José Fonseca

@gerddie , FYI, this is the issue I mentioned on https://github.com/apitrace/apitrace/pull/700#issuecomment-773411284

I think that adding an `--dump-short` option to glretrace, to tell it not to dump huge/ sounds good for starters. The 2nd time requests a state dump, the option should...

traces can be quite big, so I'm open to the idea of supporting more compression algorithms. But I dread having to support a battery of them. It seems every couple...

The issue with the benchmark from https://facebook.github.io/zstd/ is that it was carried using general purposes corpuses, as opposed to actual apitrace traces. It's possible that Zstandard does wonderfully in the...

I tried `zstd --ultra -22` and it makes a huge difference: 936M. The decompression speed is way faster too: brotli decompresses at 342MiB/s where as ztd decompresses at 1.10GiB/s. So...

By _pmem_ you mean persistent memory? I wasn't aware it could interfere with tracing. I don't have such a device. Any chance you could `git bisect` the regression? My guess...

Sorry, I realize it's been a long while, but I'll need a sample trace to investigate this further.

> I am testing angle applications, but exclusively on UWP Do you know how different is UWP from Metro? Because I already managed to get apitrace to trace Metro apps:...

@chrisdjali-wrld3d , as long as Angle support on Windows doesn't break existing stuff (egltrace on Linux and glretrace on Windows) I'm happy to accept a pull request. Regarding UWP support,...

Indeed `--singlethread` just blindly replays all calls in one thread (the main thread.) It does not attempt to do anything smarter than that. This option is just a hack, which...