math/superlu: Fix mastersite and update to 5.2.2

[rhurlin]

Summary

First and foremost, the PR should provide a fix for the MASTER_SITES that no longer work.
In the process, it was also updated and expanded a bit.

Changelog: https://portal.nersc.gov/project/sparse/superlu/changes.html#slu_change

Changes in the port beside the fix of MASTER_SITES

• Add USES=pkgconfig
• Remove PLIST_SUB (not needed)
• Update WRKSRC
• Include EXAMPLES (also Fortran examples)
• Also download the technical report, which describes the implemented ILU algorithm
• Update files/patch-SRC_CMakeLists.txt
• Update pkg-plist

Guangyuan Yang <[e-mail]> opened PR 252271, the patch is from me.

Tested on Poudriere (amd64, i386). 'portlint -AC' is ok.

Because of the fix of the MASTER_SITES, Guangyuan Yang and Fernando Apesteguia in the PR voted to not waiting for maintainers approval.

What do my mentors think about it?

[arrowd]

math/superlu/Makefile

24 ↗

(On Diff #81433)

Isn't it the default?

[rhurlin]

math/superlu/Makefile

24 ↗

(On Diff #81433)

Isn't it the default?

No, make -V WRKSRC gives .../work/superlu_5.2.2 (with underscore instead of minus) ;)

Do you think I have to wait for maintainers approval or is it justified in this case to commit the fix with update and some changes to the port?

I am not sure and of course I can wait. On the other hand, math/superlu is a dependency of math/eigen3 is a dependency of ...

[arrowd]

I'd say, commit it.