These are chat archives for MylesIsCool/ViaVersion

27th
Mar 2016
Hugo Kerstens
@hugokerstens
Mar 27 2016 00:12 UTC
Just noticed I created my account just 2 days prior to you ;)
fillefilip8
@fillefilip8
Mar 27 2016 09:09 UTC
rip pc
Mats
@Matsv
Mar 27 2016 09:37 UTC
Again?
Myles
@MylesIsCool
Mar 27 2016 12:11 UTC
just realised the stupidly asked questiosn got wiped, now restored it
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:19 UTC
Did you have to write it again or did you save the text somewhere?
Myles
@MylesIsCool
Mar 27 2016 12:23 UTC
I copied for google cache
like rest of post
How do you leave a review
on a resource
Mats
@Matsv
Mar 27 2016 12:25 UTC
Uh on the bottem of the page
you see stars
Myles
@MylesIsCool
Mar 27 2016 12:26 UTC
ahhh
thanks
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:30 UTC
How long until we push apiv2? I think it's quite stable, and there have been a lot of options/fixes which have only been added to apiv2 ;)
So even though there can be some minor glitches, the new bug fixes compensate this easily
Myles
@MylesIsCool
Mar 27 2016 12:37 UTC
so
what we doing for april fools
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:38 UTC
We should implement this already
Bring herobrine back?
Myles
@MylesIsCool
Mar 27 2016 12:39 UTC
just puts in code
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:39 UTC
The ghast sound
Myles
@MylesIsCool
Mar 27 2016 12:39 UTC
if april 1st
then randomise sent entity ids
;)
(the types)
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:39 UTC
I think servers will be happy ;)
Myles
@MylesIsCool
Mar 27 2016 12:39 UTC
the big ones especially
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:40 UTC
Yeah i can see the donations coming already
Myles
@MylesIsCool
Mar 27 2016 12:45 UTC
I would like to get the issue with the CannotCastException
fixed for APIv2
:(
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:46 UTC
ClassCastException you mean?
Do we have any stacktraces where this exception occured in apiv2?
Myles
@MylesIsCool
Mar 27 2016 12:47 UTC
well it happens in both no doubt about it
I think I might know the cause
nah wasn't right
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:52 UTC
And what about the deadlock with our bossbar api?
Myles
@MylesIsCool
Mar 27 2016 12:52 UTC
I don't have an example that deadlocks it
I thought the cast issue was caused by
info.getChannel().pipeline().fireChannelRead(idlePacket);
need someone to test with simulate-pt off
there's a chance it still could be
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:53 UTC
I'm not at home rn
Myles
@MylesIsCool
Mar 27 2016 12:53 UTC
well when someones free
Hugo Kerstens
@hugokerstens
Mar 27 2016 12:58 UTC
Isn't the issue with the sendRawPacket method? Don't we need to release the buffer after we sent it?
Mats
@Matsv
Mar 27 2016 12:58 UTC
Netty releases the bytebuf
Myles
@MylesIsCool
Mar 27 2016 12:58 UTC
netty does that
there's a whole page dedicated to when to release it
it's a very fun exciting page
Hugo Kerstens
@hugokerstens
Mar 27 2016 13:00 UTC
Ah yeah I just found an issue reported by md_5 that it wasn't the case, but it should be fixed indeed
Myles
@MylesIsCool
Mar 27 2016 13:00 UTC
hm?
wheres md_5
Hugo Kerstens
@hugokerstens
Mar 27 2016 13:01 UTC
netty/netty#1294
But it is outdated now ;)
Myles
@MylesIsCool
Mar 27 2016 15:36 UTC
I can't figure out this issue to save my life
The issue is when you're using something like armour
it doesnt fire player interact
ugh
If anyone wants to try at it go for it I believe something is still wrong with the place block packet
Mats
@Matsv
Mar 27 2016 15:39 UTC
Hm you mean equiping armor by right-clicking?
Myles
@MylesIsCool
Mar 27 2016 15:39 UTC
Well using action with right click doesnt fire interact event usually
Mats
@Matsv
Mar 27 2016 15:39 UTC
Does it fire if you right-click the ground with that item?
Myles
@MylesIsCool
Mar 27 2016 15:42 UTC
No not at all
Maybe I need to fire packets twice or something when using
Hugo Kerstens
@hugokerstens
Mar 27 2016 17:09 UTC
#279 @MylesIsCool
Myles
@MylesIsCool
Mar 27 2016 17:32 UTC
ty
Hugo Kerstens
@hugokerstens
Mar 27 2016 19:39 UTC
#280
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:06 UTC
I know we don't support reloads, but what on earth is this:
http://pastebin.com/015992wx
java.lang.ClassCastException: us.myles.ViaVersion.ViaVersionPlugin cannot be cast to us.myles.ViaVersion.ViaVersionPlugin
Myles
@MylesIsCool
Mar 27 2016 20:33 UTC
Essentially it's because
reloads cause 2 diff class loaders
which then cause issues
:(
ViaVersion release apiv2 tonight? ;)
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:38 UTC
Yeah! :D It's been a week since an update
Won't we add metrics? :( I'm quite curious to know how much we are used exactly ;)
Myles
@MylesIsCool
Mar 27 2016 20:41 UTC
I dislike metrics
cause it doesn't have good uptime
I always see error logs with 500 errors for it :(
I used to use it
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:43 UTC
Hahahah if you don't like it that's okay ;) Just a question
Myles
@MylesIsCool
Mar 27 2016 20:44 UTC
<3<3 I would if it was stable ;)
Time to write a changelog
braces self
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:45 UTC
Hahah yeah there are quite some things ;)
How do I get the server protocol number?
