Jens Petersen

Builds fine for me locally on Stackage Nightly and LTS 23 (using Fedora Rawhide). Well the next lts-23 hopefully (the current one still mistakenly pins a few gi-* packages) nix...

(No matter how wide the terminal, dnf5 update always stretches the package output to use the maximum available wide.)

I would like to propose a better behaviour: to ignore cabal.project once it is established that the immediate .cabal file is not part of that project.

Above are for Fedora Rawhide. I got similar failures on Fedora 41 & 42: https://koji.fedoraproject.org/koji/taskinfo?taskID=131255475 I tested with both ghc-9.8.4 and 9.6.6. Also same failures with 1.0.4: https://kojipkgs.fedoraproject.org/work/tasks/6452/131256452/build.log Note the...

Oh more relevant/interesting perhaps: the testsuite passes on s390x for 1.0.0

Perhaps related but the crypton-x509-validation testsuite also failed on [s390x](https://kojipkgs.fedoraproject.org/work/tasks/9160/131259160/build.log) and [ppc64le](https://kojipkgs.fedoraproject.org//work/tasks/9445/131259445/build.log). 61 out of 582 tests failed (22.27s) Test suite test-x509-validation: FAIL Test suite logged to: dist/test/crypton-x509-validation-1.6.14-test-x509-validation.log

> Doesn't this happen on `tests/KAT_Ed25519.hs`? I need to check again

More weird, on RHEL 10 (EPEL 10) the 0.34 testsuite "fails" on x86_64 but not other archs. This is with ghc-9.6.6 and package versions from Stackage LTS 22 (not latest)....

(And also again paralleling the reversal of the earlier results: crypton-x509-1.7.7 testsuite also [fails](https://kojipkgs.fedoraproject.org/work/tasks/9608/132249608/build.log) on EPEL10 x86_64) ``` Test suite test-x509: RUNNING... Test suite test-x509: FAIL Test suite logged to:...

Hmm, also for crypton-x509-validation-1.6.12-test-x509-validation: ``` : key-usage none: OK valid: OK invalid: OK key-purpose none: OK valid: OK invalid: OK exhaustive: Test suite test-x509-validation: FAIL ``` Again no result shown...