Commit Graph

  • 1b9c9dac62 Decreased the maximum number of sound spoolers to 10. What was I smoking when I set it to 24? Let's try it like this for a spell. Ray Greenwell 2003-07-28 05:32:04 +00:00
  • 5dc25a8b09 Print debug messages to STDERR. Michael Bayne 2003-07-28 04:07:52 +00:00
  • 7961683cb0 Added support for dumping a bundle metadata file directly in addition to extracting it from a bundle. Michael Bayne 2003-07-28 04:07:30 +00:00
  • 42b84f894f Made setActionSequence() more robust. Michael Bayne 2003-07-28 04:06:38 +00:00
  • fbd52c2be3 Tone down cache performance reporting; upped image cache size by a meg. Michael Bayne 2003-07-28 04:06:14 +00:00
  • f252c0f88d Added sanity check. Michael Bayne 2003-07-27 02:50:53 +00:00
  • a35edea96e We need also to clear the receivers when we transfer them from our auth user object to our chosen pirate user object. Michael Bayne 2003-07-25 20:51:08 +00:00
  • 82b616d724 Oh fucking hell. I left some debugging output in here. Michael Bayne 2003-07-24 23:51:50 +00:00
  • 991b68638e Yarrrrrrrrrrrrrrrrrrrrrrrrrrrgh! Michael Bayne 2003-07-24 23:50:32 +00:00
  • 0419aebbd5 Could they have possibly changed any other things about the javap output that would cause this class matching script to fail in unexpected ways? I'm sure to soon find out. Michael Bayne 2003-07-24 23:40:16 +00:00
  • e43f7c0de8 More god damned fiddling to make this work with 1.4.2's javap output. Michael Bayne 2003-07-24 23:19:26 +00:00
  • aced4d2fe9 More tweaks to deal with 1.4.2 javap output. Michael Bayne 2003-07-24 22:12:52 +00:00
  • 212d051080 Deal with the damned changes they made to javap output. Michael Bayne 2003-07-24 21:26:40 +00:00
  • 76af5d2353 Prevent propagation of lock release events. Michael Bayne 2003-07-22 22:54:14 +00:00
  • ef2d31db0b Canned vestigal population crap. Michael Bayne 2003-07-22 07:29:11 +00:00
  • 493c75b865 Make client and connection count information available. Michael Bayne 2003-07-22 02:50:49 +00:00
  • 245a5d5139 Do some sanity checking. Michael Bayne 2003-07-20 17:02:59 +00:00
  • f7d55b9814 More debugging for DSet duplicate business. Michael Bayne 2003-07-20 17:02:26 +00:00
  • 957faac149 Slap the CommunicationAuthorizer code in such that we block all user communication if the user is not authorized. Eric Lundberg 2003-07-18 01:58:38 +00:00
  • c2a5e8966b Initial checkin of ye ole' pluggable interface for authorizing player communication. Eric Lundberg 2003-07-18 01:54:51 +00:00
  • 7efc5dfc32 Make the auth response data available to interested parties. Michael Bayne 2003-07-17 21:39:15 +00:00
  • 829f3ad0eb Start transactions outside of the try-finally clause so that if there is a problem with starting a transaction, we don't get an exception for failing to commit the transaction. Ray Greenwell 2003-07-16 18:03:30 +00:00
  • 2726f6a29d Slight optimization to chat history pruning. Ray Greenwell 2003-07-14 23:46:03 +00:00
  • 96998e5678 Set _sourceLastMod to -1 when uninitialized instead of 0, so that we can distinguish between uninitialized and various file errors, and hopefully get us some better logging output. (If there are errors with a file, lastModified() returns 0, which would have caused resolveJarFile to think that the bundle just wasn't ready yet and not create the JarFile object, which resulted in a NPE, and not useful IOEs.) Ray Greenwell 2003-07-14 22:18:18 +00:00
  • 10ceff36d2 Log up a warning in these trying circumstances so that future bugs don't remain obscure. Michael Bayne 2003-07-13 02:59:27 +00:00
  • 030545b8eb Thanks to Mr. Lundberg's insight, we have tracked down a gross oversight in the low-level invocation services whereby after a disconnect, any InvocationReceiver registration would be wholly ignored due to the previous registrations being left in the receivers set and DSet refusing to add duplicate instances of the registration. This wasn't immediately apparent because the initial set of registrations always happens in the same order and thus are equally useful after the reconnect and any other registrations are removed during the normal course of affairs. But if a user disconnects *during* a puzzle, they will leave that puzzle's registration around and be unable to play that (or any) puzzle again until they log off and back on. Michael Bayne 2003-07-12 22:34:03 +00:00
