These always returned the same types and had the same implementation as the const overload
#jira
#preflight 644c24d1864f15ebf50883ab
[FYI] Tim.Smith
#rnx
[CL 25242982 by jamie dale in ue5-main branch]
This allows a class to change from using FBaseSparseData to FDerivedSparseData without having to fix-up all the existing accessors for GetBaseSparseData()
#jira
#preflight 644c1699877716c87826527a
#rb Tim.Smith
[CL 25241758 by jamie dale in ue5-main branch]
* All standard buffers now implement IComputeBuffer directly, and do not need to have ToSharedInstance() called to create a ref-counted version.
* Workers can now construct a socket directly, allowing multiple buffers to be attached.
* Control messages are sent whenever a receive buffer is attached, allowing the remote to wait for it to be available.
#preflight none
[CL 25169238 by Ben Marsh in ue5-main branch]
- Each artifact is named, and by default any files added to a tag with the name of the artifact will be treated as belonging to it.
- Artifacts can have include an optional set of keys that can be queried against on Horde.
- The host system (eg. Horde) is deemed responsible for archiving artifacts for later retrieval.
#preflight none
[CL 25048052 by Ben Marsh in ue5-main branch]
1) Ability to specify buffer size with FileNodeWriter methods CreateAsync and AppendAsync.
2) Ability to copy a FileNode tree directly into a stream without creating the nodes themselves.
3) When reading data, avoid creating byte[]s for the cache if the cache is not enabled.
4) Reduced default read size from 15MB to 1MB when caching isn't enabled
5) New ReadNodeAsync method that invokes a callback with the data instead of creating the node instance.
6) TreeWriter uses the current packet's full memory buffer size to detect the need to resize indead of the max packat size.
7) New bundle test for the streaming of FileNodes directly to a stream.
#rb ben.marsh
#preflight 64394d6d8901e5ef9811680f
[CL 25039272 by Tim Smith in ue5-main branch]
