=== welcome === <hoijui>zerver missing, and kloot.. we dont know if will come <hoijui>hey <abma>hi all! <zerver>hi <abma>anyone heard something from kloot? <abma>i only read his... a bit confusing message <abma>oh, before i forget it: Happy new year! <zerver>indeed, happy happy <Tobi>happy new year <jK>or in noppon: "new year! happy happy everyone!" <hoijui>have a good one! <hoijui>so... <hoijui>we folow the etherpad? <abma>yep?
=== merge or not https://github.com/spring/spring/pull/25 ? === <abma>we already talked about that, but afaik we didn't come to a decission <jK>better make it a module? <jK>none of us would maintain/use it <abma>thats reasonable, yep <abma>other thoughts? <abma>if not... next point <abma>+ thanks... i personally had no real opinion about the pull request <zerver>so it is to allow python to invoke unitsync? <jK>something like that <abma>if i understood it right, it will generate a python module which is unitsync <jK>currently python has to wrap the library and with each update it needs a new wrapper, with that branch it would be a python native module <zerver>ah
conclusion: will be added if its a submodule (as current devs can't maintain / use it)
=== Release plan === <zerver>we need 85.1 with MT fixed <jK>there are many unsynced bugs currently <zerver>yeah, I have seen some desyncs <jK>unsynced <jK>not synced <jK>the synced issues are ... better don't comment it <zerver>okay <zerver>pathfinding is... <jK>pathfinding is fine <abma>http://springrts.com/mantis/view.php?id=2865 ? <abma>( Excessive movement bouncing off buildings and wreckages ) <zerver>yeah, the obstacle avoidance/bounding is a bit weird <zerver>and it goes into infinite loop sometimes, bounding forever between two obstacles <jK>the bouncing is not pathing related <zerver>I will make MT fixed EXE pretty soon since I don't know when the other unsynced bugs will be fixed <jK>will possibly fix them this week <zerver>are they all on mantis? <jK>no <jK>there are 2: intel gpu detection fails + SPRING_DATADIR isn't used in isolated mode <jK>(even in isolated mode the user should be able to define multiple readonly directories, so isolated should only mean that any _automated_ directory including is disabled) <zerver>yeah <zerver>I did not know we had intel detection <jK>new in 85 <zerver>if (intel) exit(-1); <jK>to disable glsl shaders & loadingMT <jK>their newest generation is able to run spring at decent FPS <jK>still it fails in many ways <zerver>I'm wondering if the LoadingMT driver issue is the same as the glShareLists issue in MT <jK>possibly still need to add a warning messagebox at first start <jK>no, LoadingMT, issue is that Mesa fails at garbage collection <jK>-, <zerver>okay <jK>so when the loading opengl context is destructed, it also deletes the textures (or something similar, fact is that the textures are invalid after doing so) <Tobi>FLOZi is inquiring about http://springrts.com/mantis/view.php?id=2873 <Tobi>in short Q is: is it an intended change that if AllowUnitCreation returns false, factory just keeps trying again and again to build that unit until it finally returns true (which may be never) <Tobi>? <Tobi>or is intended behavior for the factory to skip that unit if AllowUnitCreation returns false, and go on with next order? <jK>it was a undocumented change <Tobi>intentional or accidental change? <jK>only kloot knows <Tobi>baww <Tobi>ok <Tobi>I say it is a bug then, not a feature
conclusion: many bugs should be fixed before a release
=== piwik website integration === <Tobi>abma_irc: I fixed up and merged that piwik pull request and its running now <abma>nice! <abma>uhm, how to login? <Tobi>seems I need to add users as superuser <abma>what can users do, expect adding other users? <Tobi>adminning other stuff maybe <Tobi>though I guess looking at it is most useful activity right now
conclusion: tobi set up piwik, so we have soon some hopefully nice stats... and you are tracked! :)
