| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Metrics can only be collected for a single context, with the current
implementation of INTEL_performance_query. Stop/start metrics before
and after context switches, and collect metrics from each context before
publishing.
Reorganize renders into contexts. Contexts contain renders for a GL
context, and manage state changes that must be intstrumented as contexts
switch.
Eliminate awkward checking of current context when retracing metrics.
Subsequent features (state tracking/uniforms) will be per-context.
|
|
|
|
I'm not aware of people using it, so I believe it's not worth the hassle
as I'm trying to cut maintenance burden down.
|
|
As suggested in
https://github.com/apitrace/apitrace/issues/479#issuecomment-239666769
|
|
Don't explicitly mention exceptions.
|
|
|
|
Renaming the library to something else is left to another change, as
there's already a trace lib in the wrappers dir.
|
|
|
|
|
|
|
|
|
|
|
