Still working on it. I'm only cleaning up the changes but I've been having issues with the autopkgtest infrastructure since the beginning of the week (if I trigger tests within "too" quickly, the testbed setup fails).
Status is:
- amd64 and arm64 pass,
- ppc64el fails with 2, 3, or 4 failures,
- s390x fails early with a segfault
- not sure about i386 and armhf which I considered as second-class here.
A number of tests are skipped because they are basically performance tests and therefore easily fail in CI. I need to investigate if the ppc64el failures fall into this.
The s390x issue might be related to endianness but imitating the endianness patch in libtraceevent didn't yield success and I didn't have much time to devote to that on top of the rest.
It doesn't seem reasonable to spend much more time in a row on this at the moment. We can work on better ppc64el and s390x over time. Is that OK for that MIR?
Still working on it. I'm only cleaning up the changes but I've been having issues with the autopkgtest infrastructure since the beginning of the week (if I trigger tests within "too" quickly, the testbed setup fails).
Status is:
- amd64 and arm64 pass,
- ppc64el fails with 2, 3, or 4 failures,
- s390x fails early with a segfault
- not sure about i386 and armhf which I considered as second-class here.
A number of tests are skipped because they are basically performance tests and therefore easily fail in CI. I need to investigate if the ppc64el failures fall into this.
The s390x issue might be related to endianness but imitating the endianness patch in libtraceevent didn't yield success and I didn't have much time to devote to that on top of the rest.
It doesn't seem reasonable to spend much more time in a row on this at the moment. We can work on better ppc64el and s390x over time. Is that OK for that MIR?