aboutsummaryrefslogtreecommitdiffstatshomepage
Commit message (Collapse)AuthorAgeFilesLines
...
* need to delay even more agent close after teleport to compensate for slow ↵UbitUmarov2017-05-281-3/+4
| | | | machines and user connections. This one of current tp teleport protocol flaws
* a null ref check left behind on previus commitsUbitUmarov2017-05-281-1/+2
|
* make use of a rare thing called StringBuilder on LSL_List. LSL_List uses ↵UbitUmarov2017-05-281-30/+39
| | | | may need a revision to make sure they are passed by ref and not by value, with necessary adjustments. This does not have much impact on AppDomains, since if they cross, they are always serialized. Since lists are important parts of LSL, the AppDomainLoading option needs to be replaced by something else
* make LSL_list a CLASS again. Now we need to it allover again. Scripts need ↵UbitUmarov2017-05-281-1/+1
| | | | to be recompiled ( delete contents of bin/ScriptEngines)
* minor cleanupUbitUmarov2017-05-271-0/+1
|
* minor cleanupUbitUmarov2017-05-273-86/+56
|
* lose a refUbitUmarov2017-05-261-5/+12
|
* cleanup (grid)region info endpoint; add log to try to find some xml decode ↵UbitUmarov2017-05-264-88/+75
| | | | issues
* mantis 8181: don't try to delete contents of a non exitent folderUbitUmarov2017-05-261-4/+13
|
* change opensim-ode.sh to use the ode engine defined on opensim.iniUbitUmarov2017-05-262-2/+6
|
* avoid some wrong watchdog timeoutsUbitUmarov2017-05-262-3/+2
|
* make BlockingQueue.Dequeue timeouts more coerent (just less than watchdog ↵UbitUmarov2017-05-264-10/+9
| | | | timeout)
* missing file change.. actually use watchdog threads on assetsconnector andUbitUmarov2017-05-262-5/+13
| | | | avoid null refs on JobEngine
* change wrong watchdog stopped thread check code, and don't log itUbitUmarov2017-05-262-5/+10
|
* add some checks for valid endpointsUbitUmarov2017-05-252-59/+33
|
* move assetsconnector threads to whatchdog; Abort all alive threads known to ↵UbitUmarov2017-05-256-2/+30
| | | | watchdog before exit
* Update macOS version of libode to 0.13.2. The library is build as a fat ↵Geir Nøklebye2017-05-252-0/+0
| | | | | | library with both 32 and 64 bit code. O3 optimization and built with macOS 10.11 SDK. Signed-off-by: UbitUmarov <ajlduarte@sapo.pt>
* use threads from main pool on assetsconnectorUbitUmarov2017-05-252-4/+1
|
* minor changesUbitUmarov2017-05-253-6/+13
|
* another oneUbitUmarov2017-05-252-1/+9
|
* make a few more threads background onesUbitUmarov2017-05-252-3/+3
|
* make some threads background onesUbitUmarov2017-05-253-2/+3
|
* remove use of libomv BlockingQueueUbitUmarov2017-05-253-7/+7
|
* add a bit more pre exit cleanupUbitUmarov2017-05-244-4/+23
|
* mantis 8780 replace ode bins for linux with ones compiled against older ↵UbitUmarov2017-05-242-0/+0
| | | | Glibc, Thanks Jak
* give up on the code to stop AsyncCommandManager. It fails with many scripts ↵UbitUmarov2017-05-242-2/+4
| | | | running, and its only done deep on exit code. It is a backgroud thread should not block exit and never did
* dont log timeouts on stopped threadsUbitUmarov2017-05-231-1/+2
|
* fix lock position removing a not necessary functionUbitUmarov2017-05-231-10/+1
|
* allow new thread creation if it was aborted on AsyncCommandManagerUbitUmarov2017-05-231-3/+6
|
* protect AsyncCommandManager thread from fails on the several things it checksUbitUmarov2017-05-231-6/+7
|
* only remove AsyncCommandManager thread when last instance is deletedUbitUmarov2017-05-231-4/+11
|
* dont fail saving a OAR because of a broken materialUbitUmarov2017-05-231-1/+10
|
* stop a few more threads on exitUbitUmarov2017-05-233-17/+21
|
* please DIE! PLEASE??UbitUmarov2017-05-237-7/+40
|
* no all mono versions know DnsRefreshTimeoutUbitUmarov2017-05-232-2/+5
|
* remember mono about the default for DnsRefreshTimeoutUbitUmarov2017-05-232-0/+4
|
* check XFF headers of caller, but be quiet about themUbitUmarov2017-05-223-8/+8
|
* receivers can't tell if there where proxies on path unless they lookUbitUmarov2017-05-222-22/+20
|
* remove some code for NAT that was disabled by its author Adam Frisby since ↵UbitUmarov2017-05-221-18/+1
| | | | 2009, and i don't see how to mk work with current viewers at that point at least
* fix IClientIPEndpoint broken by justin long ago.. but stop using itUbitUmarov2017-05-224-20/+15
| | | | | except on SceneBanner, later it my be also removed from there and everywhere
* PGSQL: Another missing cast in XInventoryDataGeir Nøklebye2017-05-221-1/+1
| | | | Signed-off-by: UbitUmarov <ajlduarte@sapo.pt>
* make sure we drop requests if DoHTTPGruntWork failsUbitUmarov2017-05-221-4/+9
|
* disable keepalive on llHttpRequestUbitUmarov2017-05-221-0/+1
|
* cross avatar group info also if title is empty stringUbitUmarov2017-05-211-1/+4
|
* remove LongPoll type and queues. Events should now have reduced latency ↵UbitUmarov2017-05-212-35/+3
| | | | also. About previus commit, it maybe a modified ingen-e3s-v1.33 script that is broken, and not the original version, can't tell
* add a bit more protection for broken scripts like ingen-e3s-v1.33UbitUmarov2017-05-211-0/+4
|
* on recover form nom finite avatar pos, don't add it physics it it had noneUbitUmarov2017-05-211-2/+5
|
* fix a null ref in ubOde characterUbitUmarov2017-05-201-36/+34
|
* fix a cast that makes LitJson unhappyUbitUmarov2017-05-201-1/+1
|
* mantis 8172: fix llList2Json() in case of vector or rotation typesUbitUmarov2017-05-201-4/+18
|