Gregor Milos (Grzegorz Miłoś) comments

Results 27 comments of


                                            Gregor Milos (Grzegorz Miłoś)

SwiftPM package always produces warnings when `import libzstd`

@terrelln the cause are the macros defined as `config_macros` in `module.modulemap` here: https://github.com/facebook/zstd/blob/97291fc5020a8994019ab76cf0cda83a9824374c/lib/module.modulemap#L4-L24. They aren't really config macros, as they are defined here: https://clang.llvm.org/docs/Modules.html#configuration-macros-declaration (config macros need to be "binary"...

Performance tests broken (always paste a result from 3rd Jan, actual run fails)

I just noticed that my manual tests occasionally failed. I reproduced it by running it in tight loop and eventually tracked it down to an apparent race in `TCPThroughtputBenchmark`: `self.context.writeAndFlush`...

Performance tests broken (always paste a result from 3rd Jan, actual run fails)

Agreed re the harness being immune to failures in the test binary. For reference, I filed the issue about the specific test here: https://github.com/apple/swift-nio/issues/2649 //cc @FranzBusch

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

The test failure is ... because the # of allocations for less than expected. By about 1k. That's a good thing. ``` 17:35:51 ++ assert_greater_than 155002 156050 17:35:51 ++ [[...

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

@weissi this is the change we discussed in person.

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

@swift-server-bot test this please

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

@weissi how much signal is in the performance tests? Looks like there are a few that did regress, including: ``` lock_8_threads_10M_ops | 0.957658591 | 0.873560162 | previous | 9% ```

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

And another: ``` future_whenallcomplete_100k_deferred_on_loop | 0.084619949 | 0.08085791 | previous | 4% ``` There seems to be a pattern where the tests with large queues degraded in performance. I'll take...

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

@swift-nio-bot perf test please

Track `execute()` and `enqueue()` tasks separately from scheduled tasks.

I reviewed the test responsible for: ``` lock_8_threads_10M_ops | 0.957658591 | 0.873560162 | previous | 9% ``` This doesn't exercise my code change in any meaningful way. So the degradation...