetching from an arbitrary URL should be easy but it asks me for my username/password for some reason, then says repository not found
02:02 VanessaE wish I could remember more, sorry :-;
02:02 VanessaE :-/
02:02 hmmmm this is what happens when you don't do it for like 4-5 months
02:03 VanessaE mmhmm.  I used to be reasonably good with cvs.  Now I'd have to hit up Google just to clone someone's repo :P
02:05 hmmmm this is what happens when your knowledge of git is from muscle memory instead of real understanding
02:05 ShadowNinja zat: Instead of adding a rollback.txt.done it should add a rollback.txt.notdone while converting IMO. That way it doesn't leave behind a extra file when done.
02:05 VanessaE hmmmm: yep.  same damn thing with passwords too.
02:06 zat ShadowNinja: sounds good too, but it would not let the migrator know the real state of a previous migration, after completion.
02:07 zat ah no no... having only both .txt and .sqlite would mean migration was completed previously.
02:07 zat ¬¬ lol
02:08 hmmmm yeah, so I ended up doing it your way vanessa
02:08 hmmmm I can't remember how I used to do it...
02:08 VanessaE sometimes brute force is the easiest way :)
02:08 ShadowNinja I usually do "git checkout -b feature && git pull <url> <branch>"
02:09 kahrl git fetch https://github.com/kwolekr/minetest.git <-- works for me
02:09 hmmmm that does not work for me
02:11 hmmmm any comments on sapier's heavy-handed mingw/win32 i18n fix?
02:11 hmmmm (err rather, msvc build)
02:13 kahrl I don't like that it makes it practically impossible to write a batch file that automatically restarts the minetest server
02:14 hmmmm me neither
02:40 hmmmm what are your opinions on https://github.com/minetest/minetest/pull/957 (assume selection box for fences)
02:45 VanessaE hmmmm: +100
02:48 VanessaE hmmmm: also #minetest, re: 6d code...should I also add a little something there to define the orient_flags table if one isn't passed?
02:50 VanessaE it's supposed to be optional but it looks like I forgot to add a check for it being passed at all
03:10 VanessaE hmmmm: please add this one-line fix asap:  https://github.com/minetest/minetest/pull/986
03:20 hmmmm lol
03:20 hmmmm blindly pushing to upstream is a great way to find bugs
03:20 VanessaE heh yep
03:20 VanessaE should that be local, btw?
03:20 hmmmm it's analagous to "flip a coin, when it's in the air you'll know what you wanted"
03:21 hmmmm what?
03:21 VanessaE https://github.com/VanessaE/minetest/commit/79e5a7c803a8c29c70aea8f850ce0460f7f494b5
03:21 hmmmm the orient_flags?  it doesn't need to be
03:21 VanessaE ok
03:21 VanessaE I always get confused when it comes to variable scoping where a function is involved :)
03:31 hmmmm alright seriously
03:31 hmmmm feature freeze time
03:31 hmmmm i'm getting tired of this 0.4.7 bullshit
03:31 hmmmm remember when we 'almost' released it after being late a month, about 4 months ago?
03:32 VanessaE yes
03:32 VanessaE but wait!
03:32 VanessaE [11-06 22:31] <cy1> VanessaE: https://github.com/minetest/minetest/pull/987
03:33 VanessaE bugfix to vector.lua
03:33 hmmmm wow vector.lua is really getting bulky
03:33 hmmmm I hope you guys understand that  a). you need to spend as little time in lua as possible
03:34 hmmmm b). lua is slow
03:34 hmmmm c). things that require vector calculations are typically intensive in the first place
03:34 VanessaE abso-fucking-lutely.
03:35 VanessaE (on point a that is)
03:35 VanessaE why was the vector lib even written in Lua to begin with?
