c88ca10d26
Provide a means of adjusting the select loop time.
Andrzej Kapolka
2009-11-06 22:00:30 +00:00
00c4777460
Example text.
Ray Greenwell
2009-11-06 21:50:21 +00:00
d82a926ca1
Add an example to the docs.
Ray Greenwell
2009-11-06 21:18:43 +00:00
a707c8b515
Set the traffic class for datagrams to low-delay. I doubt this will actually reduce latency, but wouldn't it be cool if it did?
Andrzej Kapolka
2009-11-06 20:56:37 +00:00
ee27342a24
Proposed change to readObject().
Ray Greenwell
2009-11-06 01:27:03 +00:00
c16de5994e
Moved these here to narya. Lately I'm trying to do things more "actionscripty", and so we just use the "natural" iteration on an object (either 'for' or 'for each'). Plus, we never use these Iterators anywhere except in presents-based code.
Ray Greenwell
2009-10-27 00:57:48 +00:00
6137fbfb79
As Charlie points out, we can use the protected constructor if we're creating an anonymous extension of DefaultMap and any other creation is better done through the static factory methods, so don't present the user with a pointless choice.
Michael Bayne
2009-10-23 04:58:49 +00:00
b124934a97
MethodQueue -> DelayUtil.
Ray Greenwell
2009-10-21 21:49:17 +00:00
90c97a54b3
Turns out USR1 is reserved by the JVM on pretty much every platform but Mac, so only expose USR2 for injected handling by subclasses.
Charlie Groves
2009-10-21 20:47:55 +00:00
9284f1b4b6
Expose USR1 and USR2 signals to users of AbstractSignalHandler if they bind instances of a SignalReceiver class for those signals.
Charlie Groves
2009-10-19 20:58:03 +00:00
644de92f7b
toString() is being overridden, mark as such.
Mike Thomas
2009-10-08 20:28:09 +00:00
ecec43418b
Fluency--.
Michael Bayne
2009-10-04 23:26:09 +00:00
5f3f1ad5c9
As Ray points out, we don't need a special interface for this. The creator is simply a function from K to V. Sorry to break users but I know you're few and being able to integrate with all the excellent Google Collections functional stuff is a juicy reward for a little egg breaking.
Michael Bayne
2009-10-01 01:07:28 +00:00
dd15495810
Avoid NPE for final datagrams sent/received after client cleanup.
Andrzej Kapolka
2009-09-28 21:51:08 +00:00
8cbc3fcf1a
Throttle datagrams as well as streamed messages.
Andrzej Kapolka
2009-09-27 02:23:17 +00:00
2d044ec3f5
Fix the docs now that fetch has shown up and break out a factory for instance Creators separate from newInstanceMap.
Charlie Groves
2009-09-26 00:35:13 +00:00
c42cb490cf
Every "case" should have a "break", even the last one.
Ray Greenwell
2009-09-24 21:17:15 +00:00
ea49d70129
* Add Msoy's MethodProfiler so I can use it in Riposte. * Apparently, a couple of whitespace cleanups my gvim did awhile ago while I was looking around in here.
Nathan Curtis
2009-09-24 20:40:48 +00:00
1c538ac78e
Drive-by logging modernization.
Ray Greenwell
2009-09-24 20:18:14 +00:00
dae0f7f44f
Fuckit, let's use uint paramters and make the default uint.MAX_VALUE. Unless I have to roll this back because the like compiler in CS3 shits its pants trying to understand it (external references for default parameter values seem to sometimes fail).
Ray Greenwell
2009-09-24 20:13:39 +00:00
3f2468e7a1
Oh crap, let's make ObjectOutputStream protect against fuckup, too. Note that we're using ints instead of units for offset/length. Does anyone really need a byte array larger than 2^31? And if so, wouldn't you then also be within sight of busting 2^32? Why not make it a Number and support up to 2^53? Number is no guarantee of a desire for a floating-point value in the flash libraries, see Graphics.lineStyle()'s first parameter. Actionscript, you fail for not having long. And... I said I wasn't going to bitch at Adobe/Macromedia any more but this is a total FAIL. 0 is a valid length. Everyone knows this. Can you imagine if they picked a different constant? length: the number of bytes to read, unless it's 9, in which case it reads *all* the bytes. Your code should check to see if you're about to read 9 bytes, and instead read 8 in one go and then a follow-on byte. uint.MAX_VALUE would have made tons more sense.
