On 04/10/12 11:58, Danek Duvall wrote:
Brock Pytlik wrote:
While the typical ON way is to put dummy targets in directories where
something doesn't apply (install_h being one notable such target), I don't
think that's something we'd want to replicate here.
Ok, that makes sense to me.
Since I'm in this space, the other thing I'd like to do is remove setup.py
lint, and instead have lint in the makefile depend on client and pylint.
Thoughts?
So the issue that I see is that the top-level lint target never bothers to
lint the zoneadm directory. The only C linting that gets done is when you
do "cd zoneadm; make lint", at which point you have a bunch of lint
warnings, and -mt getting passed erroneously to lint.
Right
I guess I don't care whether or not C linting is done exclusively through
Makefile infrastructure or not, but we'll need to continue linting the C
module source, and that's likely to be harder (or at least duplicative) to
do in the Makefiles. I guess I'd explore a hybrid approach -- either stuff
the lintable subdirectories into setup.py (which could then simply run
"make lint" in each), or jigger the makefile to run "setup.py lint" and
then descend into lintable subdirs.
FWIW, though, that needn't be part of this wad, which I'd rather get in
sooner than later. Lint-cleanliness and a proper infrastructure for it can
come soon after.
Ok, I'll make the change where we remove the exit(0) line from the code
and simply land that change as a fix for this bug, and file bugs for the
other build issues.
Danek
Thanks for taking a look,
Brock
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