03:37 hmmmm beeeeecause
03:37 VanessaE random? :)
03:37 hmmmm someone wanted to be able to do v3f(4, 3, 2) * 4
03:37 hmmmm but then shadowninja said, "oh, we don't like metatable operations"
03:38 hmmmm I think originally it was my idea because I started playing with metatables
03:38 hmmmm with the VoxelArea constructor
03:52 VanessaE you seeing #minetest ?
04:02 VanessaE cy's patch works.
04:17 hmmmm I know
05:23 Ritchie joined #minetest-dev
05:39 VanessaE hmmmm: farming mod exploit:  https://forum.minetest.net/viewtopic.php?pid=117327
05:39 VanessaE (pilzadam's territory, but you're here and he ain't :) )
06:06 darkrose joined #minetest-dev
07:42 Ritchie joined #minetest-dev
07:52 mrtux joined #minetest-dev
08:18 Akien joined #minetest-dev
08:38 celeron55 the vector library is probably faster in lua than what it would be in C++ if luajit is used
08:38 celeron55 i mean, given that it's called from lua
08:55 darkrose joined #minetest-dev
10:08 smoke_fumus joined #minetest-dev
11:17 Ritchie- joined #minetest-dev
11:22 proller joined #minetest-dev
12:12 Taoki joined #minetest-dev
12:14 Ritchie joined #minetest-dev
12:15 harrison_ joined #minetest-dev
12:16 ydooW_ joined #minetest-dev
12:22 lanxu joined #minetest-dev
12:51 iqualfragile joined #minetest-dev
13:00 EvergreenTree joined #minetest-dev
13:13 ImQ009 joined #minetest-dev
13:32 iqualfragile joined #minetest-dev
13:33 hmmmm joined #minetest-dev
14:02 iqualfragile joined #minetest-dev
14:21 iqualfragile joined #minetest-dev
15:17 zat joined #minetest-dev
15:53 zat Hey ShadowNinja
15:54 EvergreenTree joined #minetest-dev
15:55 Jordach joined #minetest-dev
16:10 PilzAdam joined #minetest-dev
16:20 Zeitgeist_ joined #minetest-dev
17:01 Calinou joined #minetest-dev
17:17 EvergreenTree Can someone have a look at #214?
17:23 EvergreenTree https://github.com/minetest/minetest_game/pull/214
17:28 ShadowNinja The vector lib is fastest in Lua because of Lua -> C++ overhead.
17:32 zat Hey ShadowNinja, did you give it a try yesterday
17:33 ShadowNinja zat: A bit, I didn't have much time.
17:33 zat ok, I just made improvements
17:33 zat now rollback_checks are instant
17:33 ShadowNinja zat: I still got the C++ exception, and it locked up my system.
17:34 zat what did you do to get that?
17:34 ShadowNinja /rollback_check 10 1000000000
17:34 ShadowNinja Or maybe with a range of 100.
17:35 zat I hope it was because of a missing key in the table
17:35 zat hold on, I am migrating
17:35 Zeg9 joined #minetest-dev
17:36 ShadowNinja By locked up I mean it seemed very good at consuming CPU cycles and memery without leaving much for other programs.
17:37 zat it was sqlite scanning a big file I bet
17:37 zat how big is your rollback.txt?
17:45 VanessaE I have #214 already in place here and it works perfectly.
17:48 ShadowNinja zat: Rollback.sqlite is 240M and contains 7826433 entries.
17:49 zat that is why then
17:49 EvergreenTree Can someone check/test #214?
18:00 EvergreenTree Ahem, I said can some check 214 on minetest_game?   :P
18:02 ShadowNinja EvergreenTree: Looks good to me.
18:03 zat Ok ShadowNinja, it is done.
18:03 zat The last commit, whenever you have time.
18:03 zat thx for your time btw.
