This makes the shebangfix.mk code a lot more complicated.

It would be better to simply add three variables, SHEBANG_FILES_INSTALL, SHEBANG_GLOB_INSTALL and SHEBANG_REGEX_INSTALL, create a new target that gets hooked up in the stage phase and enable one or two of the targets depending on which variables are set.

As a side note, if you need to "patch" shebangs after the build process has run, you are doing it wrong, you should provide the correct path, or patch the build process so that the correct shebangs are generated.

And if all this is only to fix the couple of ports that extract their distfiles directly in STAGEDIR during install, I think those should be fixed instead, not the framework.

I tend to agree with @mat on this

In D10554#219966, @mat wrote:

As a side note, if you need to "patch" shebangs after the build process has run, you are doing it wrong, you should provide the correct path, or patch the build process so that the correct shebangs are generated.

Patching at patch time is impossible with certain ports. As mentioned in the original description, for instance, print/tex-xetex installs directly to stage/ from a distribution tarball.

I agree patching earlier (in WRKDIR) is preferable.

That said, one could install from dist tarball to WRKDIR, then add an additional step to copy from WRKDIR to STAGEDIR. That adds complication to the port and twice the storage needs (and ensuing extra I/O time). I guess I'd prefer the ability to do some in-place shebang fixes in STAGEDIR.

In D10554#219965, @mat wrote:

This makes the shebangfix.mk code a lot more complicated.

It doesn't change the guts of the shebangfix.mk code at all (hence the .USE) - just facilitates a way to provide a different file list and a different directory.

It would be better to simply add three variables, SHEBANG_FILES_INSTALL, SHEBANG_GLOB_INSTALL and SHEBANG_REGEX_INSTALL, create a new target that gets hooked up in the stage phase and enable one or two of the targets depending on which variables are set.

That (a new target that is hooked up in the stage phase) is exactly what this patch does. Maybe you're saying that you think it would be better to duplicate fix-shebang via copy/paste rather than leverage .USE?

In D10554#219970, @mat wrote:

And if all this is only to fix the couple of ports that extract their distfiles directly in STAGEDIR during install, I think those should be fixed instead, not the framework.

I think the fix for print/tex-xetex would be (as I mentioned a couple comments up) to add an extra step to extract the distribution tarball to WRKDIR, then shebangfix, then copy from WRKDIR to STAGEDIR. The tarball is currently:

du -h /usr/ports/distfiles/TeX/texlive-20150523-texmf.tar.xz
1.8G /usr/ports/distfiles/TeX/texlive-20150523-texmf.tar.xz

Uncompressed that is 3.9G.
Tar extraction takes 3-4 minutes on a moderate machine.
So the extra copy takes another 3.9G on top of that and roughly a similar amount of time as the extract to copy (and a little extra time for the extra removal during clean).

If that is preferred over shebangfix in-place changes in STAGEDIR and a reduced footprint, I'll work up a patch for print/tex-xetex to do that and drop the shebangfix.mk patch.

In D10554#219975, @rfyu28uyeg_snkmail.com wrote:

If that is preferred over shebangfix in-place changes in STAGEDIR and a reduced footprint, I'll work up a patch for print/tex-xetex to do that and drop the shebangfix.mk patch.

If this patch was to fix hundreds of ports, I would say why not. The only reason a couple of tex ports do nothing but directly extract their distfile in STAGEDIR without going through the normal process is because they are just data files that need to be repackaged. I think the best course of action is to do the patching in place in STAGEDIR after extracting the files.