FTBFS due to arch check failing to match Architecture: "all" at sbuild time

Bug #690428 reported by Dimitri John Ledkov
16
This bug affects 2 people
Affects Status Importance Assigned to Milestone
launchpad-buildd
Fix Released
Low
William Grant

Bug Description

See Bug #680889 for failing to upload.

Now I'm able to upload dsc with "Architecture: mingw32-any all' - https://code.launchpad.net/~dmitrij.ledkov/+archive/ppa/+build/2097199

Such upload gets correctly dispatched to i386 to build "all" package. But the launchpad's sbuild doesn't like it:

RUN: /usr/share/launchpad-buildd/slavebin/sbuild-package ['sbuild-package', '6a57ac88751bc9657bbdc652549d3188cab7bf08', 'i386', 'natty', '--nolog', '--batch', '--archive=ubuntu', '--dist=natty', '-A', '--purpose=PPA', '--architecture=i386', '--comp=main', 'mingw-w64_1.0+bzr2290+201012142325-0~p4.dsc']
Initiating build 6a57ac88751bc9657bbdc652549d3188cab7bf08 with 2 processor cores.
Automatic build of mingw-w64_1.0+bzr2290+201012142325-0~p4 on iridium by sbuild/i386 1.170.5
Build started at 20101214-2333
******************************************************************************
mingw-w64_1.0+bzr2290+201012142325-0~p4.dsc exists in cwd
: i386 not in arch list or does not match any arch wildcards: mingw32-any all -- skipping
******************************************************************************
Finished at 20101214-2333

Or am I suppose to wait for all build nodes to be upgraded to include fix for Bug #680889?

Revision history for this message
William Grant (wgrant) wrote :

This is a problem in another layer. Does Ubuntu sbuild handle that architecture string OK?

affects: soyuz → launchpad-buildd
Changed in launchpad-buildd:
assignee: nobody → William Grant (wgrant)
importance: Undecided → Medium
status: New → Triaged
Revision history for this message
Dimitri John Ledkov (xnox) wrote : Re: [Bug 690428] [NEW] FTBFS due to arch check failing to match Architecture: "all" at sbuild time

I haven't tried plain sbuild in ubuntu. The pbuilder and
dpkg-buildpackage do handle them just fine with/without modified
dpkg-architecture ostable. My expectation for "all ming-any" to build
just "all" package. My Debian rules detect arch at build time and can
build just "all" package. I will be using xdeb to build the others.

On 15/12/2010, William Grant <email address hidden> wrote:
> This is a problem in another layer. Does Ubuntu sbuild handle that
> architecture string OK?
>
> ** Project changed: soyuz => launchpad-buildd
>
> ** Changed in: launchpad-buildd
> Importance: Undecided => Medium
>
> ** Changed in: launchpad-buildd
> Status: New => Triaged
>
> ** Changed in: launchpad-buildd
> Assignee: (unassigned) => William Grant (wgrant)
>
> --
> You received this bug notification because you are a direct subscriber
> of the bug.
> https://bugs.launchpad.net/bugs/690428
>
> Title:
> FTBFS due to arch check failing to match Architecture: "all" at sbuild
> time
>

--
Sent from my mobile device

Revision history for this message
Dimitri John Ledkov (xnox) wrote :

No, vanilla Debian/Sid Ubuntu/Natty sbuild has the same behaviour as launchpad-buildd.

After analyzing the arch_check in sbuild it seems that packages with declared architecture "all mingw32-any" would be accepted to be build on mingw32 platforms only and a job will not start on e.g. "linux-any" to build packages with architecture "all".

Similarly "all linux-armel" by analogy would only get build on linux-armel sbuild-host. Which imho is not very good: a linux-armel specific package which has loads of documentation. You want docs to be generated on e.g. linux-i386 and arch specific package on the linux-armel.

For me it is a bootstrap problem. I need "all" package to build cross-toolchain to be able to build (using xdeb) mingw32-any package. *sigh*

Does this sound reasonable? I will try to create the minimal debian source package to demonstrate this.

William Grant (wgrant)
Changed in launchpad-buildd:
assignee: William Grant (wgrant) → nobody
Changed in launchpad-buildd:
importance: Medium → Low
William Grant (wgrant)
Changed in launchpad-buildd:
assignee: nobody → William Grant (wgrant)
status: Triaged → Fix Released
Revision history for this message
Dimitri John Ledkov (xnox) wrote :

Thank you all. I do get pleasure from acient bugs being looked after and fixed =) even if 4.5 years later =)

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.