18:03 Sokomine ah yes. please have mercy with poor admins/moderators on servers and do test zats mysql-based rollback on real servers! hope to see that tested on a real server and eventually merged soon
18:03 Sokomine won't help on redcrabs server but lots of other servers could profit
18:03 thexyz EvergreenTree: I think you should make a function
18:05 thexyz similar to minetest.item_eat
18:06 thexyz but that's just my opinion, better ask other devs
18:06 VanessaE o.O a function for just two lines of code?
18:06 thexyz …And then they copypasted those two lines of code again […]
18:07 VanessaE well I guess
18:07 * VanessaE shrugs
18:07 zat VanessaE: have you got some huge rollback.txt file?
18:07 thexyz that's why I'm telling to ask other devs, I might suffer from perfectionism sometimes
18:07 zat we are looking for someone to test the new rollback.
18:08 VanessaE zat: is 122MB big enough?
18:08 Sokomine 122 mb is comparatively small
18:08 zat Mine is 1.9gb :D
18:08 Sokomine oh :-)
18:08 Sokomine that would fry any server with the current upstream rollback
18:09 zat that is why I never do any rollbacks in production
18:09 zat for rollback means hang up
18:09 VanessaE I must have nuked my old rollback logs recently.  The biggest one I have is 196MB, from a singleplayer-adapted old copy of my survival map.
18:10 VanessaE (previously, the largest one was like 2.5GB)
18:10 zat uh
18:13 werwerwer_ joined #minetest-dev
18:23 Sokomine that 2.5 gb one would have been helpful. they compress to very small size. if you could dig that out of an old backup?
18:28 zat if it exists lol
18:30 VanessaE I could try sure
18:30 VanessaE *fires up backup drive*
18:30 EvergreenTree joined #minetest-dev
18:31 EvergreenTree thexyz, Do I really need a function for just two lines of coded repeated only twice?
18:31 thexyz EvergreenTree: dunno, if you're sure it's only repeated twice then it's ok, I guess
18:32 EvergreenTree It would be useful for mods though...
18:32 VanessaE no dice, Sokomine.
18:33 EvergreenTree I think it would be fine for the mod creator to register it themselves though, would be more customizable that way
18:33 VanessaE the copies that made it into my backups (which I have long since made a point to exclude in my backup script) are smaller than the current ones.
18:33 PilzAdam thexyz, actually I suggested that the 6d facedir placement should return a function (like item_eat()) too
18:34 EvergreenTree Anyway, that's what VanessaE's 6d facedir is for
18:34 Zeg9 joined #minetest-dev
18:34 VanessaE if someone wants to modify it to do that, I'm not against it.
18:35 PilzAdam it should have been done before it was merged, though
18:37 rubenwardy joined #minetest-dev
18:37 EvergreenTree Hmm, if it is changed, you may want to do it before my pull is merged
18:38 EvergreenTree That way, no unnecessary commits are added
18:41 Ritchie joined #minetest-dev
18:41 Jordach joined #minetest-dev
18:42 ImQ009 joined #minetest-dev
18:43 VanessaE meanwhile, at some point I guess I should adapt the default stairs/slabs to use the new routine
18:44 PilzAdam I dont like that idea
18:44 ShadowNinja EvergreenTree: You should split the first line into two or three lines.
18:44 VanessaE why?
18:44 PilzAdam its just my personal preference
18:45 PilzAdam probably the same reaons why I dont like EvergreenTree's latest pull request
18:45 VanessaE it would cut down on code duplication and improve consistency.
18:46 VanessaE of course I never use default stairs, so meh.
18:48 PilzAdam btw, havent we decided to not touch minetest_game anymore?
18:48 VanessaE in the sense of adding major features, sure
18:48 VanessaE but it still has to be maintained.
18:49 PilzAdam in the featurelessness (Im pretty sure that word doesnt exist) of minetest_game horizontal trees are a major feature
18:49 VanessaE not really.  They're a logical extension of what already exists.  besides, there have been mods that add horizontal trees for ages now.
18:50 PilzAdam why do we add that to minetest_game then?
