aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/OpenSim/Framework/Communications/Cache/AssetCache.cs (follow)
Commit message (Collapse)AuthorAgeFilesLines
* * Optimized usingslbsa712007-12-271-25/+24
| | | | | | | * shortened references * Removed redundant 'this' * Normalized EOF
* Make copying of scripts into prim inventories more reliable on the first ↵Justin Clarke Casey2007-12-221-3/+45
| | | | | | | | | | attempt when the asset server is lagging by formalising the de facto polling. This may not be the best solution in the long run, but should improve things for now. This may also improve reliability when updating inventory item metadata (e.g. renaming an item) and in retrieving textures for the main map view.
* minor refactorJustin Clarke Casey2007-12-221-3/+8
|
* Misc. cleanup:Jeff Ames2007-12-191-46/+14
| | | | | | | * added Util.Clip(value, min, max) * modified asset cache's numPackets calculation to use max packet size (600) instead of 1000 * removed a few magic numbers
* Thank you Justin for:Charles Krinke2007-12-141-52/+27
| | | | | 0000188: Clean up of some AssetCache.ProcessAssetQueue() code
* patch from justincc for Mantis #163 - Opening a script in inventory longer ↵Teravus Ovares2007-12-141-11/+36
| | | | than 1500characters can cause looping crash.
* more work on texture downloading.MW2007-12-101-265/+38
| | | | | | | Refractored the TextureDownloadModule (but currently to make debugging easier, it is running as a non shared module, so this results in a instance of this module being created for each region (and a extra thread per region), this will be changed back soon. Removed the old texture handling/sending code from AssetCache. A few other small changes/fixes.
* saved OpenSim source code from the giant rampaging unterminated copyright ↵Jeff Ames2007-12-101-2/+2
| | | | notice of doom
* reversing the r2599 patch, as this started causing CreateThread errors for Sean Dague2007-12-061-6/+1
| | | | | | | | | me on Mono 1.2.4, which led to client crashes. I think the Timer.Stop() wasn't doing what was desired on Mono. The Queue refactoring should address the readability issues lbsa71 was working on as soon as I get the merge together.
* * now the throttle timer is stopped whilst processing Queue so that it won't ↵lbsa712007-12-061-1/+6
| | | | | | | | fire twice * started to refactored throttling method * some code convention refactorings
* * Refactored NumPackets calculationlbsa712007-11-231-20/+21
|
* * Refactored IClientAPI.OutPacket to require a second mandatory parameter. ↵Teravus Ovares2007-11-181-8/+8
| | | | This parameter has an enum:int ThrottleOutPacketType and contains types; Resend, Land, Wind, Cloud, Task, Texture, and Asset.
* first stab at implementation of CopyInventoryItemJeff Ames2007-11-181-0/+11
|
* Fixed bug in assetcache where certain textures were not being packetized ↵Brian McBee2007-11-111-1/+1
| | | | correctly
* * Added better logging to AssetCachelbsa712007-11-061-42/+66
| | | | | | | * AssetCache now ignores duplicate uploads * some m_ refactoring * ignored some bins
* * Diuerse beavtificatemslbsa712007-11-011-2/+2
|
* renamed FetchAsset to RequestAsset on IAssetServer, as think its a more ↵MW2007-11-011-4/+4
| | | | fitting name. As the call shouldn't actually return the data, but just add a request for the data to be sent back via the callback.
* made illogical bitwise operations logicalJeff Ames2007-10-311-1/+1
|
* * Optimized usingslbsa712007-10-301-107/+100
| | | | | | * Shortened type references * Removed redundant 'this' qualifier
* as per the "Filesystem cleanup for OpenSim repository" mailing list thread. ↵MW2007-10-291-2/+1
| | | | | | | Have flattened the OpenSim.Framework project/namespace. The problem is that the namespace is still wrong as its "OpenSim.Framework" while the directory is "OpenSim\Framework\General" , so we need to decide if we change the directory or correct the namespace. Note this has lead to a big flat project, but I think a lot of the files we most likely don't even use any longer. And others belong in other projects/namespaces anyway.
* * Removed AssetNotFound as it crashes clientlbsa712007-10-291-13/+13
| | | | | * Fixed introduced bugs in ScenePresence
* Started the process of cleaning up AssetCache and moving most of the code ↵MW2007-10-291-77/+91
| | | | into modules. Have moved TextureRequest handling (from the client) to a module. But even though to start with I just did a little bit of cleaning up of the existing code, it doesn't seem to work as good as the old code so I need to spend more time on it. So for now am committing my changes but with them not in use. So for now all Texture and asset requests are still handled by the old code in AssetCache.
* Region ground texture was not marked as temporary, when being created, and ↵Tleiades Hax2007-10-261-7/+9
| | | | assetcache did not check if an asset was temporary prior to uploading the cache to the asset server. Consequently the asset server was constantly being swamped by temporary ground textures.
* Very early first implementation of grid based assets.Tleiades Hax2007-10-261-4/+7
| | | | | Run this on a major grid, and weep
* Part of supporting gridwide assetss. A little premature release, but I'm ↵Tleiades Hax2007-10-241-23/+1
| | | | tired of conflicts.
