On 11/21/11 01:14 PM, Danek Duvall wrote:
Alan Coopersmith wrote:
On 11/21/11 10:55, Rich Burridge wrote:
Or maybe even close as Will Not Fix if the system is unlikely to work without
the
system/linker package on it?
Admittedly the bug started before Ali's linker package refactoring,
when the default install didn't include all those files, because they
were in the lint-library package instead of the linker package, so it
was causing actual
problems, which should no longer be present.
https://defect.opensolaris.org/bz/show_bug.cgi?id=18575 was the original report
before the move to bugster.
I guess the question is whether these particular files could ever be safely
placed in a package that was allowed to be minimized off the system. I
don't know why that wouldn't be the case, and so my suggestion would be to
have direct and explicit dependencies on them, should they ever move.
That said, it seems a bit unlikely that they would move, and if they do,
the final dependency on system/linker wouldn't prevent you from removing
their new package -- you'd need for gcc to incorporate system/linker at a
version that would never allow the files to be removed. While I think
that's a fine idea, I think we're a bit early in history for that to solve
more problems than it might cause.
On the other hand, the file-level dependencies in the manifest help us
understand what the underlying dependencies actually are, and that's useful
for the maintainers.
I would prefer that we track the file level dependencies because it
allows us to better understand our actual needs and because it means
that any refactoring that occurs is less likely to lead to missing
dependencies in the future.
-Norm
_______________________________________________
userland-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/userland-discuss
