Peter Hill

Ok, I can make `XZHermiteSpline` a base class of `XZPetscHermiteSpline` too to avoid the duplication of the basis calculations. We could also add `XZPetscMonotonicHermiteSpline` by simply doing the interpolation and...

I worked out why I was getting different warnings in CI than locally -- `fmt` was turning off that warning, but a bug meant they didn't restore the previous settings....

Hmm, `FV::Div_par_fvv` MMS fails for FCI, it gets order ~1.3, but there is a comment on the penalty term: ```cpp // Penalty terms. This implementation is very dissipative. ``` in...

I've tried [this version of the penalty term for `Div_par_fvv`](https://github.com/boutproject/hermes-3/commit/770d09e464c29b037acbc8d9df85413dc93e6192) as suggested by @dschwoerer, but it has the same behaviour as the version here. Does anyone have a derivation of...

I think the conclusion of `FV::Div_par_fvv` is that it's just a first order operator, at least for general oscillatory inputs

Thanks for the bug report @jmillan! Ah, yes, I see that the default value for `extra_arguments` is getting converted to `""` and that's being interpreted as the current directory. So...

> Do you mean the `""` after `-store-check-profile=clang-tidy-review-profile `, right? Yes, exactly > Since I'm not setting any extra arguments myself this means every run should fail, or that the...

Ahh, I've possibly worked it out, `sanitizer/common_interface_defs.h` comes from a system package, so I don't think it's an issue with relative paths. You should check to see if that header...

Ah, this is definitely a build system issue then, and I'm not sure there's much to do here. I'd try and limit generating the `compile_commands.json` for a single target, or...

Hmm, I can reproduce this locally by passing the `line-filter` argument -- as in, this seems to be a problem with `clang-tidy`. Even if I pass the whole file: ```...