diff options
| author | Monty Brandenberg <monty@lindenlab.com> | 2013-09-12 20:15:20 -0400 |
|---|---|---|
| committer | Monty Brandenberg <monty@lindenlab.com> | 2013-09-12 20:15:20 -0400 |
| commit | 99c60b83f15a3c40ecf0e5f144cfc060b37c4cf8 (patch) | |
| tree | 71fd7c3bb326d73255d5a63a3ad1d0d9cc9471e0 | |
| parent | 622eae65551df9a4ca6843a6a657777ff5e2140e (diff) | |
SH-4489 New debug/dev settings for control over new mesh behavior
While giving myself a full project code review, found a bug in
the MeshUseGetMesh1 setting. Mostly defaulted to old configuration
but used the GetMesh2 caps which would have been a huge DoS
resource sink. Did some documentation maintenance as well while
I was in there. More for the to-do list, etc.
| -rwxr-xr-x | indra/newview/app_settings/settings.xml | 4 | ||||
| -rwxr-xr-x | indra/newview/llmeshrepository.cpp | 17 |
2 files changed, 13 insertions, 8 deletions
diff --git a/indra/newview/app_settings/settings.xml b/indra/newview/app_settings/settings.xml index 9948f9974a..60fc2d349d 100755 --- a/indra/newview/app_settings/settings.xml +++ b/indra/newview/app_settings/settings.xml @@ -9983,7 +9983,7 @@ <key>MeshUseHttpRetryAfter</key> <map> <key>Comment</key> - <string>If TRUE, use Retry-After response headers when rescheduling a mesh request that fails with an HTTP 503 status.</string> + <string>If TRUE, use Retry-After response headers when rescheduling a mesh request that fails with an HTTP 503 status. Static.</string> <key>Persist</key> <integer>1</integer> <key>Type</key> @@ -9994,7 +9994,7 @@ <key>MeshUseGetMesh1</key> <map> <key>Comment</key> - <string>If TRUE, use the legacy GetMesh capability for mesh download requests.</string> + <string>If TRUE, use the legacy GetMesh capability for mesh download requests. Semi-dynamic (read at region crossings).</string> <key>Persist</key> <integer>1</integer> <key>Type</key> diff --git a/indra/newview/llmeshrepository.cpp b/indra/newview/llmeshrepository.cpp index 353e7e9a7f..c3d149db9a 100755 --- a/indra/newview/llmeshrepository.cpp +++ b/indra/newview/llmeshrepository.cpp @@ -282,10 +282,15 @@ // dialog. Get the structured information going into the log into a // tree there. // * Header parse failures come without much explanation. Elaborate. -// * Need a final failure state for requests that are retried and just won't -// complete. We can fail a LOD request, others we don't. - - +// * Work queue for uploads? Any need for this or is the current scheme good +// enough? +// * Various temp buffers used in VFS I/O might be allocated once or even +// statically. Look for some wins here. +// * Move data structures holding mesh data used by main thread into main- +// thread-only access so that no locking is needed. May require duplication +// of some data so that worker thread has a minimal data set to guide +// operations. +// // -------------------------------------------------------------------------- // Development/Debug/QA Tools // @@ -1052,7 +1057,7 @@ std::string LLMeshRepoThread::constructUrl(LLUUID mesh_id) if (gAgent.getRegion()) { - if (! gMeshRepo.mGetMesh2Capability.empty()) + if (! gMeshRepo.mGetMesh2Capability.empty() && gMeshRepo.mGetMeshVersion > 1) { http_url = gMeshRepo.mGetMesh2Capability; } @@ -2181,7 +2186,7 @@ void LLMeshUploadThread::doWholeModelUpload() LLCore::BufferArray * ba = new LLCore::BufferArray; LLCore::BufferArrayStream bas(ba); LLSDSerialize::toXML(body, bas); - // LLSDSerialize::toXML(mModelData, bas); // <- This will generate a convenient upload error + // LLSDSerialize::toXML(mModelData, bas); // <- Enabling this will generate a convenient upload error LLCore::HttpHandle handle = mHttpRequest->requestPost(mHttpPolicyClass, mHttpPriority, mWholeModelUploadURL, |