Minetest logo

IRC log for #minetest-dev, 2014-05-11

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

All times shown according to UTC.

Time Nick Message
00:14 sapier #1290 is supposed to fix (allmost) all the compatibility issues I caused by moving statbars to lua ... it's supposed to be removed as of 0.4.11. I said allmost, because I did not test clients prior lua hud, no idea if they will behave correct or not. AND clients supporting lua hud may have slightly changed heartbar positions. I don't see any way to fix this as it's a bug in those clients.
00:14 ShadowBot sapier: Error: Delemiter not found.
00:15 sapier ShadowNinja: is there a way to fix ShadowBot? it's not very usefull if it's not working (felt) half of times?
00:20 grrk-bzzt joined #minetest-dev
01:14 Taoki joined #minetest-dev
01:18 us`0gb joined #minetest-dev
01:54 sapier ShadowNinja: plz check #1291 and #1292
01:54 ShadowBot https://github.com/minetest/minetest/issues/1291 -- Opening online modstore crashes minetest
01:54 ShadowBot https://github.com/minetest/minetest/issues/1292 -- Async env broken (fails to handle modstore)
02:12 sapier left #minetest-dev
02:19 us`0gb joined #minetest-dev
02:32 Megaf folks, a guy is offering a build server on #Minetest
02:33 Megaf ShadowNinja: VanessaE
02:33 * VanessaE hides
03:03 us`0gb joined #minetest-dev
03:39 werwerwer joined #minetest-dev
04:14 CraigyDavi_ joined #minetest-dev
04:19 CraigyDavi__ joined #minetest-dev
04:26 cheapie joined #minetest-dev
04:37 Guest55042 joined #minetest-dev
04:42 crazyR__ joined #minetest-dev
04:48 hmmmmm joined #minetest-dev
04:48 Weedy joined #minetest-dev
05:00 hmmmmmm joined #minetest-dev
05:08 Weedy joined #minetest-dev
05:08 Weedy joined #minetest-dev
05:11 Weedy fuck it, forced card.freenode.net to 127.0.0.1
05:11 Weedy never want to join that god forsaken server ever again
05:11 VanessaE ?
05:12 VanessaE offtopic aside, perhaps you should contact the freenode staff and tell them what's going on?
05:14 Weedy now someone unban me from main channel
05:14 VanessaE done.
05:16 VanessaE bbl
07:04 nore joined #minetest-dev
07:06 restcoser joined #minetest-dev
07:08 darkrose joined #minetest-dev
07:08 Mimilus joined #minetest-dev
07:13 pi4 joined #minetest-dev
07:20 Calinou joined #minetest-dev
07:47 restcoser joined #minetest-dev
07:52 Calinou does anyone have a patch to make water not opaque unless you're digging?
07:53 pi4 joined #minetest-dev
08:23 diemartin joined #minetest-dev
08:38 jin_xi joined #minetest-dev
08:40 rdococ joined #minetest-dev
08:48 restcoser joined #minetest-dev
08:55 ImQ009 joined #minetest-dev
09:11 sapier joined #minetest-dev
09:15 SmugLeaf joined #minetest-dev
09:15 SmugLeaf joined #minetest-dev
09:17 tomreyn joined #minetest-dev
09:26 PilzAdam joined #minetest-dev
09:35 PenguinDad joined #minetest-dev
09:40 deltib joined #minetest-dev
09:41 CraigyDavi joined #minetest-dev
09:48 Calinou joined #minetest-dev
09:58 grrk-bzzt joined #minetest-dev
10:16 Taoki joined #minetest-dev
10:39 PenguinDad joined #minetest-dev
10:40 rsiska joined #minetest-dev
11:04 kahrl joined #minetest-dev
11:24 ImQ009 joined #minetest-dev
11:56 sapier what about #1241 it fixes 4 current bugs too
11:56 ShadowBot https://github.com/minetest/minetest/issues/1241 -- Add formspec toolkit and refactor mainmenu to use it by sapier
11:57 pi4 joined #minetest-dev
12:02 crazyR joined #minetest-dev
12:18 celeron55 joined #minetest-dev
12:29 Jordach joined #minetest-dev
12:53 sapier ShadowNinja: #1293
12:53 ShadowBot https://github.com/minetest/minetest/issues/1293 -- main_menu_script setting doesnt work any longer
13:09 CraigyDavi A new bug with long chat messages #1294
13:09 ShadowBot https://github.com/minetest/minetest/issues/1294 -- Problem with long chat messages
13:11 sapier CraigyDavi: that's not a bug but was intentionally added about half a year ago
13:11 CraigyDavi Why's that?