  • 48c9556d18 Fine grained scene block resolution instrumentation. Michael Bayne 2003-07-12 04:16:32 +00:00
  • eec42e8f42 Report our fully qualified tile id more human readibly. Michael Bayne 2003-07-12 04:16:10 +00:00
  • 13ded21e51 Added summarizeState(). Michael Bayne 2003-07-12 03:59:54 +00:00
  • 55a40d7161 We need to know about unregistrations as well. Michael Bayne 2003-07-11 03:13:30 +00:00
  • 7047bcfaf5 Log when we register an invocation decoder because we're seeing circumstances where they aren't around when they should be. Michael Bayne 2003-07-11 03:02:46 +00:00
  • d37c56b077 Added more substantial funny business checks. Michael Bayne 2003-07-11 01:49:52 +00:00
  • e09ef8bbb2 If asked to find a marshaller for a pure interface, use the marshaller for Object. Ray Greenwell 2003-07-11 01:22:16 +00:00
  • cfe4b49c6e Key is Comparable. Ray Greenwell 2003-07-11 01:21:43 +00:00
  • 71cd8b421a - Use the newer binary search method of locating entries in get(). - Require a Comparable key in get() and containsKey(). Ray Greenwell 2003-07-11 01:21:23 +00:00
  • e85e6ad617 Catch our funny choking getOccupantInfo and log the failure prior to allowing things to get out of hand. Since getOccupantInfo may normally return null, doing so in our strange failure case will not cause further freakout. Michael Bayne 2003-07-11 00:47:33 +00:00
  • e140fd6301 Use the AttachableURLFactory. Ray Greenwell 2003-07-09 18:44:52 +00:00
  • 4c342c28f3 We need to potentially decode the length after every read. The original code did that due to eventuall getting out of the loop, the "fixed" code failed to do it if we failed to fill our buffer up on the first read. Michael Bayne 2003-07-09 05:13:34 +00:00
  • 0ebfeadd21 Seems to be some kind of issue with moving the length == -1 check inside the loop, so I'm rolling back to the previous version since I can't even log in to my server with this change. Mike if you want to take a look or tell me what jimmyjamming needs to be done to make this change work, that would be great. Eric Lundberg 2003-07-07 22:35:38 +00:00
  • 17b567ba18 There was a circumstance where the framed input stream could hang if the frame size was exactly the same size as the buffer and no future messages came down the pipe to wake us up again. This probably would have had to happen very early in the session as the buffer gets bigger and bigger to accomodate the largest frame seen during the entire session and asynchronous messages are coming in more frequently as well. Thus it's extremely unlikely that this ever actually hung. In any case, we now log that circumstance (and do the right thing) so we can see if the log message turns up in future bug reports. Michael Bayne 2003-07-06 18:39:55 +00:00
  • 5c34de0283 Cache tilesets loaded out of component bundles. Michael Bayne 2003-07-02 00:01:17 +00:00
  • 4862311e28 Not to NPE if we get back no resource bundles. This would be a catastrophic fuckola, but we can wait until later to freak out. Michael Bayne 2003-07-01 21:56:10 +00:00
  • 993c79e9d2 Added interface method documentation, rejiggered formatting. Michael Bayne 2003-07-01 21:53:02 +00:00
  • 00286b608e Added interface to all other packages to plug into the chat provider and provide an auto response to a tell. Eric Lundberg 2003-07-01 21:50:18 +00:00
  • 81460a9496 Added getGameConfig(). Michael Bayne 2003-06-29 23:11:23 +00:00
  • 78cf95ed49 Some minor reformatting. Michael Bayne 2003-06-26 19:57:24 +00:00
  • f26dc2ab9c Adapters! We loves the OOP. Michael Bayne 2003-06-26 18:24:31 +00:00
  • ec6cf00d6f Ask the place object if it needs to be broadcasted on before doing so. Ray Greenwell 2003-06-25 17:17:41 +00:00
