Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 24 May 2003 03:24:47 -0700
From:      Peter Wemm <peter@wemm.org>
To:        Mikhail Teterin <Mikhail.Teterin@murex.com>
Cc:        ports-committers@FreeBSD.org
Subject:   Re: cvs commit: ports/graphics/lcms Makefile distinfo 
Message-ID:  <20030524102447.1C7662A8B1@canning.wemm.org>
In-Reply-To: <200305221159.22458.Mikhail.Teterin@murex.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
Mikhail Teterin wrote:
> For the record, I -- the port's maintainer -- was notified of the
> package building failure by Kris as far back as May 8th. By May 9th
> I tried to reproduce the self-test failure on all architectures using
> the machines available to a ports committer as listed on
> 
> 	http://www.FreeBSD.org/internal/machines.html
> 
> Namely: beast (Alpha), panther (sparc64), and one of the plutos (ia64)
> -- the other pluto had no /usr/ports/Mk. I could NOT reproduce the
> problem and asked him to let give me access to one of the machines
> the bug shows up. Kris never explained, why that is not possible and
> insists, he is tired of my complaining.
> 
> Although this appears to be an endiannes problem, as far as I am
> concerned, the bug only exists on the package-building cluster. The
> minor technical issue is being severely aggravated by the evidently
> political one...

The bottom line is that it was a release build blocker.  Since we are
in a release build cycle, all other issues are secondary.  Kris, with
his portmgr hat on, did what was needed in order to build the ports
so that we can get the release out the door.  Now is not the time
to play 'spot the black helicopter'.

> I have little idea, how the build-cluster is set up, but I need more
> cooperation in finding and solving this strange problem.

The build cluster is set up as a chroot environment with a pristine
'make world' installed in it, and any listed binary packages pre-installed.
This is well documented.  The difference between the build cluster and
the reference boxes is that the packages on the reference boxes are
relatively infrequently updated.  Perhaps your test builds worked because
they used an old pre-existing build in /usr/local.

But regardless, now is not the time for finger pointing.  The fact is that
it was holding up the release build.  It doesn't matter if it builds OK on
everything from FreeBSD/vax through FreeBSD/atari2600.. It wasn't working
on the build cluster, and that is the bottom line.

> 	-mi
> 
> On Wednesday 21 May 2003 06:39 pm, Kris Kennaway wrote:
> = kris        2003/05/21 15:39:38 PDT
> = 
> =   FreeBSD ports repository
> = 
> =   Modified files:
> =     graphics/lcms        Makefile distinfo 
> =   Log:
> =   With my portmgr hat on, revert this port to 1.09 and bump PORTEPOCH.
> =   The port fails to pass the internal self-tests on 64-bit platforms
> =   running very recent 5.1 systems, but it is required to build KDE and
> =   therefore critical for the release.
> =   
> =   Revision  Changes    Path
> =   1.9       +2 -1      ports/graphics/lcms/Makefile
> =   1.6       +1 -1      ports/graphics/lcms/distinfo
> 
> 

Cheers,
-Peter
--
Peter Wemm - peter@wemm.org; peter@FreeBSD.org; peter@yahoo-inc.com
"All of this is for nothing if we don't go to the stars" - JMS/B5



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030524102447.1C7662A8B1>