18:51 VanessaE eh?
18:51 VanessaE because this is not really the territory of an external mod?
18:51 VanessaE because there was no way at all before?
18:52 VanessaE why are you so opposed to allowing the default game to evolve just a little?  especially if it does so in ways that don't break compatibility?
18:56 VanessaE (inb4 <PilzAdam> bye  -->  * PilzAdam has quit (bye) )
18:56 PilzAdam hm?
18:58 VanessaE well this really doesn't belong in this channel, but you have a habit of signing off suddenly when someone starts putting questions to you that you don't like :)
19:02 zat joined #minetest-dev
19:08 EvergreenTree Hm, so what should be done about 214?
19:08 VanessaE anyhow, can we just cut the bullshit, as was discussed yesterday, and get this thing updated and released?
19:09 VanessaE EvergreenTree: you may as well functionalize it.
19:10 EvergreenTree Fine, but it would just be an extra commit, and is it really necessary?
19:10 sapier joined #minetest-dev
19:10 EvergreenTree If I did, where would the function go?
19:10 VanessaE make your commit and then,  git rebase -i HEAD~2
19:11 VanessaE stick the function somewhere in one of the other lua files in default, maybe functions.lua
19:11 darkrose joined #minetest-dev
19:11 VanessaE once you rebase -i HEAD~2, git push -f  and your pull will show as one commit instead of two
19:14 sapier I'll gladly rebase my changes hmmm if shadows fixes solve the backtrace issues it's worth it
19:14 zat joined #minetest-dev
19:15 lanxu joined #minetest-dev
19:15 sapier kahrl what about disabling the i18n fix in dedicated server mode I guess a server could live with broken locale
19:16 sapier btw the most simple fix would be just use a sane compiler to build minetest ;-)
19:20 ShadowNinja sapier: Yes, it does. I am currently busy making sure the errorhandler is removed from the stack after it is used. (And removing other unused things from the stack ahile i'm at it)
19:20 ShadowNinja while*
19:24 VanessaE ShadowNinja: did you see my messages?
19:27 VanessaE the short version is that, with one exception (the current release of wield3d), cy's patch to vector lib fixes all of the related crashes I've been seeing, and your path produces reliable backtraces for me once I sorted out an install-prefix glitch.
19:29 ShadowNinja VanessaE: Yes, I did.
19:29 VanessaE ok
19:31 mrtux joined #minetest-dev
19:40 sapier ahh :-) I just was wondering if that's correct not to pop it ;-)
19:40 sapier I prepare async for integration
19:43 Zeitgeist_ joined #minetest-dev
19:46 sapier if you found out how much needs to be popped from stack tell me shadow ;-)
19:51 EvergreenTree thexyz, Added a wrapper for rotate_and_place for rotating the logs
19:52 EvergreenTree in 214
19:55 ShadowNinja sapier: https://github.com/ShadowNinja/minetest/commits/pcall_errfunc More to come soon most likely.
19:56 EvergreenTree Added a function thexyz, happy now?   https://github.com/minetest/minetest_game/pull/214
19:57 sapier yes but where to look for the right way to pop it ;-)
19:58 thexyz EvergreenTree: that's not exactly what I meant, and I think PilzAdam wanted coroutine (for some reason, dunno why)
19:59 thexyz you should be able now to do "on_place = rotate_node"
19:59 EvergreenTree Tried that, didn't work
19:59 thexyz what the error was?
20:00 VanessaE http://pastie.org/8463495
20:00 ShadowNinja sapier: I use lua_pop or lua_remove, depending on whether it is at the top of the stack.
20:01 sapier a single one right?
20:01 ShadowNinja VanessaE: I updated pcall_errfunc. Pull it and see if it causes any crashes.
20:02 sapier is forum modlist at heavy load?
20:02 ShadowNinja sapier: Sometimes multiple lua_remove's in a row, but I just change the number for lua_pop, although I think I only had to pop multiple items once in that commit.