* * Return of R2162. /Take that SVN!/Adam Frisby2007-10-221-2/+2
|
* revert r2162 as it completely clobbered all the work onSean Dague2007-10-221-2/+2
| | | | | | | the ChatModule by MW and myself. Couldn't find Adam online after that rev went in.
* * Major ass commit.Adam Frisby2007-10-221-2/+2
| | | | | | | * Sqlite Storage Engine now supports terrain -- however be aware that every terrain revision stored will at 512KB to your database file. At the moment it is storing every revision from the first. * Fixed an issue where by noverbose mode would display lots of useless junk. Noverbose mode is now quite usable. * Fixed a whole bunch of console message issues such as naming and categorisation
* * Total refactoring of Asset Server for massive winlbsa712007-10-191-2/+7
| | | | | | * There is now a AssetServerBase * lolcat in ur assets
* * Encapsulated all CommunicationsManager serviceslbsa712007-09-241-5/+4
|
* Hopefully fixed the build error that I introduced with my last commit.MW2007-09-151-1/+1
|
* Testing to see if this fixes the editing appearance crash, or makes it (or ↵MW2007-09-151-1/+9
| | | | anything else) worse.
* move the locks closer to the data, sqlite for assets now works like a champSean Dague2007-09-111-9/+3
|
* mass update of urls in source code to new websiteSean Dague2007-09-101-1/+1
|
* hooked up sdague new sqlite asset database provider to the old asset system. ↵MW2007-09-101-6/+6
| | | | | | | | So we can still use sqlite for assets while we wait for the rest of the new asset system to be wrote. Needs more testing, so if it causes problems will have to swap back to db4o.
* Some work on Module loading/management.MW2007-09-041-1/+6
| | | | | | | | Some more modules templates classes (hoping that someone will pick some of these and work on implementing them). Early version of the "Dynamic Texture Module", although currently there are no render modules included (so not really functional without them). Added osSetDynamicTextureURL script function, for attaching a dynamic texture to a prim. Some work on the console command handling. Added "change-region <regionname>" and "exit-region" so that after the use of change-region, the commands entered will apply to that region only. Then use exit-region to return to the top level (so commands then function as they did before and either apply to all regions or to the first region) (Note: this hasn't been tested very much)
* Added some place holder classes for various modules.MW2007-08-301-36/+39
| | | | | Some work on the asset cache, can people please test this. including on one of the public systems so we can see if it causes problems with multiple users.
* Deleted a few old files that are no longer used.MW2007-08-271-3/+3
| | | | | | Deleted the GridInterfaces projects, and for now moved the old local asset server into Framework.Communications, as we prepare to rewrite the asset cache and asset server. Deleted Framework.manager as I am sure this is no longer in use.
* Stopped EntityBase from implementing IScriptHost, as don't think it should, ↵MW2007-08-271-1/+1
| | | | | | | multiple objects are based on entitybase and they all don't want the baggage from IScriptHost. SceneObjectPart already implements it anyway. Added llGetOwner function, and tested the ll functions that I added in last commit.
* Another small tweak to image sending.MW2007-08-271-1/+2
| | | | | Implemented a few ll Functions, llSetObjectName llGetObjectName, llLoadURL (all currently untested).
* Tweaks to assetcache to make texture downloading better (I hope)Brian McBee2007-08-271-13/+14
|
* See if this fixes the 0x0 texture size problems, and speeds up texture ↵MW2007-08-261-2/+3
| | | | downloads in general a little bit.
* Another attempt to fix the image sending bug (next week, I intend to rewrite ↵MW2007-08-261-32/+69
| | | | | | | | the assetcache and asset server). Attempt to fix bug # 326. (crashing when using save-xml and hollow prims) Attempt to fix bug # 328 (limit of 50 items in a folder)
* Second attempt to commit : Test to see if this change improves or makes ↵MW2007-08-251-1/+1
| | | | worse the texture sending bug
* A little bit of cleaning up.MW2007-08-221-1/+1
|
* Taking Prims (SceneObjectGroups) in and out of inventory should now work and ↵MW2007-08-161-26/+43
| | | | if left in inventory will still be there after restarts. (as with the rest of inventory it will only fully work in standalone mode with account authentication turned on).
* Making sure my local working copy is in sync with svn before I start the job ↵MW2007-08-091-3/+5
| | | | of enabling the new SceneObject classes.
* Fixed the asset download problem I think, solution might not be ideal but ↵MW2007-08-081-12/+9
| | | | works for now.
* Hopefully improved texture downloading (including the speed at which they ↵MW2007-08-081-35/+83
| | | | download), but still a few problems, it seems that the client has a quite short timeout for receiving a texture and if the whole texture isn't sent within this time, the client will request the texture again, With quite small textures this is fine, but it seems that with larger textures we can't send them fast enough and a infinite loop develops where the client keeps requesting a texture and we keep trying to send it, but are never fast enough. So I've for now put code in that so that the server will try to send a texture only once and then after that will ignore future requests from that client for that texture.
* Deleting objects should now work. But beware they aren't send to your trash ↵MW2007-07-291-0/+1
| | | | folder or anything so there is at the moment no way to recover deleted objects.
generated by cgit v1.1 at 2024-07-16 11:13:44 +0000