various fixes to lltrace
start() on started recording no longer resets
fixed various instances of unit forgetfullness in lltrace
recording split now has gapless timing
scene monitor now guarantees min sample time
renamed a bunch of stats
added names to debug thread view on windows
removed LLThreadLocalSingleton
collapsed all thread recorder classes to single type, LLTrace::ThreadRecorder
moved fasttimer stack head to llthreadlocalsingletonpointer via ThreadRecorder
fixed copy construction behavior of Recordings to not zero out data
split measurement into event and sample, with sample representing
a continuous function
created separate constructor for static allocation of sampler buffer
fixed start/stop/resume semantics of samplers and added sampler time interval tracking
Two fairly simple conflicts: dead stats sending code in the
texture fetch code (new llcorehttp library) and the cleanup
code in llappviewer was moved around in 3.4.x.
Non-LLThread-based threads need to do some registration or LLMutex locks taken out in these
threads will not work as expected (SH-3154). We'll get a better solution later, this fixes
some things for now. Tracing of operations now supported. Global and per-request (via
HttpOptions) tracing levels of [0..3]. The 2 and 3 levels use libcurl's VERBOSE mode
combined with CURLOPT_DEBUGFUNCTION to stream high levels of detail into the log. *Very*
laggy but useful. Simple GET request supported (no Range: header). Really just a
degenrate case of a ranged get but supplied an API anyway. Global option to use the
LLProxy interface to setup CURL handles for either socks5 or http proxy usage. This
isn't really the most encapsulated way to do this but a better solution will have to
come later. The wantHeaders and tracing options are now supported in HttpOptions giving
per-request controls. Big refactoring of the HTTP resource waiter in lltexturefetch.
What I was doing before wasn't correct. Instead, I'm implementing the resource wait
after the Semaphore model (though not using system semaphores). So instead of having
a sequence like: SEND_HTTP_REQ -> WAIT_HTTP_RESOURCE -> SEND_HTTP_REQ, we now
do WAIT_HTTP_RESOURCE -> WAIT_HTTP_RESOURCE2 (actual wait) -> SEND_HTTP_REQ. Works
well but the prioritized filling of the corehttp library needs some performance
work later.
this is a combined mergeback of the following branches as per QAR-841:
maint-server/maint-server-1 (absorbed by maint-server-2)
maint-server/maint-server-2
maint-server/maint-server-3
havok4/havok4-8
havok4/havok4-9
yes dataserver-is-deprecated
Merge patches from 1.19.1 Server branch:
* QAR-293 Fix for hardcoded TTLs in web dataservices
* DEV-10826 fix for busted binary and notation parse if handed an unlimited parse size
* Bounce apache2 processes before starting backbone/dataserver/simulator
* Changing web-ds TTL in a way that any query that has 0 < ttl <=60 will have ttl = 61.
* Partial reversion of multiagent-chat to 1.19.0 to address fast memory leak
* Fixed minor, non user facing bug in multiagentchat
* set-classified-stats: Rewrote to use new MDB2 query reformatting syntax
* Fixed possible bad conversion of vivox data
* DEV-550, caching changes to DirClassifieds Query
* QAR-240 (DEV-8488) Prevent residents from purging stuff that isn't trash on the backend
* More mem leak fixes for multiagent-chat
* QAR-274 Fetch inventory descendents over TCP (via HTTP cap) instead of UDP
* DEV-10151: Sometimes group IMs appear to be person to person IMs
* QAR-321 Changes to crash_reporter
* DEV-11004 Speed up people search query using FORCE INDEX (PRIMARY) on the username table if the first-name query fragment is >= 3 chars
* DEV-11004 Speed up people search query using FORCE INDEX (PRIMARY). Web service version of this, must use two named queries because we need to change the query based on input string length.
Richard Linden