13:11 sapier spam prevention
13:12 sapier if you find a legit reason to write that much text we can discuss about it again
13:13 [PavelS] joined #minetest-dev
13:13 sapier problem with chat messages like that is they're sent to all clients, thus allowing messages like this can be easyly abused to do a denial of service attack to server
13:14 sapier and it's limited server side so client can't really know how much of the text is sent
13:14 CraigyDavi Hmm well I usually type quite a lot and ita rather annoying when people don't get the lst bit of my message
13:15 sapier you know irc does exactly same thing ? ;-)
13:16 CraigyDavi What about a config option? You can set how many characters people on the server can type before it starts cutting it off
13:16 sapier is this really worth adding a config option?
13:17 sapier I don't know the exact size but it's within a range almost noone would write
13:18 CraigyDavi enable_chat_spam_prevention = false. I would use that.
13:18 sapier noone would realize that using this setting may cause it's server to be more vulnerable to denial of service
13:20 CraigyDavi Well I sometimes copy and paste long bits of text into the chat
13:20 sapier even if customization is our goal CraigyDavi I don't believe making each pixel customizable will be the way to go, especially as that setting will cause speed penalty for each sent chat message
13:20 CraigyDavi Maybe put a comment above it saying "Can cause server to be more vunreble to denial of service"
13:21 sapier but of course that's my personal opinion
13:21 celeron55 half a year ago? wasn't that discussed like a month ago?
13:21 sapier a month only?
13:21 celeron55 i have always opposed it
13:21 sapier thought it's been more time
13:22 celeron55 where is the message limited now?
13:22 celeron55 i didn't even realize anything was committed related to that
13:22 CraigyDavi I have only noticed this when i updated by server recently. So yes it must have been in the recent month.
13:23 sapier for what I remember I wasn't very positive about adding it too. But I really hate changing things back and forward in 4 week intervals ;-)
13:23 celeron55 is it limited on the client or the server? i can't find the commit right now
13:24 sapier as you can see it on client I'd guess it's server side
13:25 sapier ok hard limit is 64k
13:25 celeron55 but i propose that the message length limit is tied to some privilege that is commonly used to discern between trusted and untrusted users
13:25 sapier we can't send bigger messages
13:25 sapier but there should be a smaller limit too
13:26 sapier wchar is 4 bytes isn't it?
13:26 celeron55 or dunno really
13:27 sapier well 16k characters would still be more then CraigyDavi tries to send
13:27 celeron55 when i was asked about this, i wanted to have a limit that isn't per-message but limits the amount of characters per minute or so
13:27 celeron55 that would be fair for everyone
13:27 celeron55 those who post long messages don't post them often anyway
13:27 sapier for what I remember ShadowNinja did suggest it .. but I'm not sure about it
13:28 sapier lol "writeU16(buf, (u16) (messagesize & 0xFF));" that's the bug
13:28 celeron55 in any case i think it is dumb that users have no way to remove the per-message limitation if they want to
13:29 celeron55 wtf? it is a bug?
13:29 celeron55 well this is the crappiest code in a long time
13:30 celeron55 and you broke it
13:30 sapier hmm could be my code ;-) .... I usually use the & 0x... notation on casting to make sure even silly compilers get it right
13:32 celeron55 https://github.com/minetest/minetest/commit​/f4f98c9550325aa8178f99cd32ea8806669aa280#d​iff-34f48ad91ac6c202ac60b0348ae90e30R2057
13:32 sapier Already about to fix it
13:32 celeron55 limit the length, don't & it
13:33 sapier it's not been ment to limit the lenght
13:33 celeron55 a bad idea in the first place
13:33 celeron55 i know; that makes it worse
13:33 sapier no that just makes it a regular bug
13:33 celeron55 it's a bug in code that was designed to do the wrong thing
13:34 celeron55 make it do the right thing and remove the bug
13:34 sapier I'll not rewrite the chat system
13:34 celeron55 what?
13:35 sapier adding a chat message rating system isn't a small fix
13:35 celeron55 is checking if a length is over 0xffff and using 0xffff if so "rewriting the chat system"?
13:35 sapier no that's the bug
13:35 celeron55 no, the bug is ANDing it with 0xff; the design error is ANDing it instead of limiting it
13:35 sapier tough't you're talking about the x messages per second thing
13:36 celeron55 the fix to the bug is using 0xffff instead of 0xff and the fix to the design is to limit it instead of ANDing
13:36 sapier there's an assert in front of it, using & 0xFFFF is paranoia code anyway ... thus beeing even more silly to cause a bug by it
13:37 sapier hmm thinking about it that assert is questionable too
13:40 celeron55 fun fact: the original code is 3.5 years old and worked fine enough until that commit
13:40 celeron55 (it only had the design error)
13:41 sapier well in this case you coul even consider it to be good I added the bug because now we can fix it
13:41 celeron55 generally in minetest code (that is newer than 3.5 years) the serializeString function is used, which throws a SerializationError if the length is more than 0xffff
13:42 celeron55 but that's not for wchars
13:42 celeron55 oh there is one for those too
13:43 sapier I don't know why I did use 0xFF there I used 0xFFFF everywhere else
13:44 sapier do you want it to be fixed to send maximum possible or just keep it the way it was before?
