leigh mcrae 14d494474c Cached the tracker type (ELLMTracker) in FLLMTracker, instead of passing it to all the member functions. ...

- The FLLMTracker is initialized with the ELLMTracker, so there is no reason not to cache it.
  - This change makes the code easier to read because there isn't a redundant ELLMTracker being passed around.
  - Also it's useful when debugging the code too see which tracker you are in.

#tests Built most Win64 targets. Ran editor. Did a local ReplayRun Did a Horde Minimal Code Only Build with most targets and LLM.
[REVIEW] [at]matt.peters [at]simon.orr [at]joe.barnes [at]alex.nischwitz [at]robert.millar [at]gwennael.arbona

[CL 26283163 by leigh mcrae in ue5-main branch]

2023-06-28 09:05:11 -04:00

..

Internal

Moved file cache into ioscod and eliminated base class .h/.cpp

2023-05-24 06:13:15 -04:00

Private

Cached the tracker type (ELLMTracker) in FLLMTracker, instead of passing it to all the member functions.

2023-06-28 09:05:11 -04:00

Public

a workaround to avoid "deadlock by busy-waiting" in ShaderCodeArchive: the existing IAsyncReadRequest::WaitCompletion() didn't guarantee completion, only that the request is done and the result is ready. It was used in IMemoryReadStream::EnsureReadNonBlocking() to ensure the following read will be non-blocking, and this can fail and actually block on reading -> busy-waiting -> deadlock. The new method EnsureCompletion() was added instead of "fixing" existing WaitCompletion() because other use cases don't need completion guarantee and the new method is theoretically slower

2023-06-28 07:23:02 -04:00

Tests

Added PREPROCESSOR_VA_ARG_COUNT(...) and PREPROCESSOR_APPEND_VA_ARG_COUNT(Prefix, ...) macros, which evaluate to the number of arguments passed to it and a token made by appending a prefix to the count respectively.

2023-06-23 14:59:15 -04:00

Core.Build.cs

[Backout] - CL26223564

2023-06-24 17:52:48 -04:00