20:05 sapier ok I guess I've found the issue
20:05 sapier I'm gonna push a new version of async pull request
20:05 sapier if someone wants to try this intended to be added in 0.4.7
20:05 sapier 0.4.8 of course
20:07 sapier I added a copy of your error handler to async this way it's easy to replace it by your one shadow
20:07 VanessaE ShadowNinja: ok, deployed.
20:08 ShadowNinja Ok, I intend to push my patch soon.
20:09 VanessaE ok, that broke something.
20:09 thexyz sapier: there are some other services which create excessive load right now, sorry for that
20:09 VanessaE 15:09:33: ERROR[main]: ERROR: An unhandled exception occurred: ServerError: LuaE
20:09 VanessaE rror: allow_metadata_inventory_put should return a number
20:10 sapier sorry thexyz I don't know what you're talking about atm
20:11 thexyz 00:02  <+     sapier> | is forum modlist at heavy load?
20:11 sapier oh :-)
20:12 sapier n problem seems to work again I fot some timeouts when trying to read the modlist
20:12 VanessaE ShadowNinja: the rest of that crash ^^^^^ : /home/vanessa/Minetest-related/minetest_core/src/main.cpp:1877: int main(int, ch
20:12 VanessaE ar**): Assertion '0' failed.
20:13 ShadowNinja VanessaE: Ok, one minute...
20:14 RealBadAngel joined #minetest-dev
20:14 VanessaE sure.
20:14 RealBadAngel hi
20:14 VanessaE hey RBA
20:15 RealBadAngel could somebody update polish translation?
20:15 VanessaE thexyz: regarding EvergreenTree's patch, this was the error he was getting:  http://pastie.org/8463495
20:15 VanessaE thexyz: I am clueless when it comes to using the short form of a callback, which was formerly this:  http://pastie.org/8463502#7   What did he do wrong?
20:16 RealBadAngel ive translated missing texts ages ago, and now windows builds are scaring me with old ones or missing and empty buttons...
20:17 thexyz VanessaE: meeeh
20:17 thexyz on_place = rotate_node
20:17 thexyz that's all
20:19 EvergreenTree WEll, it worked
20:19 EvergreenTree applying
20:20 VanessaE oh well hell
20:20 VanessaE I feel stupid now
20:20 VanessaE maybe call it rotate_node_wrapper
20:20 VanessaE (or not, doesn't much matter)
20:23 EvergreenTree Applied
20:23 EvergreenTree Happy now PilzAdam?
20:25 EvergreenTree :)
20:28 EvergreenTree thexyz, better?
20:28 ShadowNinja VanessaE: Possible fix pushed.
20:29 thexyz sorry, too busy
20:29 RealBadAngel are weblate translations pushed manually? who or what is responsible for that?
20:31 sapier rba you most likely need to update again once the modmanager fix is pushed
20:31 thexyz RealBadAngel: I am
20:32 RealBadAngel what modmanager has to do with translations??
20:32 VanessaE ShadowNinja: ok, I'll grab it
20:32 RealBadAngel thexyz, could you update polish one then, please?
20:33 RealBadAngel its still incomplete, but way better than before
20:33 sapier if I remember correct there is at least a button whos caption is changed
20:33 sapier but you're right a single button isn't a reason to wait :-)
20:35 thexyz RealBadAngel: sorry, tomorrow
20:36 EvergreenTree thexyz, is 214 better?
20:36 RealBadAngel thexyz, sure, i just wanted to highlight that it is updated
20:37 RealBadAngel and thx
20:38 proller joined #minetest-dev
20:44 EvergreenTree thexyz, I said, is https://github.com/minetest/minetest_game/pull/214 any better?
20:45 thexyz EvergreenTree: sorry, didn't I say I'm a bit busy?
20:46 EvergreenTree Can someone else look?
