I'm not sure how "searching for errors" can take you more than three seconds in this case, and it still makes no sense to refer to startdir in comments.

I think that this may be an issue of misinterpreting the error message, rather than a problem of finding errors in comments. The current error, "E: File referenced in $startdir" can be interpreted as the build system not allowing files to be in the $start directory, when, in fact, it is simply indicating $startdir is not a valid variable. As such namcap correctly flags any line containing $startdir.

This is a change to the PKGBUILD structure, which is why I misinterpreted the error. PKGBUILD removed $startdir as a valid variable, yet it still interprets it without reporting an error, I'm sure this was done as to not break existing PKGBUILDS.

Perhaps a simple solution would be to simply change the error to indicate that "$startdir is no longer a valid PKGBUILD variable", which would point package builders in the correct direction of using something like "../", to reference to their root build directory.

OK, so it was unclear what the error message was trying to say- I can try and fix that up.

And your last comment is completely wrong- that is less correct than using $startdir. You should almost never ever have to reference the root build directory...do you have some valid reason for doing so?

If I want to add a script to start a daemon in /etc/rc.d. I have been including the file in the root of my build and then using the install command to move it into /etc/rc.d in the $pkg directory. How should I be doing this?

Thanks.

Never mind. I see that the build sytem creates symlinks in the $srcdir back to these files. The moral of the story is to make sure that the first time that you build a package, use a build from core as a sample template. Just because something else works, doesn't mean that it's right!

Thanks,
Kurt