13:45 celeron55 maximum possible
13:45 celeron55 it makes sense for chat
13:46 sapier ok done
13:46 sapier I'd still wanna know why I added that bug at exactly that location
13:54 pi4 joined #minetest-dev
13:57 Megaf sapier: just one more little bug for you, I won't even open an issue
13:57 sapier tell me
13:57 Megaf sapier: run minetest client, then when you are on the client tab, where you put the server address and stuff, hit the up arrow, and see what happens
13:58 sapier already fixed in main menu cleanup pull
13:58 Megaf really?
13:58 Megaf I got that yesterday
13:58 sapier yes it's been reported within another issue yesterday
13:58 sapier well I rebased it today
13:59 Megaf ok
13:59 Megaf I will test it again later on, just reinstalled my OS here
14:04 Selat joined #minetest-dev
14:20 smoke_fumus joined #minetest-dev
14:25 SmugLeaf joined #minetest-dev
14:48 PilzAdam I made my own simple voxel based 3D engine in C++ with OpenGL and SDL; I wonder if I should build it into Minetest or make my own game out of it
14:49 sapier are there some screenshots?
14:49 ShadowNinja sapier: It might be failing to retreive the title.  I'll add a bit more info to the message. -- main_menu_script has been renamed to main_menu_path, which should point to a directory with an init.lua. -- I don't remember commiting my message length limit commit, but the protocol limits it to 65535 chars.
14:50 sapier ShadowNinja:  we already fixed the chat bug ... well your change causes android to have to create a COPY of all files
14:51 PilzAdam sapier, http://i.imgur.com/YcYRhCR.png
14:51 ShadowNinja sapier: What do you mean?
14:51 sapier or ignore the get menupath feature which is absolutely useless if I have to use hardcoded menu paths again to access those files for some slightly different menu
14:51 Megaf PilzAdam: That's really cool! Well done
14:52 sapier android menu uses exactly one different tab then normal menu and is only a different init.lua script
14:52 PilzAdam (its a really short render distance in the screenshot)
14:52 Megaf PilzAdam: so, did you still use Irrlicht?
14:52 ShadowNinja sapier: It isn't hardcoded, it just points to a directory now.  And the entrance point is init.lua.
14:52 PilzAdam Megaf, no, OpenGL and SDL
14:52 sapier as I can't specify the menu script after your change I either have to set all paths relative to builtin .. or copy everything
14:52 Jordach joined #minetest-dev
14:52 Megaf really cool
14:53 sapier ShadowNinja:  that's exactly what is absolutely useless to menus co-using original menu code
14:53 ShadowNinja sapier: Instead uf using a seperate script make init.lua detect android.
14:53 sapier adding same crappy worarounds as in previous menu again?
14:54 ShadowNinja mainmenu/init.lua should probably use get_builtin_path() so that a seperate menu can call it after doing setup.
14:54 sapier init is quite different from minetest to androd menu so it's basically a silly thing like
14:55 sapier If android then
14:55 sapier else
14:55 sapier end
14:55 sapier both parts containing about 200 lines of unrelated code
14:56 sapier it's been you prising the benefits of splitting code not beeing connected
14:58 sapier while this is a code organization issue only it's way less critical then the async api crash you introduced
14:58 sapier bugs happen but please fix them
14:59 sapier PilzAdam: do I understand correct you did write an additional fork? ;-)
15:01 sapier ShadowNinja:  renaming engine to core didn't work either for mainmenu
15:02 sapier it's up to you if you wanna fix the compatibility issues or just change the api but don't ever tell me "it did work before"
15:02 ShadowNinja sapier: It isn't done by the main init.lua.
15:02 sapier try entering modstore in current master you'll see the compat problem
15:03 PilzAdam sapier, not a fork
15:03 PilzAdam it's completely written from scratch
15:03 sapier tell me more about what you did PilzAdam
15:03 sapier what did you rewrite?
15:03 PilzAdam I installed SDL, opened a text editor and wrote a voxel engine
15:04 sapier so basicaly irrlicht?
15:04 PilzAdam and now I wonder if I should rewrite Minetest to use the OpenGL renderer
15:04 sapier hmm sdl ... irrlichts software renderer?
