diff options
| author | Monty Brandenberg <monty@lindenlab.com> | 2014-09-05 17:37:21 -0400 |
|---|---|---|
| committer | Monty Brandenberg <monty@lindenlab.com> | 2014-09-05 17:37:21 -0400 |
| commit | ee4cd594623eff6f04cfc2bafb937f0cfd631080 (patch) | |
| tree | 7d3a8770fca691abc66a910ea64dc79a28ef46f6 /indra/newview/llmeshrepository.cpp | |
| parent | bbf9de9c6717f38a77a39d42d8493d275d558db9 (diff) | |
| parent | 0c20beda6800149ee71a307ca4e943b5bba56908 (diff) | |
Merge. Refresh from viewer-drano-http-4 after pipelining work.
Diffstat (limited to 'indra/newview/llmeshrepository.cpp')
| -rwxr-xr-x | indra/newview/llmeshrepository.cpp | 13 |
1 files changed, 12 insertions, 1 deletions
diff --git a/indra/newview/llmeshrepository.cpp b/indra/newview/llmeshrepository.cpp index 6477389d4c..a6707392fe 100755 --- a/indra/newview/llmeshrepository.cpp +++ b/indra/newview/llmeshrepository.cpp @@ -2644,6 +2644,17 @@ void LLMeshRepository::cacheOutgoingMesh(LLMeshUploadData& data, LLSD& header) } +// Handle failed or successful requests for mesh assets. +// +// Support for 200 responses was added for several reasons. One, +// a service or cache can ignore range headers and give us a +// 200 with full asset should it elect to. We also support +// a debug flag which disables range requests for those very +// few users that have some sort of problem with their networking +// services. But the 200 response handling is suboptimal: rather +// than cache the whole asset, we just extract the part that would +// have been sent in a 206 and process that. Inefficient but these +// are cases far off the norm. void LLMeshHandlerBase::onCompleted(LLCore::HttpHandle handle, LLCore::HttpResponse * response) { mProcessed = true; @@ -2685,7 +2696,7 @@ void LLMeshHandlerBase::onCompleted(LLCore::HttpHandle handle, LLCore::HttpRespo if (par_status == status) { - // 216 case + // 206 case response->getRange(&offset, &length, &full_length); if (! offset && ! length) { |