20:46 thexyz it's probably easy to miss my message across all those
20:46 thexyz since PilzAdam suggested you to use closure please ask him instead of me
20:46 EvergreenTree Sorry, didn't mean to bother you
20:47 EvergreenTree PilzAdam, is https://github.com/minetest/minetest_game/pull/214 better?
20:47 PilzAdam shouldnt that be done in builtin?
20:48 EvergreenTree IS everything else okay?
20:48 PilzAdam I mean minetest.rotate_and_place() should return a function
20:50 Zeg9 joined #minetest-dev
20:53 ShadowNinja It should be in builtin, but returning a function won't work with infinitestacks and orient_flags.
20:57 lanxu joined #minetest-dev
20:58 RealBadAngel PilzAdam, could you update shaders branch with my latest commit?
20:59 ShadowNinja RealBadAngel: Too late.
20:59 RealBadAngel damn, power off :)
21:08 sapier https://github.com/minetest/minetest/pull/884 any second core dev to agree?
21:18 darkrose joined #minetest-dev
21:19 Ritchie- joined #minetest-dev
21:23 sapier everyone gone now? ;-)
21:27 IceCraft joined #minetest-dev
21:29 RealBadAngel sapier, idk what this code is about
21:29 RealBadAngel sorry
21:29 ShadowNinja sapier: Seems good to me.
21:30 sapier ;-) rba it's about releasing 0.4.8 I don't need those changes but some ppl  complained about the ui months ago. this fixes their issues now they're gone
21:31 sapier ShadowNinja thanks but you don't count as two agrees if you agree twice ;-)
21:31 * ShadowNinja shrugs
21:36 RealBadAngel sapier. i cant say anything if i dont test test the code. i could just look at it and say its nice, but it wont be fair
21:37 sapier of course and I don't remember you beeing one of those people requesting the change
21:38 RealBadAngel ask me something related to what im workin on and i will have something to say
21:39 RealBadAngel when i started to commit to mt i thought i can review any code
21:39 sapier while I understand this I don't think this is gonna work if all core devs only review things they're working at right now ;-)
21:39 RealBadAngel but thats not true
21:39 sapier why not?
21:40 RealBadAngel im focused now only on OpenGL, i can barely see whats going on in cpp now
21:40 sapier most of this commit is lua
21:41 RealBadAngel ok, i will read it
21:42 VanessaE ok.  thexyz:  please check #214 on minetest_game and #988 on the engine
21:42 sapier you don't have to do it if you don't think you can contribute, yet I think a different point of view sometimes even helps to find errors ... developers usually get blind to their own errors
21:43 RealBadAngel thats true
21:43 RealBadAngel but lookin in the very same direction is needed too
21:43 sapier of course
21:44 darkrose joined #minetest-dev
21:45 RealBadAngel i could give you a credit, with my vote, saying i do believe that you know what are you doing
21:45 RealBadAngel lookin at all your previous work, you deserve it
21:45 sapier I don't always think myself I know what I'm doing but thanks :-)
21:46 RealBadAngel theres no 100% perfect code
21:46 RealBadAngel it can be always made better
21:46 RealBadAngel even if you think atm its PERFECT
21:49 RealBadAngel but still, coding things all over is a way to get close to it
21:50 RealBadAngel and theres no better way than users telling you whats wrong
21:52 Miner_48er joined #minetest-dev
21:55 VanessaE hrm, #988 is wrong.  evergreen forgot to put minetest. in front of the function name,
21:56 RealBadAngel sapier, see? better to ask some1 who know things ;)
21:57 sapier :-)
21:57 sapier yet there are things evereyone can recognize e.g. documentation errors ;-)
22:02 VanessaE thexyz: see my comment attached to #988.
22:30 harrison joined #minetest-dev
22:31 harrison joined #minetest-dev
22:50 sapier left #minetest-dev
23:57 Taoki RealBadAngel: Hi

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