15:04 PilzAdam or if I should turn it into my own game
15:05 PilzAdam SDL is just used to create the OpenGL context
15:05 sapier PilzAdam: did you think about how much work has to be done to get a engine ?
15:05 sapier not the first cool 50% but the follow up 50%
15:05 sapier well silly question :-)
15:06 sapier you probably did
15:06 ShadowNinja sapier: http://ix.io/cfK/diff
15:06 sapier did you test it?
15:07 ShadowNinja sapier: Yes, it starts to load and then crashes with: *** Error in `bin/minetest': realloc(): invalid pointer: 0x0000000003560eb0 ***
15:07 sapier ok then you're as far as I am, that's the one caused by async replacement
15:08 sapier I didn't do futher investigation but may be related to stack depth
15:10 ShadowNinja sapier: Protip: Enclose code in ``` in github issues.
15:10 sapier btw as this init file is already as ugly as what I don't wanna do in mainmenu you can fix compatibility by honoring menu script in there too
15:12 ShadowNinja sapier: Hmmm, that seems to happen deep in Lua.  Maybe the errfunc is invalid?
15:13 ShadowNinja Well, I'll push that patch I just pasted.
15:13 sapier please don't
15:13 sapier It's just causing additional rebasing on fstk
15:14 sapier and it's irrelevant there as I alreade removed all remains of engine
15:14 ShadowNinja sapier: It only changes one line in the mainmenu.
15:14 ShadowNinja sapier: Ah, O.K.
15:15 Megaf sapier: open minetest client, click on the server address input box, then press up arrow and see what happens
15:15 sapier isn't this same thing I already fixed in fstk?
15:15 Megaf sapier: I just did a clean install of mintest from git
15:15 Megaf still happening
15:16 sapier it's not in master
15:18 sapier ShadowNinja: what about #1241 I fixed any damn comment except of metatables ... I will only change this if you take responsibility for fixing any future bug in there because I really don't wanna look for metatable issues
15:18 ShadowBot https://github.com/minetest/minetest/issues/1241 -- Add formspec toolkit and refactor mainmenu to use it by sapier
15:20 hmmmm joined #minetest-dev
15:28 Zeitgeist_ joined #minetest-dev
15:28 Zeitgeist_ joined #minetest-dev
15:53 Calinou joined #minetest-dev
15:55 Weedy joined #minetest-dev
15:55 Weedy joined #minetest-dev
16:01 Eater4 joined #minetest-dev
16:06 cheapie joined #minetest-dev
16:17 rsiska joined #minetest-dev
16:28 zat joined #minetest-dev
16:49 MegAFK !up minetest.megaf.info 30005
16:49 ShadowBot MegAFK: minetest.megaf.info port 30005 is down
16:49 MegAFK liar
16:50 MegAFK I dont get it
16:50 MegAFK wrong channel again
16:50 MegAFK left #minetest-dev
16:55 xyz joined #minetest-dev
16:55 ImQ009 joined #minetest-dev
16:55 xyz sapier: why are you opening an issue and then deleting comments from it?
16:55 xyz sapier: just merge your code to master instead of wasting our time
16:56 sapier because that commit wasn't about the neverending "We need to replace formspecs" discussion
16:57 xyz so it's okay to just delete it?
16:57 sapier to me a commit comment isn't a forum thread discussing various topics but related to this particular commit
16:57 xyz just because you didn't like it?
16:58 xyz it wasn't a commit comment
16:58 xyz it was an issue comment
16:58 xyz an issue about replacing formspecs with formspecs
16:58 sapier no I'm perfectly fine about the discussion I just wont take part of it
16:58 sapier because I'll not write a 100% compatible yet perfectly new implementation of irrlicht gui, qt/gtk whatever toolkit and define a new gui description language too
16:59 xyz sapier: why did you remove my comment then?
17:00 sapier for what I remember it was first comment of a discussion how formspec replacement should look like
17:00 CraigyDavi joined #minetest-dev
17:00 sapier that's not topic of this pull request
17:00 xyz sapier: so you think it's okay to just remove comments?
17:00 xyz and wipe out the whole discussion?
17:00 xyz wtf
17:00 sapier as of my definition it's spam within that pull request
17:01 xyz your definition
17:01 xyz you're not the only collaborator
17:01 xyz you're not the minetest project leader
17:01 xyz yet you get to decide what is spam and what is not
17:02 sapier true, but I'm the only one who seems to be willing to stand the attacks of those who don't contribute but only do bashing
17:02 xyz so i guess this comment attacked you in some way?
