Created attachment 178501 [details]
I tested all the edge cases I could think of, and find, I don't expect much failures.
There is a problem with the handling of strip flags
emulators/xen-kernel and sysutils/xen-tools fail to patch for instance.
there must have been a bug in the old code, with some variables being left over from the previous patch, fixing these two.
Yeah, the strip instruction was not wiped between patches, so the -p of one watch would be applied to the next if it did not have one.
Created attachment 178702 [details]
A few other failures (exp-run is still running):
Created attachment 178747 [details]
Spaces in a WRKSRC directory where not an edge case I found, fixed. (along with a couple of other issues)
There are some ports that use or set PATCH_ARGS and PATCH_DIST_ARGS directly
Ah, yes, I had it on my todo list and forgot about those.
(In reply to Antoine Brodin from comment #7)
Is it worth also checking for Makefile.local in the external tree? It makes sense to want those in VCS too and it'd allow for altering PATCH_ARGS at the same time.
Created attachment 178749 [details]
I don't think it is, and in any way, it would be a different changeset.
Also, I don't remember anyone asking for that.
In a general case, we try to not add functionalities nobody asks or want, because once we have them, we have to maintain them :-)
Exp-run looks fine.
A commit references this bug:
Date: Mon Jan 16 16:47:02 UTC 2017
New revision: 431680
Extract do-patch into a separate script.
Submitted by: mat
Exp-run by: antoine
Sponsored by: Absolight
Differential Revision: https://reviews.freebsd.org/D9029