(sorry for the late answer, this message must have fallen through the cracks during some mail problem, only saw it on bugs.d.o now)
On Tue, Jun 07, 2005 at 11:11:30AM -0400, Michael Stone wrote: > On Tue, Jun 07, 2005 at 04:34:38PM +0200, Michael Banck wrote: > >However, sarge is now out and efforts towards porting d-i to hurd-i386 > >are underway, so this issue should be fixed ASAP. Therefore, I propose > >to not run the test suite on hurd-i386 for the time being, until the > >issues are sorted out. > > Fine, as long as hurd isn't an official, releasable port. I will not > release another coreutils that doesn't run through the test suite (it > turns up too many problems in the build chain that could lead to some > nasty bugs.) Totally understandably. Note that the test suite failures are minor and well understood[1]. I propose to use a similar mechanism as for glibc, i.e. to not run the test suite when 'nocheck' is present in DEB_BUILD_OPTIONS. This will assure that every official build daemon will run the test suite, while making it possible for porters to circumvent it manually when needed. The patch in #278915 does this, I can NMU coreutils with that if you want. cheers, Michael -- [1] coreutils' uptime never got used on the Hurd so far, and is not being built; it is questionable whether the test suite should fail for that at all. There are long known differences between what Mach and POSIX think about priorities and their range, making the nice test fail, but people are looking into this to get some kind of mapping in place. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
