science/mcstas-comps: Update to 2.5 and unbreak

[kai]

Summary

PR 234590

While I'm here:

• Add post-install target to strip binary files and create symbolic links to ${PREFIX}

/bin

• Add shebang fix for sasview_generator.py

Test Plan

• poudriere (11.2-, 12.0-RELEASE, 13.0-CURRENT@r342772 amd64 + i386) -> OK
• portlint -> OK (with a warning, that portname shouldn't contain hyphens)
• Runtime tests -> OK

Minor fixes from original patch:

• Fixed MASTER_SITES by changing ${PORTNAME} to ${PORTNAME:C/-comps\$} (I might also use just "mcstas" instead of ${PORTNAME:C/-comps\$} to make it more readable if desired)
• Removed STRIP=#none because it had no effect and always produced non-stripped binaries

[tcberner]

science/mcstas-comps/Makefile

1

^nope -- unless you have explicit permission by Erik B. Knudsen

[tcberner]

science/mcstas-comps/Makefile

18

CMAKE_ON=enable_mcstas

should work too, and reads nicer ^^

20

^ why +=

[kai]

• Re-Added "# Created by" line manually. The creator is still the actual maintainer so that line didn't get my full attention when I created this review.
• Removed CMAKE_ARGS completely as it seems not to be used by the configure/build process. I have compared the build logs with CMAKE_ARGS and CMAKE_ON enabled and in both cases a warning from CMake was generated that the argument "enable_mcstas" is not used.
• Changed "+=" into "=" for PLIST_SUB

[kai]

science/mcstas-comps/Makefile

1

Good catch, the creator is still the actual maintainer thus that line didn't get my full attention. I have now restored that line to avoid confusion.

18

That's neat and reads indeed nicer. But I have removed it completely because it seems to be unused during the whole configure/build process.

20

Oops, indeed. There's no reason to use += .

[miwi]

LGTM