Time  Nick          Message
00:03 ShadowNinja   Taoki: Why not sypport it if all it takes is an #ifdef?
00:03 Taoki         ShadowNinja: I was wondering whether an #ifdef can be used. But I'm not sure how to do that for Irrlicht.
00:03 Taoki         In either case, it means the improvement won't work for people with Irrlicht 1.7. If that is acceptable I don't mind.
00:04 Taoki         Not sure about my workaround to the restarting animation glitch, that might work. But non-looped animations will not.
00:16 ShadowNinja   Taoki: #if (IRRLICHT_VERSION_MAJOR == 1 && IRRLICHT_VERSION_MINOR > 8) || IRRLICHT_VERSION_MAJOR > 1
00:17 Taoki         Thanks, I'll try this
00:18 Taoki         ShadowNinja: Will another Travis build be triggered if I force update the branch? I won't make another commit since I'll then be requested to squish them and don't know how especially on Github
00:20 ShadowNinja   Taoki: Yes, Travis knows to rebuild.
00:20 ShadowNinja   (and I can force it anyway)
00:20 Taoki         Nice. I'll hopefully find a solution in a few minutes
00:26 Taoki         Fixed and pushed. It should work now... unless the start / end frame checks also require 1.7, then I have to do it for those as well
00:26 Taoki         Travis compilation is in progress, no need to retrigger it
00:35 Taoki         Awesome, no more issues now :)
00:43 Taoki         ShadowNinja: Is the formatting for the #if tags good? Those I can break into multiple lines if it's too long, and if the compiler doesn't mind of course
00:45 ShadowNinja   Taoki: They're fine.
00:45 Taoki         Nice. I can put the code aside for now in that case. I assume others will have to test it before it's merged so
01:29 paramat       now merging game#412
01:29 ShadowBot     https://github.com/minetest/minetest_game/issues/412 -- mapgen.lua: Add dirt/sand/gravel blobs to mgv5/mgv7. Use blob ore for clay. Tune flowers by paramat
03:03 Zeno`         ShadowNinja, Does %g work better?
03:06 Zeno`         nvm
03:07 ShadowNinja   Zeno`: No, it's even worse.  Things as short as 12345678 use exponential notation (Currently gets to 12345678912345678 or so without losing precision).
03:08 Zeno`         yeah :)
03:11 Zeno`         you could always specify precision but I don't really know what the problem was, so *shrug*
03:30 ShadowNinja   Zeno`: The problem is that large integers (used for node hashes) need to be serialized precisely, but floats need to work too.
03:32 Zeno`         Ah, I see now
05:30 hmmmm         jesus christ I can't stand it
05:30 hmmmm         I don't want to be here right now but minetest is headed toward disaster
05:31 hmmmm         at least respect my last wish:  stop with the protocol breakage.
12:04 sfan5         is there not going to be a 0.4.13 or why are we planning to merge 0.5 pulls into master soon?
12:07 Krock         network changes are coming and 0.5.0 would be good for a compatilibity break
12:25 sfan5         apparently not
12:25 sfan5         nrz announced in the forum that we will be merging incompatible changes (that are for 0.5) on 1st march
14:17 Wayward_One   did i miss something? i was under the impression the next release would be 1.0
14:18 nrzkt         no :p
14:18 nrzkt         but this is a version bump, yes
14:18 nrzkt         and this release will happen when the roadmap will be done, or nearly done
14:19 Wayward_One   ah, ok
14:20 nrzkt         nore, Zeno`, no problem if i add the following line in README.txt into "In case you downloaded the source code": If you want to use source code for your public server, please follow a stable-X branch instead of master, which could break your players' experience
14:21 nore          perhaps put something explicit about protocol incompatibility
14:23 nrzkt         then what do you suggests ?
14:31 nore          nrzkt: "the server on the master branch is incompatible with stable clients"
14:31 nore          just add something like that after what you say
14:32 nrzkt         could be incompatible ? because it's not everytime :)
14:33 nore          I suggest adding that when we merge the first incompatibility
14:33 nore          and removing it after the 0.5 release
14:39 T4im          since there's no backporting to the stable branches you should probably keep master as long compatible as possible and put the breaking changes into their own branch
14:39 nrzkt         backport could be possible, bugfixes and maybe some little functions
14:42 T4im          but unlikely to happen.. also keep in mind that master usually is associated with "current stable version" as its usually the default branch (and thus the default one when checking out), I would expect quite a few servers out there to run on some version of master due to that
14:42 nrzkt         informations have been posted about htat.
