Clemens Brunner

I'd try to avoid the copy by always computing the `:full` solution, and if `mode === :same` or `mode === :valid` I'd return just the portion of the array as...

Additional comment: I would prefer to use strings for the three possible parameter values, i.e. `"full"`, `"same"`, and `"valid"`. AFAIK symbols seem to be a bit misused for this purpose...

What performance difference do you expect for comparing an argument once? According to what I read (especially the [SO answer](https://stackoverflow.com/a/23482257/1112283) by @StefanKarpinski), using symbols as arguments only make sense when...

> A very prominent example from Base is the head field of Expr, but most users rarely get in touch with it. This is what symbols were made for -...

> First, I noticed that `xcorr` using `Symbol`s, so I thank `conv` should also use `Symbol`s. You're right, I didn't see that. Then yes, to be consistent both functions should...

Out of curiosity, what backend is flexiblas using? I think the default is OpenBLAS, could you run the tests with the Netlib backend?

> With the Netlib backend (`export FLEXIBLAS=NETLIB`), the two failures reported here on `ppc64le` and the one failure reported in #10984 on `aarch64` disappear Cool, so those two architectures now...

> Correct, on Netlib BLAS. Does this project target only Netlib BLAS, or should these tests pass on other BLASes? Or do you think these are actual bugs in OpenBLAS?...

> Yes 18 digits is fine for that test, feel free to open a PR OK! You could set up a [self-hosted runner](https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#architectures), they support Linux ARM64. Ideally, you could...

Looks good to me!