Date: Sun, 11 May 2014 17:22:07 -0600 From: Warner Losh <imp@bsdimp.com> To: Craig Rodrigues <rodrigc@FreeBSD.org> Cc: freebsd-current Current <freebsd-current@freebsd.org>, "jenkins-admin@freebsd.org" <jenkins-admin@freebsd.org> Subject: Re: Build failed in Jenkins: FreeBSD_HEAD-modules #386 Message-ID: <9A75A814-1B7B-45DB-960D-346A23F9F99A@bsdimp.com> In-Reply-To: <CAG=rPVeR3yWsYfu-USFbbVoZB4Fua1V6hBnMgZF=%2BtiXFMq_dw@mail.gmail.com> References: <254143261.0.1399675169702.JavaMail.jenkins@jenkins-9.freebsd.org> <CAG=rPVdHR29qMsvZVfVxBgoX4Jg7sXXWePZ9%2BvvyPOcLqL5R8A@mail.gmail.com> <26E92F41-F849-43E2-AC1F-553F1AB7B11F@bsdimp.com> <CAG=rPVd8JVLxV1pFUVWsZBX1OnFAEUU6MeBugO3JH7Yf0gSXnA@mail.gmail.com> <130417FC-CE39-43DE-B87F-DF2185FC94DC@bsdimp.com> <CAG=rPVeR3yWsYfu-USFbbVoZB4Fua1V6hBnMgZF=%2BtiXFMq_dw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] On May 11, 2014, at 5:00 PM, Craig Rodrigues <rodrigc@FreeBSD.org> wrote: > On Sun, May 11, 2014 at 3:25 PM, Warner Losh <imp@bsdimp.com> wrote: >> >> On May 11, 2014, at 1:43 PM, Craig Rodrigues <rodrigc@FreeBSD.org> wrote: > >>> make[1]: "/builds/FreeBSD_HEAD-modules/sys/modules/aac/../../conf/kern.opts.mk" >>> line 24: Could not find ../../share/mk/bsd.mkopt.mk > > Instead of explicitly specifying the path to "../../share/mk", > should some additional logic be added to respect > MAKESYSPATH? The only logic that can work is to test MAKESYSPATH and respect it if it is defined, and use the relative path if not (since we don’t require MAKESYSPATH to be defined to do a out-of-tree module build, and could be built on a system that isn’t running the latest current). > Some people building products do build the kernel and modules out of the > FreeBSD source tree, and put their mk files in a directory > which is not src/share/mk. That’s not really a supported configuration. Using JIRA to force it is rather an annoying use of JIRA since it makes making JIRA happy hard to test and outside the normal work flow. This will cause pain in the future, and I also object to my changes being characterized as “wrong” when such a non-standard setup has never been explicitly supported, required or part of anybody’s normal workflow. Grump. I’ve prepared a compromise commit, but I’m not happy about it, and think it is a horribly ugly kludge. But I can think of no other ways to keep both use cases working. Warner [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Comment: GPGTools - https://gpgtools.org iQIcBAEBCgAGBQJTcAYfAAoJEGwc0Sh9sBEAQVEQAMMKrYKtwrxN9JUq4D1G6gIQ E7xyxKkCLjvbQObAxGhi85aWat4t5x4BJEOjnADT9QGrWAxncYZER+Xinq5H2AiJ KB44lh3KPPfAmlxSE6+xfpCJAaALQc0owacsRNnEKlYLphZYbrJuNlndUYGXG+xz s2ofvUKqiR4pDP55MopDqPcxZyFZyyVDJkDrpDLJYDjRgjIWJVCs+rkUwxA8V6Fd tkb4nYP/3FfwuwoiUc914SIUGMgOh9yV6DCDxD9AqGByM4JfY8Ckk3QO70O02p53 KxLRcaJZThr+9K4B6tuHz6t24BbO1iUUE/ktg3Am8A1oFOIU2+j5cTVaa1874EgH SnINPdJ31zUADiOMBPejb15jpBcKDOrhUOm7zmHYdwnHR5cE166aU8KGanI9TjSR CvSm4SflreAKTkV7orMvg1MGFHE04I6bv9gCd/g6vVA/zA3cgOqzHMCTQxWyhRVJ 8+SPyf6R51c/X5SRF+isNs1OH3trFZuZBnjSqbovBlwN+DH98OFd1nP8c/+Xgjgo K+Jiyc7QsGH1GADr5cWq9kB1tyOTlNxtry5VzGVWtbN+ga28qaHIlVuQad3gLJCO TtyMo8czJUtpXgh+cYnLyT1A8oLIdSGQLbHsGhpypK5K42G/uRJMKJLZHi1i2osW d++f/cbwsVOO9zzRMEsp =6eCh -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9A75A814-1B7B-45DB-960D-346A23F9F99A>