Ray Greenwell
2009-09-24 20:09:35 +00:00
33d19f7a69
Another fix for potential streaming errors.
Ray Greenwell
2009-09-24 19:37:34 +00:00
730d448e40
As far as I can figure, ArrayMask objects are used even when the Array length is 0, so let's make sure we don't cause any streaming errors here.
Ray Greenwell
2009-09-24 19:35:35 +00:00
a6a008cb50
Turns out we want to highlight the difference between normal getting and getting that adds the default value to the map. I opted for fetch() which at least means something close to get() and can be documented as also inserting the default value. I also considered obtain() and acquire() but those both felt like they had something to do with locks. apply() seemed too functional. wangle() seemed like a strong contender for a while.
Michael Bayne
2009-09-22 20:25:05 +00:00
3c159a6497
Widened
Charlie Groves
2009-09-22 00:05:18 +00:00
f75d46f75f
Add a better toString, for debugging.
Ray Greenwell
2009-09-17 23:16:22 +00:00
53f1dff95d
These are all final.
Ray Greenwell
2009-09-17 21:58:30 +00:00
f9bc385c99
Some examples.
Ray Greenwell
2009-09-17 21:53:02 +00:00
a24b0d2d03
So, these classes like Integer and langBoolean exist solely for streaming to our java server, so let's just make them streamable directly. Does away with the need for custom Streamers. Also made them Hashable, in case we read a StreamableHashMap containing Integers as keys, for example. Less code is better code. (Why was BooleanStreamer added the other day? langBoolean was already Streamable and was already in Translations)
Ray Greenwell
2009-09-17 21:42:25 +00:00
055aac9b55
Use Joiner instead of StringBuilder, in some places. Use String concatination in others (it's faster in actionscript than in Java)
Ray Greenwell
2009-09-16 21:13:07 +00:00
9f967dd575
This is already output by SimpleStreamableObject...
Ray Greenwell
2009-09-16 20:03:53 +00:00
e0a4340947
add missing BooleanStreamer
Tom Conkling
2009-09-15 21:53:26 +00:00
293ec52255
Some Galileo inspired raw type and dead code elimination.
Michael Bayne
2009-09-14 22:20:16 +00:00
26d982ae54
Document those so eclipse with strict docbook warnings doesn't get cranky that they're never used.
Dave Hoover
2009-09-14 22:18:31 +00:00
66af5d397b
There's no especial need to disallow use of the constructor.
Michael Bayne
2009-09-12 01:17:49 +00:00
bba467be39
Slight description editing.
Michael Bayne
2009-09-12 01:15:53 +00:00
9b946dfbd5
DefaultMap extracted from Yohoho, simpl- and and MDB-ified. For Charlie's inspection and approval and Matt's eventual use.
Michael Bayne
2009-09-12 01:13:19 +00:00
32db838be2
Hmm don't throw the ArgumentError inside our own catch, even though we should have re-thrown it.
Ray Greenwell
2009-09-11 16:49:37 +00:00
4fe12a043d
Even when fed an object graph with no loops we could have a stack overflow in our recursive checking. Try to catch it and cope. I have a bad feeling this won't actually work in some envs.
Ray Greenwell
2009-09-11 15:30:43 +00:00
f4ab06b619
Integers and Booleans are final. This fixes the Integer[] streaming problem I was having with Bite Me.
Tom Conkling
2009-09-11 02:13:03 +00:00
00569e651f
Added a system for tracking the messages sent to and received from the server for the purpose of statistics tracking.
Andrzej Kapolka
2009-09-11 00:19:10 +00:00
05e37f4f42
This log message is not strictly necessary. The connection manager immediately logs that it's listening on its configured ports, when start() is called. It does so on a separate thread, but I don't believe the subtle ordering possibilities there are something that's actually causing/likely to cause problems and/or needs to be called out.
Michael Bayne
2009-09-10 19:42:18 +00:00
179c66f46a
Missed a spot.
Michael Bayne
2009-09-10 19:39:03 +00:00
03ff9f8220
PresentsDObjectMgr.isRunning() is really isNotShutdown() and we need to preserve those semantics. Sigh.
Michael Bayne
2009-09-10 19:38:36 +00:00
f92350f1b3
Bah, don't include the PresentsDObjectMgr's state when trying to assert our invariants.
Michael Bayne
2009-09-10 19:36:33 +00:00
a8d3227fb5
You too. Naughty misreporter.
Michael Bayne
2009-09-10 19:32:06 +00:00
1c60b9cc75
Promote the datagramHostname to the bindHostname and use it for our TCP connections as well.