  • 205a25ce10 Allow editing of floats and doubles. Michael Bayne 2003-06-25 03:13:56 +00:00
  • 8f0ec22267 Undid my last revision; the X/Y coordinates of a pathable are actually the hotspot, which does not need to be the upper left coordinate. So we let the sprite do the right thing and we just center on the hotspot. Ray Greenwell 2003-06-24 18:02:58 +00:00
  • faee2e4088 Whoops, we add new history messages to the end, so start pruning from the beginning. To do this more efficiently, we end up having to do some jockeying. Michael Bayne 2003-06-24 05:19:57 +00:00
  • 71109615bc Added setMirage(). Ray Greenwell 2003-06-24 02:25:10 +00:00
  • 0d79024b5c When centering on a pathable, center on its center rather than its upper-left coordinate. Ray Greenwell 2003-06-24 02:24:54 +00:00
  • b8b76e091c Preserve command events as is. Michael Bayne 2003-06-23 20:16:07 +00:00
  • 25ebfba64f For some strange reason, reusing Colorization instances causes character creation to break. Michael Bayne 2003-06-23 18:06:57 +00:00
  • 28a9441143 Allow the ImageManager to function if we're running in headless mode. It will create all images as 32-bit ARGB. Michael Bayne 2003-06-23 17:59:04 +00:00
  • 13fba9919e Make our AI tick frequency public. Michael Bayne 2003-06-21 15:32:29 +00:00
  • 2f63091f59 No show all. Michael Bayne 2003-06-20 21:50:56 +00:00
  • fc9f2d88f3 Fire a special scene object action when an object is clicked or object radial menu item is selected. Michael Bayne 2003-06-19 22:20:02 +00:00
  • 629c40f631 Increased the padding around the icon and between the icon and text. We're going to need to make this customizable by derived classes at some point. Michael Bayne 2003-06-19 22:12:51 +00:00
  • e7a7c2ba09 Pulled icon lookup into method that can be overridden. Michael Bayne 2003-06-19 22:03:14 +00:00
  • 3d8e053a7f Added the ability to add dynamic bundles which can be downloaded and added in while the game is running. Ray Greenwell 2003-06-19 01:30:27 +00:00
  • b4f2c4c526 Added isUnpacked(), which lets us know if the bundle is downloaded and ready to go. Ray Greenwell 2003-06-19 01:12:59 +00:00
  • 5398ce5656 Added support for adding a new bundle on the fly. Ray Greenwell 2003-06-19 01:12:32 +00:00
  • ab2a627f1a Report our time calculations always so that we can do a better job of debugging weird server delta calculations. Michael Bayne 2003-06-18 17:17:21 +00:00
  • 629a61d6b5 Report the error directly when we fail to create a jar file. Michael Bayne 2003-06-18 16:54:17 +00:00
  • cbe229ccb7 Revamped BundledTileSetRepository to make it easier to add bundles on the fly in the future; it also now provides efficient tileset name to tileset id mappings which is exported via the TileSetRepository interface. Lastly, TileManager now relies on TileSetRepository to efficiently provide tiles by name or id and no longer maintains its own cache (because the BundledTileSetRepository already maintains the same mappings). Michael Bayne 2003-06-18 05:48:45 +00:00
  • 572570de75 Added a static createBundle() to create a jar file from an already prepared TileSetBundle. Ray Greenwell 2003-06-18 00:34:47 +00:00
  • 3fb9a4272d Maximally compress jar files that we create. Ray Greenwell 2003-06-17 23:29:33 +00:00
  • 8e158dd1e8 Let's get the class name in there as well. Michael Bayne 2003-06-15 16:47:33 +00:00
  • aacca4ca02 Only log if we have a user message but no speak object. Michael Bayne 2003-06-14 06:23:41 +00:00
  • b9271233ac Har, lets be scanning down, not up. Eric Lundberg 2003-06-14 01:55:42 +00:00
  • 3701cef1d7 Only track UserMessage since we only care to log things other players said; prune any big history when a new message is noted. Michael Bayne 2003-06-14 01:18:27 +00:00
  • 1523629ef6 Provide the option to deliver listener body oids or usernames. Michael Bayne 2003-06-14 00:59:24 +00:00
  • f1ac86936e Be speaker aware. Big brother loves us. Michael Bayne 2003-06-14 00:55:40 +00:00
