[Reproducible-builds] please don't build on archs that packages don't declare support for

[Holger Levsen]

Hi Michael,

thanks for reaching out to us and reporting this…! 

On Samstag, 6. Februar 2016, Michael R. Crusoe wrote:
> I think you all are the bee's knees. However, some of the packages I
> maintain aren't compatible with 32 bit architectures, like armhf, and are
> explicitly marked as such. Yet the reproducible builds infrastructure still
> tries to build on the incompatible architecture and I get an error on
> tracker.d.o and elsewhere.
> 
> Here is an example:
> https://tests.reproducible-builds.org/rb-pkg/unstable/amd64/khmer.html
> 
> Package: khmer
> Architecture: amd64 any-arm64 any-mips64 any-mips64el any-ia64 any-ppc64el
> any-sparc64
> (from https://sources.debian.net/src/khmer/2.0%2Bdfsg-4/debian/control/ )

khmer is being tried to build (on armhf), because it has "all" in the 
Architecture field, at least as seen on 
https://tests.reproducible-
builds.org/rbuild/testing/armhf/khmer_2.0+dfsg-3.rbuild.log where the line 
reads: 
"Architecture: any-amd64 any-arm64 any-mips64 any-mips64el any-ia64 any-
ppc64el any-sparc64 any-s390x all ppc64"

This line is coming from the .dsc file, see 
http://httpredir.debian.org/debian/pool/main/k/khmer/khmer_2.0+dfsg-3.dsc

I have no idea why the .dsc file differs from the control file.

> Could, perhaps, this not happen?

That's surely the idea as you can see if you compare 
https://tests.reproducible-builds.org/unstable/amd64/index_not_for_us.html 
with
https://tests.reproducible-builds.org/unstable/armhf/index_not_for_us.html

however there's a bug somewhere.


cheers,
	Holger
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 828 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.alioth.debian.org/pipermail/reproducible-builds/attachments/20160208/0f1c6744/attachment.sig>