Michael Bayne
2009-09-10 18:34:42 +00:00
701d885222
Keep track of the total number of units run independently of Stats so it is not reset every 3 minutes by pulse.
Jamie Doornbos
2009-09-09 23:59:02 +00:00
6f7c9831d8
Comment tweak.
Par Winzell
2009-09-09 14:50:30 +00:00
f3fd4d28a5
We need to use the thread context classloader if one has not been specified.
Michael Bayne
2009-09-07 20:39:32 +00:00
2acedb91a4
Better protection against divide-by-zero - if we have less than a second, we'll just round up to 1 second.
Mike Thomas
2009-09-04 22:42:27 +00:00
3aec11f519
This is cloggin' the bureau loggin'. I'm pretty sure it's the BaseGameConfig. Typically, clients don't add directly to a StreamableHashMap received from the server, why would they? So let's just treat it as a String-key map and shut it up.
Ray Greenwell
2009-08-30 20:40:37 +00:00
bf5d3c666f
Avast, thar be bugs here too.
Ray Greenwell
2009-08-30 00:37:08 +00:00
e27f0ed5e2
No more setHours
Charlie Groves
2009-08-26 22:51:48 +00:00
90e11d00b7
More fluency, less effluent
Charlie Groves
2009-08-25 23:14:59 +00:00
7a95df0c26
Added import for javadoc.
Michael Bayne
2009-08-25 22:57:34 +00:00
2b5b0389e0
These classes moved from flash-utils back into narya. Not generally useful to your everyday flash coder. MessageBundle and MessageManager are built on top of Flex Resources, and were already being excluded from our asdocs, and TimeUtil uses MessageBundle.
Ray Greenwell
2009-08-25 19:42:47 +00:00
da0e42a2b4
Inline references to other classes? That is not the actionscript way.
Ray Greenwell
2009-08-25 19:24:21 +00:00
5e0eb1d970
Fix the asdocs here. But wait, why is this even showing up in the SDK docs?
Ray Greenwell
2009-08-25 18:56:18 +00:00
1083490e73
Use newMapOf().
Ray Greenwell
2009-08-24 21:36:19 +00:00
f8d5e679fd
We use StreamableHashMap sensible and only store one kind of key. Let's guess the type based on the first key unserialized, and guess if it was empty.
Ray Greenwell
2009-08-24 05:22:05 +00:00
c9ea3709f7
Use new Maps factory. The old HashMap class was a frankenstein one-size-fits-all mess of: - storing simple objects - storing Hashables - having the capability to provide external functions to hash and check equality of any other type of key.
Ray Greenwell
2009-08-24 02:54:23 +00:00
e6fa8e77f1
Obsolete.
Ray Greenwell
2009-08-22 00:12:51 +00:00
6aa154575e
Look ma, just lib.
Ray Greenwell
2009-08-21 23:30:11 +00:00
24c16141f6
I do not think we need this anymore.
Ray Greenwell
2009-08-21 22:54:12 +00:00
b5de96dc83
It's an ecological disaster! I'm trashing the Environment! No more *-env libs.
Ray Greenwell
2009-08-21 22:46:50 +00:00
44480ada4a
Use ClassUtil.isAssignableAs(). It works in thane now.
Ray Greenwell
2009-08-21 22:45:36 +00:00
655b70b0e2
Don't include the ooolib classes in our lib, right?
Ray Greenwell
2009-08-21 21:30:35 +00:00
4848fe4b29
Now that ByteEnum has been liberated from depot, lets support streaming them as bytes. It's better than writing a modified UTF String, but it'd be even nicer if we could avoid the short to read the class name (Enums are final), but I suppose we still have to know whether one is null or not. Hm.
Ray Greenwell
2009-08-19 02:18:14 +00:00
00416168bb
Nix some warnings. One shouldn't be coped with, and so let's let it NPE. The other happens now in normal circumstances in msoy, and I'd rather not surround the call to super with an if to avoid this warning, and potentially kablooie in the future if other functionality is added to leftLocation in this base class...
Ray Greenwell
2009-08-18 21:34:51 +00:00
b04e27f8fa
Prune spurious semicolons
Dave Hoover
2009-08-17 21:45:13 +00:00
a3b20e176b
Scratched an itch... Removed the unholy 'alreadyDisplayed' hack and associated return value from ChatDisplay. The localtype should be used to determine whether something should be displayed. If not, why should ChatDisplay registration order matter? It shouldn't. If there are two ChatDisplays that could potentially show the same message and you only want one of them to, sort that out in your own fuckin' code. I'm not changing the Java side right now because I'd have to patch-up a bunch of projects...