14:44 T4im          doesn't change my point :P just adds a warning sign to it
14:45 T4im          on the bright side, that might mean less mobile clients on servers for a while, hehe
14:46 nrzkt         no
14:46 T4im          there will be servers updating to current master, even if that breaks old clients (including mobile ones I assumed)
14:46 nrzkt         this means servers admins must be careful :) stable-0.4 contains all the compatiblity, and informations have been posted for how switch to stable-0.4 branch
14:47 T4im          still doesn't change my point about it still happening.. information provided or not :p
14:47 nrzkt         no problem :)
14:48 T4im          new mod features leads to mods using new features... requiring new minetest server versions.. so there will be servers updating just to use updated mods eventually..
14:48 T4im          happened in the past
14:49 nrzkt         you are right
14:51 nrzkt         then the modders must be careful and manage their compatibility layer too, because it's not our job, we will give high level API in 0.5 to improve mods performance, and it's why there is a break. In FOSS you couldn't keep a compatibility layer everytime, it's unmaintainable at a moment. Network layer is a great part, and because everything will be rewrite we couldn't provide a compatible layer. If you are an admin you
14:51 nrzkt         must be careful about your product.
14:52 T4im          saying "its not our problem" does not solve the problem :P in this case it creates it
14:53 nrzkt         maybe, but, because it's a FOSS you can create the compat layer if you want :)
14:53 nrzkt         i cannot provide a compat layer because _INIT packet must be changed.
14:53 celeron55     where's this forum post that was talked about?
14:53 nrzkt         https://forum.minetest.net/viewtopic.php?f=18&t=11289
14:55 T4im          nrzkt: I wish there was a way for a mod to add such a layer.. but I wouldn't even know how to querry what minetest version the server is running without adding heuristics, that you simply cannot demand from most modders... I think only debug builds offer that information to mods...
14:56 nrzkt         a mod couldn't manage network layer.
14:56 nrzkt         and modders are not affected by this network change, we are not modifying the LUA api
14:56 T4im          not yet ;)
14:57 nrzkt         you need to have a LUA api to know which is the MT server version ?
14:58 nrzkt         also we could add a thing to main.cpp to show the version and warn users
14:58 T4im          well depends.. as long as there's only lua api additions, there are easier ways to test.. but if there should be behavior changes (and I opened some issues to point out there are some required as they currently are not working as documented), you might have to
14:58 nrzkt         i don't think we must change the LUA calls, but add some newer
14:58 T4im          the api has broken calls atm
14:59 Zeno`         apparently there *is* a way to change the protocol without breaking protocol
14:59 Zeno`         I haven't looked though. But apparently the mechanism is there (or so I've been told)
14:59 nrzkt         oh T4im, maybe, but not because of me
15:00 nrzkt         in fact Zeno` protocol version handling at this moment isn't very reliable because it's based on packet length...
15:00 T4im          I know.. but your point was, mods should manage those cases if they are fixed after compatibility breakage
15:00 nrzkt         mods will not be affected by protocol break.
15:00 T4im          currently the only way to use those fixes is by ignoring any compatibility issues
15:01 T4im          they won't... but servers will in reuslt
15:01 Zeno`         I've been meaning to ask. When was the March 1st deadline/cutoff discussed and set?
15:02 Zeno`         nrzkt?
