I read measuring-test-coverage-of-rust-programs and became interesting is any reasons to not implement workaround described in article as
cargo test --with-coverage
carog run --with-coverage
?
Or may be cargo team wait fix of this ?
I read measuring-test-coverage-of-rust-programs and became interesting is any reasons to not implement workaround described in article as
cargo test --with-coverage
carog run --with-coverage
?
Or may be cargo team wait fix of this ?
Just wanted to add that this would be a really, really nice feature to have out of the box (or even just in a single crate install).
In Python, this is really easy to get going with pytest-cov or even just the coverage package directly. Letting a new user write a test and run something as simple as cargo test --with-coverage
would make coverage very accessible.
I agree, it would be really nice if this were a first-class feature of rustc and cargo.
Relatedly, my blog post was up for one day before being broken by 1.19
[Now it seems impossible to get accurate coverage of unit tests. -Zno-landing-pads
was the previous method, but -Z
isn’t supported any more. The stable replacement -Cpanic=abort
isn’t supported for unit tests.]
From one hand this is bad, but from other hand coverage task may run once per day/half day on nightly compiler, just to measure coverage, I doubt that numbers would be different.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.