gamma0987

I'll have a look into it but I actually tend to leave such things to the final user.

Sorry, but I don't think a custom global allocator should be part of iai-callgrind. So, I'm closing this issue.

What exactly do you mean with regression?

Sorry, that it took some time. In `v0.17.0`, I've added a command-line argument `--show-only-comparison` and an env var `GUNGRAUN_SHOW_ONLY_COMPARISON`. This option suppresses the usual self comparisons between the old and...

@tgross35 Thanks for your suggestion. I think both, `iter` and `file` have their advantages. The problem with an iterator is, that in contrast to `file` and `args` an iterator needs...

@tgross35 In v0.17.0 (which closed #326), I've added an `iter` argument for the benches attribute (`#[benches::some_id(iter = ...)]`). `iter` supports anything that implements `IntoIterator`. Let me know if this is...

Hi, thanks for your interest in iai-callgrind. There is no public interface of a `LibraryBenchmarkGroup`, `LibraryBenchmark`, ... because we expand the `#[library_benchmark]` into a new structure at compile time and...

You're welcome :) Yeah, the truncation of the description should be adjustable. It was never meant to identify a benchmark, but I guess that changed with the file parameter.

I don't know your setup, but I don't see the hard reason for having a customizable id beyond the customizability that already exists. Maybe you can explain your use case...

Ok, I thought you wanted to migrate your existing time based benchmarks. > I’m trying to use [Typst](https://github.com/typst/typst)’s existing integration test cases as cases for benchmarking Maybe you shouldn't do...