15:04 nrzkt         we can change it if you want :) but it was because the patch was discussed and over discussed and it's now ready for merge, and i will have time in march to do a great PR 4/4, after this will be difficult :)
15:05 Zeno`         I'd just like to know where it was discussed and how the decision was made :(
15:06 Zeno`         Because there are a few core devs here who knew nothing about that decision until they read the minetest forum
15:06 nrzkt         it was a little bit partial, because of my timetable :s
15:06 rubenwardy    #2320 does compile of me
15:06 ShadowBot     https://github.com/minetest/minetest/issues/2320 -- Fix superflous shader setting updates by gregorycu
15:07 rubenwardy    "error: ‘v3f’ has no member named ‘getAs3Values’" is the problem in travis, it is probably an issue with different libraries
15:07 rubenwardy    * versions
15:07 nrzkt         irrlicht 1.7, not 1.8
15:07 Zeno`         nrzkt, some warning or discussion might have been nice :(
15:07 nrzkt         then we can change this if you want, tell me
15:08 nrzkt         after the merge of network PR 3/4 we must merge the damage patch and the inventory patch too
15:09 Zeno`         What I am saying is, is that there seemed to be a lack of communication
15:09 Zeno`         but perhaps I was asleep when it was discussed
15:10 nrzkt         no, you drunk but doesn't forget anything :)
15:10 T4im          no drinking and deving, Zeno` :x
15:10 Zeno`         heh
15:11 nrzkt         xD
15:15 celeron55     nrzkt: i think we must have the incompatible version in a separate branch initially
15:16 celeron55     here are two reasons why:
15:17 celeron55     1) master has been compatible with the stable version since forever; people aren't going to catch up fast enough (this not a major reason, but a reason nonetheless)
15:18 celeron55     2) master has been compatible with itself since forever; this is probably impossible to get through to users who want to use a bleeding edge version
15:18 sfan5         ^ ack
15:18 nrzkt         then we should maintain 3 branches ?
15:18 sfan5         there is nothing to maintain in stable-0.4
15:18 sfan5         -> "only" 2
15:19 celeron55     i mean, the branch that includes the initial incompatible network changes is going to continuously change in incompatible ways because the problems aren't limited to the initial breaks
15:19 celeron55     if it doesn't need to continuously change, then  we don't actually have big enough problems to mandate a breakage in the first place
15:20 celeron55     a separate branch will also serve as a reality check for whether there actually are enough fixed problems to make it worth being incompatible
15:22 T4im          oh, that ^ is probably the best reason yet.. allows you to change network layer more flexible with continuous breaking changes without having to hold yourself back until the final break is merged into master, nrzkt
15:22 celeron55     once we agree that it indeed has improved things by an amount that makes it worth being incompatible, and we agree that it doesn't need to break much anymore, it can be rebased on top of master
15:23 nrzkt         okay celeron55
15:24 celeron55     it does not seem that there's much else going on regarding to the network layer, so keeping the network modification branch up-to-date doesn't seem like a too large issue
15:26 nrzkt         in fact some PR are related to protocol incompat and nobody must change the Send and read functions, and this wouldn't be too difficult
15:26 celeron55     even while mostly you are working on it, the branch should be in the organization repository so that others also can use the rare chance of radically fixing things
15:27 celeron55     and reviewing it becomes easier and whatnot
15:28 rubenwardy    #2375
15:28 ShadowBot     https://github.com/minetest/minetest/issues/2375 -- Fix superflous shader setting updates by rubenwardy
15:31 celeron55     something like "network-rework" might be a good branch name
15:32 nrzkt         okay, if you want celeron, i'll create it next week
15:35 rubenwardy    It works! #2375
15:35 ShadowBot     https://github.com/minetest/minetest/issues/2375 -- Fix superflous shader setting updates by rubenwardy
15:36 celeron55     once the branch exists, it becomes everyone's reponsibility to use it to fix long-standing issues and to judge when it's fixing enough things to be worth the incompatibility
16:00 celeron55     someone asks by email for the option to turn player name visibility completely off
16:00 celeron55     maybe someone could implement that, altough it could always be cheated with a modified client
16:01 celeron55     they want to play hide and seek 8)
16:02 VanessaE      actually, bring able to turn player names off with a keypress is useful when taking screenshots
16:02 rubenwardy    Related: https://github.com/minetest/minetest/issues/1345
16:02 VanessaE      (e.g. for a "server overview" image, since it's not hard at all for players themselves to get lost in the image, but the player names stand out regardless)
16:03 nrzkt         #2376
16:03 ShadowBot     https://github.com/minetest/minetest/issues/2376 -- Add find_surface_nodes_in_area LUA call which permit to only get the nodes which are under air nodes by nerzhul
16:07 nrzkt         please review it, it permits to improve performances of plantlife, but not only if there are other mods which needs to know surface nodes
16:31 nrzkt         celeron55: i changed the topic on the forum :)
19:22 celeron55     somebody who says he is a noob and uses dali linux asks by email for an x86 binary package with libraries included
19:22 celeron55     why is a noob using dali linux, is my question
19:23 celeron55     well anyway, if somebody is interested in making such a package, there's one person in this world who needs it
19:27 shadowzone    x86..