  • 61168c737c Don't double note UserMessage instead specifically note tells() which is the only circumstance where we're not getting sent our own message back to us. Michael Bayne 2003-06-14 00:55:19 +00:00
  • e6dc5c38d7 The basis for logging all recent chat on the server. Now I have to make every type of chat object known to man capable of reporting who's listening. Michael Bayne 2003-06-14 00:47:16 +00:00
  • 7d98a59c1c Enable unit profiling and keep an eye out for exceptionally long running units. Michael Bayne 2003-06-12 22:10:07 +00:00
  • 105aa7384d Added toArray(). Michael Bayne 2003-06-12 05:24:08 +00:00
  • 1c7fbbdccf Need to name this differently to avoid conflicts. Michael Bayne 2003-06-11 17:06:42 +00:00
  • 575337b429 Added support for dispatching scene updates made to the currently occupied scene. Michael Bayne 2003-06-11 04:14:11 +00:00
  • 087f26b41b Further work on dynamic scene updating. Michael Bayne 2003-06-11 02:48:07 +00:00
  • d9e38c1d35 Don't hork if there is a config for a music key, but it's empty. Ray Greenwell 2003-06-07 00:41:25 +00:00
  • 092299d23f Updated SpeakMarshaller import, which I missed the other day because I was just search .java files. Ray Greenwell 2003-06-06 23:58:53 +00:00
  • ef6fe1c6ee Whoops, left in a call to System.gc(). Michael Bayne 2003-06-06 00:20:41 +00:00
  • e6c3d54905 Disabled scene memory usage reporting as we get a good enough idea from periodic cache stats at the moment. Michael Bayne 2003-06-05 18:41:36 +00:00
  • d0ad3559cd Use SoftReference rather than WeakReference, the former apparently being subtly more appropriate for what we're doing. Michael Bayne 2003-06-05 17:32:03 +00:00
  • 52e5206ff4 Keep a single colorization record around for each ColorRecord rather than creating the same Colorization instance over and over each time we're asked for one. Michael Bayne 2003-06-05 17:31:21 +00:00
  • d1e2c532e4 Fixed typo with previous checkin. Must get sleep. Ray Greenwell 2003-06-05 00:33:43 +00:00
  • 7dd942b673 Start out the ping interval id at -1 so we can tell if it needs to be unregistered. Ray Greenwell 2003-06-05 00:29:45 +00:00
  • 6cb3d605d7 Be sure to unregister the interval, even if the the reader exits before we logoff. Ray Greenwell 2003-06-05 00:26:16 +00:00
  • 10f5536a6a Not to create a key every time we do a lookup. Only create a key when we know we're adding a new entry. Michael Bayne 2003-06-04 22:00:59 +00:00
  • 35f7f0806b Make sure that the transparent color is first in the colormap so that Java's png encoder doesn't get confused and merge black pixels into the transparency. Ray Greenwell 2003-06-04 05:56:58 +00:00
  • 5f515133c3 Chat revamp, phase 2. - FeedbackMessage has been subsumed into SystemMessage, which now has three modes: INFO, FEEDBACK, or ATTENTION. - Cleaned up methods for sending those system messages. - on the server: sendInfo(), sendFeedback(), sendAttention() - on the client: displayInfo(), displayFeedback(), displayAttention(). Ray Greenwell 2003-06-04 02:50:19 +00:00
  • 01b123b9f8 Added some utility methods for selecting a random element from an array or iteration of known size. Michael Bayne 2003-06-04 00:03:58 +00:00
  • 674523b365 Tell feedback message. Ray Greenwell 2003-06-03 23:46:10 +00:00
  • d3f3995cb7 Allow an optional ResultListener to be passed into requestTell that will tell us whether or not the tell succeeded. Ray Greenwell 2003-06-03 23:36:18 +00:00
  • 40afe3cb94 Chat revamp, phase 1. - Repackaged crowd/chat - All messages are delivered to the client via ChatMessage messages, including tells. Ray Greenwell 2003-06-03 21:41:33 +00:00
  • 368cd940b6 Make all of our object radial menus the same size; having them scale to the size of the object is not at all useful. Michael Bayne 2003-06-02 18:57:15 +00:00
  • ba73cc55df Deal as gracefully as possible if we don't have any data for the final field or fields of a streamed object. This allows us to add new fields to the end of a streamable without causing all existing serialized versions to become unreadable. Michael Bayne 2003-06-02 17:24:18 +00:00