Age | Commit message (Collapse) | Author |
|
|
|
trivial cleanup
Move LLDiskCache::updateFileAccessTime to LLFilesystem as it's the only user of that function.
Change mCacheDir and LLDiskCache::metaDataToFilepath to statics.
|
|
and llunittype.h for now
|
|
* More deterministic vsize calculation. Add control for choosing downscale method.
* Quick hack to make GLTF preview work again
|
|
|
|
- this time based on Henri's suggestion in this discussion: https://bitbucket.org/lindenlab/viewer/commits/e28c1b46e9944f0215a13cab8ee7dded88d7fc90#comment-10537114
|
|
file in the cache. I am still unable to reproduce this problem, although others have so I am trying a more aggressive approach for them to test. I tried this as part of last speculative fix but saw what I thought were scary side effects in the log/console. Turns out, those are there in the release viewer so probably unrelated to this change.
|
|
cache - since I am unable to repro, this might be enough
|
|
system issues - taken from https://vcs.firestormviewer.org/phoenix-firestorm/changeset/104a8600946be01e2de44d10ad069ba854272d1f
|
|
results in a version of the DRTVWR-519 that matches what was presemt before it was deployed as a release viewer *plus* 3 small fixes from Maxim (See commits). This branch can now be used for additional fixes before eventually being used to release D-519 as normal
|
|
DRTVWR-519"
This reverts commit e61f485a04dc8c8ac6bcf6a24848359092884d14, reversing
changes made to 00c47d079f7e958e473ed4083a7f7691fa02dcd5.
|
|
|
|
renaming files if destination file exists
|
|
things about between llfilesystem and lldiskcache - there is still some bookkeeping work left but this is the first version that appears to work and actively manage the cache
|
|
single read and a single write function
|
|
functions in llfilesystem
|
|
new lldiskcache implementation
|
|
since it's responsible for performing file system operations and (will eventually) delegrate to a separate disk cache component to save/load data and keep track of metadata etc.
|