On 11/24/22 18:23, Drew Parsons wrote:
On 2022-11-24 17:54, IOhannes m zmölnig wrote:hi, getfem-5.4.2+dfsg1-1 has been uploaded recently (thanks). unfortunately, it fails to build on s390x (see [1]), resulting in #1023788 which in turn now threatens one of my packages to be removed from testing. i would like to avoid that. it's only been two weeks that #1023788 was filed, but given that the freeze is rapidly approaching and there hasn't been *any* reaction so far on the bugreport, i thought i'd check with you people...Hi IOhannes, upstream is actually quite active, but s390x is not the simplest architecture to fix problems for.
i thought so too (at least for the lay person that has no access to a porterbox :-)).
but recently (when debugging yet another BigEndian issue (they seem to pop up right before the freeze)), somebody mentioned how to use docker to run s390x via qemu.
here's a write up i created with step-by-step instructions (few parts are specific to the software we wanted to debug):
https://github.com/csound/csound/issues/1650#issuecomment-1324921328 i think this is a viable solution for helpful upstreams.
Could perhaps just skip testing on s390x. I figure not so many people will be using this software on it.
maybe. but who knows.i strongly believe this is a BigEndian issue, and i think that software in general should care for BigEndian (even though it seems there are currently few BE architectures in wide use)
gfamsrd IOhannes
Attachment:
OpenPGP_0xB65019C47F7A36F8.asc
Description: OpenPGP public key
Attachment:
OpenPGP_signature
Description: OpenPGP digital signature