17:02 xyz well, i'm sorry
17:03 sapier and xyz I'm annoyed about beeing attacked for things not beeing related, that's why I removed it there
17:03 xyz you expect people to contribute but you treat them like shit
17:03 sapier if there was a move option I'd have moved it but I don't know about any
17:03 xyz then you whine that you always get attacked
17:03 xyz i wonder why
17:04 sapier well xyz I'm here you're free to describe your new formspec variant
17:04 PilzAdam sapier, don't delete comments
17:04 sapier and how is it related to a cleanup of current mainmenu?
17:04 xyz sapier: it's clear that there's no point in discussing anything with you
17:04 xyz judging by that attitude
17:04 PilzAdam (except the are spam or about sfan5's avatar)
17:05 sapier PilzAdam: so we use any pull request to discuss about things not related to it?
17:05 PilzAdam no
17:05 PilzAdam just say that its offtopic, don't delete it
17:05 sapier well then I was wrong, true, I'm gonna try this next time ... but I really doubt that's gonna work
17:07 sapier I'm gonna look if I can restore those comments
17:09 sfan5 you can't
17:09 sapier Well I guess I've got the emails beeing sent by them, then I'm gonna add those offtopic things this way
17:10 sapier especially as constructive comments as ":-1:, formspecs should just die"
17:10 Exio4 it should be
17:10 Exio4 er
17:10 Exio4 sorry, wrong channel
17:12 xyz sapier: the king of minetest has decided my comment isn't worth attention?
17:14 sapier xyz contrary to you I didn't decide to run away but am still here and face all the personal attacks
17:15 xyz sapier: nice i hope you're having fun
17:15 sapier An yes I sometimes make mistakes too
17:16 sapier especially after fighting for hours and without understandable reason
17:17 Jordach sapier, why did you have to break custom huds again with the luastatbars
17:17 sapier Jordach: if they're broken that's not intentionally
17:18 sapier what's wrong about them Jordach?
17:18 Jordach hold on!
17:18 sapier custom ones are supposed to behave exactly as before, that's why I did it
17:19 Jordach https://cdn.mediacru.sh/2h-3QeEKzXUC.png
17:19 Jordach still doesn't disable the annoying old stuff
17:20 sapier well the fix for this wan't merged by now becaus it's only gonna fix 95% of old clients... I asked for comments but as usual noone did answer
17:20 sapier 95% meaning old custom hud is supposed to be identical, but health and breath bar are located a little bit different
17:21 sapier only if you connect a old client to a new server
17:22 sapier Jordach: what do you prefere?
17:23 Jordach sapier, bars that scale (including higher res images into a slot, like they used to in 20120112)
17:26 sapier Jordach: that's exactly what I added
17:27 sapier if you want to try it use a dev version do singleplayer, fixing the mixed mode is ready to pull
17:29 sapier talking about it, any comments to #1290?
17:29 xyz sapier: it's not running away but advancing in the opposite direction by the way
17:29 ShadowBot https://github.com/minetest/minetest/issues/1290 -- Fix old client showing duplicated health bar on new server by sapier
17:30 sapier well xyz I decided to try to proceede minetest within community which is way more hard then developing alone, I'm not sure if this is better way to do it any longer xyz
17:31 xyz sapier: yes, in fact, you're wasting your time
17:32 sapier xyz so you think it's better to do your own stuff then contributing?
17:32 xyz sapier: you already are doing your own stuff
17:33 sapier that's not been my intention and never will be, you might have noticed none of my commits went in without major changes due to others comments
17:34 sapier that's annoying of course ... and may be reason for me beeing a little bit harsh the last week
17:35 sapier but you told about another direction, tell me more about it I'm interested
17:36 GrimKriegor joined #minetest-dev
17:37 xyz okies, I'll assume you only ignore/delete my comments
17:37 Jordach sapier, lemme test that client
17:39 sapier xyz I usually try to consider any comment, but I'm not perfect too
17:48 sapier xyz? I can't not ignore comments not beeing made :-)
17:49 xyz and those that are made you will delete because you didn't like them
17:49 xyz surely it's a wonder no constructive comments appear
17:51 sapier I don't wanna discuss about ", formspecs should just die" is a constructive comment ... and I already excused for deleting them ... if this didn't reach you ... sorry I'll not delete comments again ... none never ;-)
17:53 xyz what's the status of android port by the way?
17:54 sapier actually all this fighting is about android ports as those fixes are the generic parts of it
17:54 sapier It's feature complete, of course there most likely are still some bugs
17:55 sapier and you can do polishing and better graphics too
17:55 Jordach sapier, your branch is a dud
17:55 Jordach https://cdn.mediacru.sh/9PzBbtCzwJCz.png
17:55 ShadowNinja sapier: What's the link to the latest debug build?
17:55 sapier jordach what version do you use?
17:56 sapier http://animalsmod.comuf.com/​downloads/Minetest-debug.apk
17:56 sapier Jordach: server side
17:56 Jordach sapier, what about clients acting as the server
17:56 xyz what about new network protocol?
17:57 sapier xyz I'm absolutely positive about it as long as we do a proper benchmarking before
17:57 xyz so, no updates from you?
17:57 sapier my enet branch as well as tcp branch are there
17:57 sapier they've been ready for about half a year now
17:58 xyz well, so?
17:58 xyz there's no point in having this code if it just rots somewhere in some branch
17:59 sapier Well I didn't merge it because of same thing I face now, I'd have been told to takover everything, I don't want to do this, I never did. It's not even been my idea to merge the fixes for old protocol. Of course I didn't fight against merging it.
18:00 sapier Jordach: can you tell me what mods are installed?
18:00 ShadowNinja sapier: It looks like this on a nexus 10: http://i.imgur.com/mdrCGlx.png
18:01 xyz since you didn't answer my question i'll just assume there's been no updates on that front
18:01 sapier hmm nexus is a high dpi screen?
18:01 ShadowNinja sapier: And the serverlist slider is far too small.
18:02 ShadowNinja sapier: Yes, very high.  300dpi or so.
18:02 sapier no I didn't do any additional update as I've been busy merging android fixes
18:02 hmmmm joined #minetest-dev
18:02 sapier hmm seems there are two issues, first, formspecs not scaling to dpi and second  some bug in scaling up formspecs to big screens
18:03 sapier and the issue about tab detaching from shown box of course
18:03 sapier can you run minetest ?
18:04 sapier xyz do you have a new protocol capable of handling old clients too?
18:04 Jordach sapier, https://cdn.mediacru.sh/kKBljGR2q41W.png
18:04 xyz sapier: no, we plan to drop compat
18:05 sapier Jordach: for wat I see the heartbar is same and food bar is same to but located different
18:05 xyz also looking for other things to improve
18:05 sapier xyz who is "we"? and what to you plan to use?
18:06 ShadowNinja sapier: The game runs, but at only 14 FPS.  This is a big screen, but it's a high-end android device.
18:06 sapier how about the gui controls do they scale to a usable size?
18:07 sapier Jordach: your first screeshot seems to indicate hideing built in heartbar doesn't work correct
18:07 xyz sapier: FM, meaning me and proller
18:07 xyz sapier: protocol-wise? enet + msgpack
18:08 sapier not worst choice of course
18:08 sapier isn't celeron working on something similar to msgpack too?
18:08 xyz no idea
18:08 sapier talking about enet, I haven't found a way to benchmark it by now
18:08 proller bicyclepack + bicyclenet
18:09 xyz sapier: my goal is to remove as much code as possible
18:09 xyz while adding new deps when it's reasonable
18:09 sapier xyz sane goal, basicaly I try to do same thing but with quite different methods
18:10 xyz basically you add code instead of removing it?
18:10 proller sapier, different methods = add 1000 lines of code with every "cleanup" ?
18:10 ShadowNinja sapier: Nope, the serverlist slider is far too small.
18:10 sapier on first glance this is true yes, but if you have a look at connection handling I separated low level code from core thus making it replaceable without changes in core code
18:11 ShadowNinja sapier: Unified Inventory is scaled well though.
18:11 ShadowNinja sapier: There are also lots of "ERROR[main]: Name of selected formspec: " lines.
18:11 sapier ShadowNinja: I guess I know what's happening, I was forced to make mainmenu fixed size on pc ... thus it's not scaling ... I'll change this for android mainmenu
18:11 ShadowNinja (In the chat)
18:12 sapier yes I already fixed this in dev branch ShadowNinja
18:13 ShadowNinja And UI needs better icons, it looks terrible on x+hdpi devices.
18:13 sapier :-)
18:13 sapier that's what I meant with polishing
18:13 sapier I'm not a graphics guy, I need help for those things
18:14 xyz on first glance it looks like i'm adding 1000x lines of code but in fact i make things replaceable where they don't have to be
18:15 sapier xyz it's about making minetests core structure to be more modular to not make a packet cause a lock for luaengine to be take
18:16 xyz ah, i just was thinking about that thingy that supported three protocols at once
18:16 sapier basicaly it's meant to be able to remove the big map lock by some future time
18:16 xyz what a useful feature, would like to spend time developing
18:16 ShadowNinja sapier: The text is also very small.  Barely readably for me, and my eyes are pretty good.
18:16 sapier xyz it's never been meant to support 3 protocols, but yes it's supposed to support two, legacy and new one for a limited time
18:18 xyz this feature is no less great
18:21 Calinou https://github.com/minetest/minetest/issues/826 → a core dev should reproduce this and remove the “unconfirmed bug” label
18:21 sapier of course, especally as it's adding a usefull benchmark platform to compare different protocols too ... you don't have to rely on your stomach but can test it
18:22 sapier Jordach: are you still there?