Ray Greenwell
2009-08-15 00:06:12 +00:00
1b17350ee0
Allow command handlers to be unregistered.
Ray Greenwell
2009-08-14 23:45:47 +00:00
6e69b98459
Tweaky tweaky.
Michael Bayne
2009-08-14 22:29:02 +00:00
4a734d00ef
Drive-by logging varargification. Don't ask what I'm doing driving in this neighborhood.
Ray Greenwell
2009-08-14 22:12:56 +00:00
ee4369df0f
Freak out not if we have no client reference.
Michael Bayne
2009-08-14 21:36:13 +00:00
04316d8620
My actionscript-fu has improved since I originally wrote this.
Ray Greenwell
2009-08-14 03:10:48 +00:00
49f5d7fed8
We don't read creds on the client.
Ray Greenwell
2009-08-13 23:34:12 +00:00
e469bb4abc
I'm pretty sure we never read Credentials on the client...
Ray Greenwell
2009-08-13 23:32:30 +00:00
7489a64555
Every byte is precious! No need to tell the server the callerOid, it already knows it. This saves 4 bytes for every listener you pass into an invocation request.
Ray Greenwell
2009-08-13 19:15:21 +00:00
8284198e57
Don't do it twice. Yes, I'm embracing making things more "actionscripty", including some implicit coercian to a Boolean, which can actually be quite dangerous... but only with Strings and Numbers and whatnot.
Ray Greenwell
2009-08-13 06:23:27 +00:00
ac2a633047
Regenerated services, and the callers that love them.
Ray Greenwell
2009-08-13 06:18:26 +00:00
20f3c6add6
Change how the actionscript services are generated, so that the client is not required.
Ray Greenwell
2009-08-13 06:17:57 +00:00
fe3d89132d
A little bit of tweakery so that the invocation director to use is known by any invocation marshaller, and so the client does not need to be passed in.
Ray Greenwell
2009-08-13 06:17:28 +00:00
505a9c3a53
Make MessageBundle.getResourceString public so MessageBundles can be composed instead of requiring inheritance to customize them, and open up the bundle creation process in MessageManager to subclass customization
Charlie Groves
2009-08-12 20:04:36 +00:00
8ca961fc9d
Wait for everything to clear before starting the connection manager, and split out the code called when starting said manager so subclasses can override that
Charlie Groves
2009-08-12 20:02:01 +00:00
4780fefb73
Wee cleanup.
Ray Greenwell
2009-08-12 18:22:46 +00:00
5aefb2a2a6
Check that interdependent invokers are empty when shutting down the PresentsInvoker in addition to checking the PresentsInvoker itself and the DObjectMgr. Expose waiting for all theses queues to empty on PresentsInvoker such that Presents servers can do things like wait for all the queues to empty before opening to the public.
Charlie Groves
2009-08-12 08:29:49 +00:00
5d8a93632e
Allow callers to detect when a client is about to die even though it is connected and logged on. This allows handling of the edge case where a client is getting throttled.
Jamie Doornbos
2009-08-05 22:12:16 +00:00
4dc85174ba
Fix for NPE in closing datagram channel.
Andrzej Kapolka
2009-08-04 19:41:01 +00:00
5099dc9d96
Mark run() as an override since it is.
Mike Thomas
2009-08-04 18:23:55 +00:00
c3e31440ab
Comment some fiddliness.
Michael Bayne
2009-08-04 16:48:21 +00:00
9ce8a7ac71
Moved our peer-aware cron service into Presents from MSOY.
Michael Bayne
2009-08-01 23:01:40 +00:00
d0d5c2805e
StringUtil got moved to flash-utils.
Michael Bayne
2009-07-30 21:55:40 +00:00
c0499674e9
Allow an InvocationListener interface to extend another InvocationListener interface. The way this works is a little tricky, but it all works out.
Michael Bayne
2009-07-28 21:34:13 +00:00
5dc71a9b8e
Preserve varargs in generated marshallers and providers to keep Eclipse from getting in a snit about it
Charlie Groves
2009-07-24 23:39:35 +00:00
805ef99374
Various utility classes moved to flash-utils.
Michael Bayne
2009-07-21 20:56:35 +00:00
Andrzej Kapolka