=== luasocket integration === <abma>btw: i tried to add luasocket to spring <abma>but imo it would be better to write the required functions ourself as it would be so much work, to restrict access to the existing functions <abma>also some workerthread would be required, to make it easily useable <abma>it could be used as it is, as it seems all functions can be run non-blocking <abma>but i guess this would make the lua-code ugly / hard to maintain <abma>https://github.com/abma/spring/commits/luasocket <abma>but uhm... imo not very useful this way <abma>i would prefer it to be done with some lua-callbacks, like: DataReceived(connectionid, data) or so and use a worker thread <abma>also luasocket seems to not support ipv6 <Tobi>no worker thread needed, just use O_NONBLOCK and and check every frame <Tobi>to call a callin there must be synchronization anyway <abma>hmm <abma>if i understood the docs of luasocket right, nonblock can / is already be done... <Tobi>yeah but you wanted to add call-ins on top <abma>i'm unsure what to do <Tobi>well <Tobi>it may give a nicer API, but then those other Lua-only parts of luasocket aren't usable <abma>for me it currently looks like with this implemntation many has to be coded on the lua side <Tobi>though then again they are quite probably not really usable anyway because they are blocking <Tobi>at least in my quick scan of the docs I didn't see a way to do http/smtp/ftp in non-blocking way <abma>yeah, this has to be rewritten non-blocking in lua... <abma>maybe i should just try to write a test-download widget <Tobi>then again that whitelist thing is also still open <abma>with socket.select() it should be possible to make it nonblocking <abma>yep <Tobi>and it seems pretty hard to solve decently <Tobi>yes <Tobi>don't even need select if you wait for one socket only <abma>just call receive with zero timeout? <Tobi>yep <Tobi>if it returns EWOULDBLOCK (or whatever Lua equivalent is) then there is no data <abma>ok, will try then the next days <abma>and i know about the security problem <Tobi>I don't really know what to do with this whitelist, just opening sockets completely seems unresponsible, but whitelist can almost not be stored anywhere <Tobi>cause Lua is a leaky bucket already <Tobi>config -> can be overwritten by Lua <abma>yep, i already added a mantis report about that <Tobi>any file -> can be overwritten by Lua <Tobi>I thought in the past there was already a check for file access to be limited to spring dir <abma>yep, triggers an assertion currently <abma>is it possible to use include() in a lua-widget? <abma>VFS.include() + "absolute" vfs path VFS.include('LuaUI/Widgets/include/myinclude.lua') for example <Tobi>not sure <Tobi>gn <abma>good night! <abma>uh, sorry was confusing, this is how it seems to work...
conclusion: luasocket integration seems be reasonable
Last edited by abma on 08 Jan 2012, 18:05, edited 1 time in total.
<zerver>pathfinding is... <jK>pathfinding is fine <abma>http://springrts.com/mantis/view.php?id=2865 ? <abma>( Excessive movement bouncing off buildings and wreckages ) <zerver>yeah, the obstacle avoidance/bounding is a bit weird <zerver>and it goes into infinite loop sometimes, bounding forever between two obstacles <jK>the bouncing is not pathing related
Good to see someone has the record straight, good on ya JK. I have not seen this bug but people are mighty buttmad about it. IS there a fix coming?
how do I trigger this bouncing? i have not seen it yet.
just play some games, eventually you will notice units being thrown around when exiting factory or just bumping into stuff.
Was not as funny as I expected... bouncing is not appropriate.. it is closer to headbutting or humping. I was thinking bouncing like the ghey stuff that happens when a unit lands after falling. This is just a unit being pushed back... lame.
Joined: 22 Feb 2006, 01:02 Location: cheap kitchen
Quote:
it is closer to headbutting or humping.
That is what is supposed to happen when it works normally.
trap (surround) a commander with eg laser towers and use übermicro to make him step over the buildings.
also this: yes, i tested: it can happen in gundam too, mech walked into a turret and got bounced some hundreds meters. (too lazy for gif, next engine version probally fixes it anyway so enjoy pw-cannons while you still can!)
The bouncing problem has been reduced to rapid jittering with the occasional several elmo jerk. This is fine, we've never had much better behaviour in a release without unit that embed themselves in structures.
Users browsing this forum: No registered users and 1 guest
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum