Martin Pool comments

Results 246 comments of


                                            Martin Pool

Run all workspace tests

Hey, thanks for pointing out this use case in jj (which is a project I was excited to read about, too.) I think something like `--test-all-packages` or `--test-package=A,B,C` would make...

Run all workspace tests

cc @panarch

Build libusb and librtlsdr - use submodules or a curl+makefile

For what it's worth it does seem to build on macOS using the Homebrew versions of libusb and rtlsdr: #44 For Windows and cross-platform support another option might be vcpkg...

Nextest profile to use

You should be able to say `--cargo-arg=--profile=mutants`

v25.0.1 match arms and guards are mutated in excluded functions

The immediate cause here is that the `exclude_re` matches against the mutant name, and these mutants don't include the impl/fn name. So I think there are two possibly complementary fixes:...

v25.0.1 match arms and guards are mutated in excluded functions

Great, thanks for letting me know.

False positives replacing equivalent expression, like `Duration::ZERO` and `Default::default()`

Adding a table of heuristics that are known to be equivalent seems like the most immediately tractable thing. I don't know about running the code as a way to detect...

False positives replacing equivalent expression, like `Duration::ZERO` and `Default::default()`

That would be a step beyond adding more heuristic name-based assumptions that, say, `new` typically means the same as `default()`, etc. I think it would be important to be clear...

False positives replacing equivalent expression, like `Duration::ZERO` and `Default::default()`

If you did not see it you might read about my overall thinking on this balance.

False positives replacing equivalent expression, like `Duration::ZERO` and `Default::default()`

> My experience is that there‘s a lot of small functions that don’t feel worth writing tests for, because there is a sense in which the most reasonable test would...