Hello there,
it seems (pending Reproducible-Builds results for CUPS 2.3.3op2-3) that I
might have gotten to let CUPS be reproducible.
\o/
Let's wait on
https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/cups.html
This was done at the expense of three reprotest-specific patches:
- https://salsa.debian.org/printing-team/cups/-/commit/697065c444602c691dd364a315680a12f17cc59c
Let the httpAddrGetHostname() test fail under reprotest. The FAIL is still
reported, but not counted.
This has been a longstanding undebuggable error (for me); it seems CUPS
expects hostnames to be setup in a way that is incompatible with
reprotest. It _does_ work fine in pbuilder and sbuild; so the test passes
on Debian buildds.
- https://salsa.debian.org/printing-team/cups/-/commit/70966a573839b0e5c3574d51ce7ef1a358fcb285
Do not run the test/run-stp-tests.sh script if run as root. As above,
fail with a message, but no abort.
Similarly, the tests will run in pbuilder, sbuild,and official buildds,
only not in reprotest.
- https://salsa.debian.org/printing-team/cups/-/commit/f92992e02e7cfb62679ca70c0741da0adecf32b3
Run testlang for each provided CUPS locale and only for these.
I reported this one over at https://github.com/OpenPrinting/cups/pull/85 ,
where Michael Sweet and me disagree on the proposed patch (or, to be more
precise; they want a MacOS fix that I can't provide).
As things stand, it's not ideal (as not all tests run under reprotest), but
it seemed important to bring the Salsa CI to a fully-green status, and
(perhaps) make CUPS reproducible in time for Bullseye!
Fingers crossed,
OdyXAttachment:
signature.asc
Description: This is a digitally signed message part.