biology/hyphy: Update to 2.5.46hf1

[jrm]

Yuri and Robert, adding you since you both seem interested in this port. poudriere testport looks good for 12/13/14 adm64 and 12/13 i386. I don't have any arm or lower tier hardware to test, so if you're interested in those platforms, please chime in. I'm still looking into occasional seg faults running the FEL and now CONTRAST-FEL tests.

[fuz]

Will do a test, possibly later today. Currently busy making some nice chili.

[diizzy]

Upstream seems to be very keen on adding compiler (optimization) switches which will break (such as -march=native -mtune=native) and potentially also override ports framework.
Having a quick look at CMakeLists.txt it looks like you could rip out much of the optimization and rely on CPUTYPE instead?

biology/hyphy/Makefile

2

Doesn't this get strangely translated for pkg? Upstream calls this just 2.5.45 (according to their release page) so just define the tag manually
https://github.com/veg/hyphy/releases/tag/2.5.45hf1

[jrm]

biology/hyphy/Makefile

2

% make -VPORTVERSION
2.5.45.h1
% pkg version -t hyphy-2.5.45.h1 hyphy-2.5.46
<
% pkg version -t hyphy-2.5.45.h1 hyphy-2.5.45
<

They call the release 2.5.45, but it's tagged as 2.5.45hf1. There is a 2.4.45 tag, but it's associated with an earlier snapshot when there were problems (with SSE optimizations). As long as upstream doesn't do something foolish like tag something newer with the label 2.5.45 we should be fine. I'd rather not make up a version for the port/package. This is the least worst solution that I can think of.

[fuz]

Build-tested succesfully with Poudriere on armv7 arm64 FreeBSD 13.1.

[diizzy]

biology/hyphy/Makefile

2

Ok, I guess that's fine.

[jrm]

Upstream has updated to 2.5.46

[jrm]

In D37975#862788, @fuz_fuz.su wrote:

Build-tested succesfully with Poudriere on armv7 arm64 FreeBSD 13.1.

Are you up for trying those again with 2.5.46?

[jrm]

Upstream fixed all the tests and re-tagged.

https://github.com/veg/hyphy/issues/1555