18:22 ShadowNinja sapier: And icons don't render well: http://i.imgur.com/laWrVFg.png
18:22 ShadowNinja BBL.
18:22 sapier ShadowNinja: can you check if those icons are non npot2?
18:28 xyz sapier: have you tested it?
18:28 sapier define "it"?
18:29 xyz sapier: different protocols
18:30 sapier I tried legacy, tcp and enet
18:30 sapier as of bandwidth restult was legacy <big gap> enet <smaller gap> tcp
18:31 sapier I couldn't tell anything about jitter because enet does provide jitter data but it's invalid
18:31 sapier there's no way to have a jitter of 0 on a real network connection
18:32 sapier I did ask you for help to find what's wrong there too
18:34 xyz i guess i didn't really care
18:34 xyz nor that i do now
18:35 sapier you remember accusing me of ignoring comments?
18:37 xyz yes and?
18:38 sapier I'd not block enet once we (at least somehow) know if and how much better it is and if it's not way worse then tcp
18:38 sapier and message format ... why mess up different layers? to me this is a completely different topic
18:38 EvergreenTree joined #minetest-dev
18:44 BlockMen joined #minetest-dev
18:49 PenguinDad joined #minetest-dev
18:50 Jordach sapier, ?
18:51 sapier I'm still looking for what happens for you because I don't seem to be able to reproduce that issue (double hearts)
18:52 PenguinDad sapier: I noticed this too in BFD
18:53 sapier yes on mixed (old/new) this is still happening as the fix isn't merged yet
19:09 crazyR joined #minetest-dev
19:39 sapier Jordach: I believe it's fixed now, I tried with better hud
19:39 Jordach sapier, with two clients or something else?
19:42 sapier well if server runs the broken version it's not gonna work but with my branch its supposed to work with old/new combinations too
19:46 Jordach sapier, i'm not sure on how hud does the "statbars"
19:46 Jordach BlockMen, ^
19:46 Jordach except the old hearts are gone
19:47 BlockMen Jordach, it hide the engine statbars and adds new
19:47 Jordach ah, but they don't scale correctly, unlike the stuff that was there before
19:48 Jordach they still sit at 16px no matter the window size
19:48 BlockMen Jordach, yes, needs to be updated to scale correct
19:49 Jordach ah
19:49 BlockMen i will release an update if all bugs are gone
19:58 EvergreenTree joined #minetest-dev
19:59 Jordach thank you :)
19:59 Fresh_me_ joined #minetest-dev
20:05 ImQ009 joined #minetest-dev
20:38 ImQ009 joined #minetest-dev
20:45 EvergreenTree joined #minetest-dev
20:45 cheapie joined #minetest-dev
20:57 Megaf joined #minetest-dev
20:58 Megaf why does it takes so long to migrate from one db to other one?
21:00 Megaf sapier: ShadowNinja 18:00:40: ERROR[main]: MAINMENU ERROR: /home/megaf/MinetestClient/bin/.​./builtin/mainmenu/init.lua:911: attempt to get length of field 'favorites' (a nil value)
21:01 Megaf http://paste.debian.net/98914/
21:02 sapier did you try fstk addon branch?
21:02 sapier it's supposed to be fixed there
21:07 Megaf sapier: I'm always on the latest GIT
21:08 Megaf on the master branch
21:09 sapier ok then it's not fixed there
21:12 Megaf There is where it's supposed to be fixed...
21:13 sapier of course, but I won't fix a bug in code to be replaced anyway
21:14 * Megaf see that using master branch for his server was a terrible choice
21:15 sapier well master is unstable
21:15 sapier not having a stable release for quite some time is an issue we suffer
21:15 Megaf master is the only place we can find bug fixes...
21:15 Megaf or it was
21:16 Megaf I used to use stable version, and every single time I complained about a bug you would always say that it was fixed already on master
21:16 Megaf damn you minetest developers!
21:16 Megaf ^ celeron55
21:16 sapier yes ... we're not consistent about this
21:16 Megaf I'm fed up
21:16 Megaf goodbye
21:17 Megaf left #minetest-dev
21:19 us`0gb joined #minetest-dev
21:20 celeron55 i always find it funny when people ragequit blaming people who do stuff completely free out of passion as a hobby
21:22 sapier hmm I didn't take his comments to be serious but ironicaly .. have I been wrong
21:22 sapier ?