Myles
@MylesIsCool
Mar 27 2016 20:48 UTC
ProtocolRegistry.SERVER_VERSION
something like that
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:49 UTC
Ah thanks, knew it was somewhere ;)
Myles
@MylesIsCool
Mar 27 2016 20:49 UTC
It's resolved on boot though quite late
it's -1 until resolved
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:49 UTC
Yeah I know
Myles
@MylesIsCool
Mar 27 2016 20:49 UTC
it's late cause it works based on server having serverping init'd
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:49 UTC
Yeah I know, I studied that code quite a bit
But, shouldn't we in the future keep all players ported, so that we can modify packets of player using the server version
Because for example a server updates to spigot 1.9, they are not able to use the bossbar api anymore
Myles
@MylesIsCool
Mar 27 2016 20:52 UTC
shrug their fault ayoooyoyoyo
I'm not sure hm
Hugo Kerstens
@hugokerstens
Mar 27 2016 20:53 UTC
Yeah it could cause more issues as we can fuck up 1.8 player connections
But it could be a config option ;)
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:07 UTC
Maybe update the 'slogan', or whatever it is called ;)
Connect to a 1.8 server via a 1.9/1.9.1 client
Myles
@MylesIsCool
Mar 27 2016 21:08 UTC
what needs changing
:)
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:08 UTC
Well it is for 1.9 servers now as well ;)
Why can't people read, seriously.............
Myles
@MylesIsCool
Mar 27 2016 21:13 UTC
stab stab stab
on plus side we have a nice api
so one of the guy has issues with lilypad not working properly, and I used the via api to easily patch that, (You can simply patch a protocol to itself!)
it's like having protocol lib lol
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:24 UTC
Yeah we can do essentially the same as protocollib/protocolsupport
Myles
@MylesIsCool
Mar 27 2016 21:26 UTC
protocolsupport is interesting
it converts packets to like a middle packet
:P
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:26 UTC
Yeah but that can be way less efficient in some occasions
Myles
@MylesIsCool
Mar 27 2016 21:27 UTC
Yeah but it's aim is to like back support
;)
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:29 UTC
Yeah that's true. But if we are like 2 versions further, we are quit inefficient as well. Then the packets would have to go through like 3 transformers
But we could always make more direct transformers, for example for 1.8 to 1.10
Myles
@MylesIsCool
Mar 27 2016 21:30 UTC
Realistically
the performance between doing it direct
and doing 1.8 -> 1.9
and 1.9 -> 1.10
shouldnt be too much
mainly because packetwrapper allows reusing without reading from bytebufs
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:31 UTC
Yeah that's a nice addition ;) Are bytebuf so inefficient then?
Myles
@MylesIsCool
Mar 27 2016 21:32 UTC
Well the main issue is
if you had multiple pipelines
without wrappers you'd end up
having to make new bytebufs everytime
and that's expensive cause bytebufs are really fancy byte arrays
The aim of the protocol api was to remove the layer of directly using bytebufs (I know we still use them a few places but we'll work on that and one day not at all haha)
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:34 UTC
Yeah the improvements are very very large
They may seem small, but as soon as we have pipelines of multiple protocols it's really useful
Myles
@MylesIsCool
Mar 27 2016 21:34 UTC
I mean I haven't fully profiled this version but I believe it to be somewhat memory efficient
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:36 UTC
Well the actual operations we are doing are not very intensive, so compared to some other plugins we are nothing when comparing the used resources
Myles
@MylesIsCool
Mar 27 2016 21:43 UTC
I just realised
on git
i'm on a 29 day streak
damn myles
good work
Hugo Kerstens
@hugokerstens
Mar 27 2016 21:46 UTC
Hahahah nice! :D
I didn't do a commit each day :(
Hugo Kerstens
@hugokerstens
Mar 27 2016 22:41 UTC
Version 0.7.0 has already around 10% of the 0.6.7 requests at spiget.org ;)
Ughh that phrasing, but ~1700 requests voor 0.6.7, and 170 for 0.7.0
Myles
@MylesIsCool
Mar 27 2016 22:43 UTC
;)
Mats
@Matsv
Mar 27 2016 23:33 UTC
Oh I missed the party
rip
But nice, 15+ likes already :+1: