On Mon, Oct 20, 2025 at 12:31:47PM +0200, Rafael Sadowski wrote:
> On Mon Oct 20, 2025 at 12:20:25PM +0200, Theo Buehler wrote:
> > On Mon, Oct 20, 2025 at 11:20:22AM +0200, Rafael Sadowski wrote:
> > > This is cmake4 related. Simple update glm to 1.0.2. OK?
> > 
> > > +CONFIGURE_ARGS +=      -D=GLM_BUILD_TESTS=ON
> > 
> > Is that first = a typo or is that weird cmake syntax?
> > 
> 
> Yes it's a copy&waste typo but it enables the test, which is
> funny. Anyway new diff below:

should the static casts not be to <long long> to match the %lld?
(ll corresponds long long, d means signed):

> --	std::printf("inverse<%s>(%f): %lu\n", Message, static_cast<double>(Diff), EndTime - StartTime);
> -+	std::printf("inverse<%s>(%f): %lld\n", Message, static_cast<double>(Diff), EndTime - StartTime);
> +-	std::printf("inverse<%s>(%f): %lu\n", Message, static_cast<double>(Diff), static_cast<unsigned long>(EndTime - StartTime));
> ++	std::printf("inverse<%s>(%f): %lld\n", Message, static_cast<double>(Diff), static_cast<unsigned long>(EndTime - StartTime));

I don't know how to test this beyond building the direct consumers
(which I assume you have), so I'm a bit reluctant to give an ok.

You will need to bump all direct consumers, as usual for header-only
libs.