21:22 celeron55 dunno, it's funny anyway
21:22 celeron55 it also embraces the feeling of freedom in a weird way
21:25 celeron55 i might actually suggest that you fix the problem temporarily in master if you're not going to merge the big change in a reasonable time
21:26 celeron55 it's well known that master is used by many people so it should be kept working by reasonable effort
21:26 sapier if noone does comment on it I wont merge it ever
21:26 sapier and fixes like that one are quite risky to be lost as git is unable to merge them
21:27 sapier and I didn't change anything in this location so it's most likely broken for ages why hurry now?
21:28 sapier and to be honest I'm usually not very glad to merge patches causing additional work I already have done at least twice
21:28 celeron55 whatever as long as you have considered it
21:29 celeron55 doesn't seem to be a prominent issue
21:31 sapier yes I did consider but that doesn't necessaryly mean my consideration is correct, especially as I may not have all information of a bug
21:33 celeron55 you might have to poke some modders to get comments on the interface
21:34 celeron55 i'm not really interested about the internals as it doesn't touch anything on the C++ side and i have never studied the lua main menu code to begin with
21:34 sapier the only response I got sofar is "formspec should die" ... not completely wrong but useless too
21:35 celeron55 can you write an example using a plain formspec and the same thing with that api, and then throw that at people's faces?
21:35 celeron55 maybe that will lead to useful discussion
21:36 celeron55 even i might comment on such
21:36 sapier that's a good idea ... quite a lot of work ... but maybe it's gonna help
21:37 celeron55 it can be quite small i guess, just something that shows the general structure of code
21:38 sapier I'll try, but I fear the benefits start to become obvious on having different tabs only ... well lets see
21:38 celeron55 it's really hard to get people actually testing things that aren't in master but it seems to have to be handled on a case-by-case basis like this
21:40 celeron55 maybe if a windows build was made and an official request for testing was posted on the forums... but that's a lot of work too
21:41 AspireMint joined #minetest-dev
21:42 AspireMint left #minetest-dev
21:42 sapier well the other way is merging code you know to most likely contain bugs you haven't found by now and fix it once they occur.  That's gonna be quite tough. And I don't wanna do this for the rest of those fixes required for android port.
21:42 celeron55 these two discussions actually show very clearly what's wrong: people expect that master is stable, but then they expect that they don't need to test anything outside of master 8)
21:44 celeron55 maybe giving some kind of recognition to people who test things outside of master would work
21:44 celeron55 currently they really don't get any in the changelog or so
21:45 sapier what do you think about using the "signed off" thing for this?
21:45 celeron55 umm... what would that do?
21:46 sapier those signing of would be mentioned in commit? whouldn't they?
21:46 celeron55 i don't think commits are important
21:47 sapier maybe I missunderstood what you meant with "recognition"
21:48 celeron55 i mean like getting the community to know that they did a part in making the feature happen
21:49 celeron55 anyway, that's probably a minor thing
21:49 sapier some sort of top-testers list?
21:49 celeron55 though it's hard to tell; someone could make a poll about this (where's our community manager? 8))
21:50 celeron55 i was thinking just appending the names to the developers
21:50 celeron55 like "formspec toolkit thing (sapier, <somebody who tested it>)"
21:54 sapier I'd not have any problems with giving credit to testers too
21:55 sapier I'd support this suggestion in a pull
21:55 sapier poll
22:02 BlockMen celeron55, how does that work with the LIGHTBANKs?
22:06 kaeza joined #minetest-dev
22:07 SoniEx2 can we get mod/game/whatever hotswapping + /load + /reload + /unload?
22:08 sapier SoniEx2: very unlikely the next months
22:09 SoniEx2 sapier, uhh why?
22:09 sapier mods aren't separated from other data thus we can't remove them
22:09 SoniEx2 uhh ok
22:09 sapier once their part of script api engine it's (almost) impossible to remove
22:09 sapier we'd have to fid a way to separate them first
22:10 sapier and that's most likely quite a lot of work
22:11 sapier but I suggest writing a feature request anyway SoniEx2 maybe someone is looking for work and finds it
22:11 SoniEx2 ok
22:12 sapier but I don't wanna tell you wrong chances aren't that big to get it soon
22:14 celeron55 simply put: doing that is completely against the current design of the lua system and even building it from the ground up to support that would be a questionable effort
22:14 celeron55 however!
22:15 celeron55 it could be simulated by making it possible to unload everything and then load everything back but with something removed
22:15 celeron55 it could look to the user as just like /load, /reload and /unload and wouldn't be slow
22:16 celeron55 but that does require things that aren't there and aren't planned otherwise (like the protocol for making clients understand what's going on)
22:39 Taoki joined #minetest-dev
22:57 sapier left #minetest-dev
23:19 tomreyn joined #minetest-dev
23:44 NakedFury joined #minetest-dev

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