Re: armhf regressions in a few packages

Bus error tends to be about unaligned memory access. I personally don't fix such issues, but I do report them upstream and then disable that architecture for building.

For bioinformatics, there is no reason to support armhf.

--
Michael R. Crusoe

On Sat, Nov 21, 2020, 22:11 Nilesh Patra <npatra974@gmail.com> wrote:

Hi,

I've seen some packages failing _only_ on armhf when I added autopkgtests to them. Murasaki[1] and lighter[2] for example.
They seem to fail with this common error message: <pid> Bus Error -- <command>

Could it be because of lack of RAM/memory issue there?

Couple of days ago, Praveen (in CC) reported to me an error with golang-github-honnef-tools[3]
But on running in an armhf porter box, it works just fine as I'm told.

I've also noticed that the tests in these packages need a _relatively_ higher computation power.
I'm unsure if this is a fault in the tests, or a limitation at CI's end. Please let me know what you'd think about this and plausible solutions to this.

[1]: https://ci.debian.net/data/autopkgtest/testing/armhf/m/murasaki/8213738/log.gz
[2]: https://ci.debian.net/data/autopkgtest/testing/armhf/l/lighter/8203401/log.gz
[3]: https://ci.debian.net/data/autopkgtest/testing/armhf/g/golang-honnef-go-tools/8213705/log.gz

Regards,
Nilesh