On 2020/02/20 16:27, Marc Espie wrote: > Now that we have a settled down workflow, I think this should work just > fine, and avoid regenerating that Makefile 30 times in php... > > Specifically, that stuff only depends on static port information, it can't > be generated before fake because it lives in WRKINST, and it only needs > to be re-gen'd each time we run update-plist. > > That's assuming you don't change PKG_ARGS, but then, it's likely lots of > other things will fail as well.
There is a possibility that somebody may edit the PLIST without using update-plist (e.g. run update-plist, make packages, notice a problem with the plist, edit by hand, repackage) - would it make sense to clean it here too? Otherwise, makes complete sense to me. Index: bsd.port.mk =================================================================== RCS file: /cvs/ports/infrastructure/mk/bsd.port.mk,v retrieving revision 1.1517 diff -u -p -r1.1517 bsd.port.mk --- bsd.port.mk 14 Feb 2020 13:06:29 -0000 1.1517 +++ bsd.port.mk 20 Feb 2020 16:02:04 -0000 @@ -3223,7 +3223,7 @@ _internal-clean: . endif .endif .if ${_clean:Mpackages} || ${_clean:Mpackage} && ${_clean:Msub} - ${_PBUILD} rm -f ${_PACKAGE_COOKIES} + ${_PBUILD} rm -f ${_PACKAGE_COOKIES} ${_WRKDEBUG}/Makefile ${_PFETCH} rm -f ${_CACHE_PACKAGE_COOKIES} rm -f ${_UPDATE_COOKIES} .elif ${_clean:Mpackage}