19:28 celeron55     he also says blender has such a package and it works for him
19:32 EvergreenTree I was wondering, should something as small as #2378 warrent a pull request?
19:32 ShadowBot     https://github.com/minetest/minetest/issues/2378 -- Add src/network/CMakeFiles/ and src/client/CMakeFiles to gitignore by 4Evergreen4
19:34 kilbith       does anyone can disable the Debug build option on Launchpad please ?
19:34 kilbith       kind of slowness
19:36 est31         kilbith, you mean at startup?
19:36 kilbith       yes
19:36 est31         thats because of the unittests
19:36 est31         nrzkt is working on a solution
19:36 est31         but right now there is a discussion on whether to compile them into release builds
19:37 kilbith       and perhaps a bit less FPS than my own compiled build
19:37 kilbith       (Release type)
19:38 est31         https://github.com/minetest/minetest/commit/b3aeba684b4e27e9fdf2c07949750bb567e40d6b
19:39 kilbith       that build is intrinsically slower than the release one anyways
19:39 kilbith       "Debug build is slower, but gives much more useful output in a debugger."
19:39 kilbith       > http://dev.minetest.net/Compiling_Minetest
19:40 est31         thats true
19:41 * est31       wonders whether it might be a good idea when the network breaks to discontinue launchpad daily build repo, or to force a manual step.
19:41 est31         as in renaming the package
20:45 manax         Can someone explain the separation of code in player vs localplayer?  I'm looking at getStandingNodePos(), which seems like it should be in player, but it requires crouch stuff, which also seems like it should be in player... So, wondering.
20:53 Wayward_One   i think i may have found another memory leak... almost ran out of memory after 30 minutes of playing, recent git, only minetest and irc running
20:53 * Wayward_One git pulls to be sure
21:28 Wayward_One   yep, seems to be a memory leak... just git pulled, joined a server about 10 minutes ago, and memory use is up from about 400Mb to 1.8Gb, and it's gradually climbing
21:36 est31         it seems that the reason #2378 travis failed was because of an offtime of one of sfan5's servers: https://travis-ci.org/minetest/minetest/jobs/51748219#L4425
21:36 ShadowBot     https://github.com/minetest/minetest/issues/2378 -- Add src/network/CMakeFiles/ and src/client/CMakeFiles to gitignore by 4Evergreen4
21:49 sfan5         est31: i restarted them
21:54 est31         sfan5, ok,added comment so that nobody thinks this is the PR's fault
22:27 Fritigern     Interesting, MT4.11 will be in the Ubuntu 15.04 repositories. I guess that's a lot better than it was...
22:28 est31         but its not 0.4.12 :(
22:31 Fritigern     Of course not. Because there is not 4.12 yet
22:33 kilbith       Fritigern, don't be so sure of that ;)
22:34 Fritigern     I just pulled and built, and I am at 4.11-255-gcd4324e
22:34 Fritigern     So yeah, i am very sure
22:34 est31         thats because of an error
22:34 est31         there is 0.4.12
22:34 est31         ~topic
22:34 ShadowBot     est31: Minetest core development and maintenance. Chit-chat goes to #minetest. Consider this instead of /msg celeron55. http://irc.minetest.ru/minetest-dev/ http://dev.minetest.net/
22:34 est31         ~topic #minetest
22:34 ShadowBot     est31: Welcome to #minetest, the official channel of Minetest | Version: 0.4.12 (2015-02-18) | Responses may take a while, so be patient. | Rules: http://wiki.minetest.net/IRC#Rules | Development: #minetest-dev | Server list: http://minetest.net/servers | IRC logs: http://irc.minetest.ru/minetest/
22:34 est31         see?
22:35 est31         Fritigern, its issue #2357
22:35 ShadowBot     https://github.com/minetest/minetest/issues/2357 -- Make git version discovery work on master by est31
22:36 Fritigern     Wow, what a mess