-
Notifications
You must be signed in to change notification settings - Fork 943
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Some tests fail on PowerPC [84% tests passed, 4 tests failed out of 25] #1088
Comments
This might be related to #1097. Can you please retry with this change. |
@COM8 Thanks, I will try this today. |
@COM8 Does not work yet, at least with no extra flags: #1097 (comment) |
Do I understand it correctly you are using macOS on powerpc 32bit? So I can setup a VM for debugging. |
@COM8 With 5772153 patch added (on top of the earlier), the build succeeds with no issues (i.e. I do not have to disable errors/warnings), however looks like test performance regressed.
(I interrupted two tests, since they apparently froze; if long execution is expected, please let me know.) |
Interesting... For the CI I can not observe such a change. Here for example old as well as new fedora debug and release test runs with openssl take the same amount of time ~2:45min. But thanks for testing my changes! |
@COM8 That may be somewhat non-trivial: you will need either to install 10.5 ppc in Qemu/UTM or install 10.6 with Rosetta in a VM on x86 machine. The second option is likely easier, but then everything needs to be built from scratch, which gonna take quite a while. I do have a working set-up in VM on Intel which supports building and running ppc binaries, so if you will really want to try that, I should be able to provide whatever may be needed, but it is pretty involving; I do not really expect anyone to go that far :) |
Let me try not interrupting anything then (as long as it a single test does not free beyond 10 min). Will update on the outcome in a while. |
Nah, it does not become better. A bit odd that outcomes are not consistent: on one go tests may take much more time to fail or just freeze, the same tests may not freeze on another go. |
Thanks for the update! Regarding the ppc thing: PPC, 32bit and Apple might be a bit too many variables to have to deal with. |
Debian supported ppc32, though not in the latest release, AFAIK. It should be possible to run it via Qemu on x86: https://linuxjedi.co.uk/2022/02/23/emulating-ppc32-linux-in-x86_64-linux Some other Linux distros running on ppc32: https://www.websentra.com/best-powerpc-linux-distros/ OpenBSD, NetBSD and FreeBSD support ppc32 (all in the latest releases, FreeBSD seems to have dropped it in current, OpenBSD and NetBSD keeps supporting it). |
OK, thanks for the pointer. For now I consider it out of scope. |
Description
P. S. I had to interrupt
cpr_put_tests
since it apparently freezes. (Or is it supposed to take a long time to run?)Specifically, here are cases that fail:
Example/How to Reproduce
Run the test on [32-bit?] powerpc.
Possible Fix
No response
Where did you get it from?
GitHub (branch e.g. master)
Additional Context/Your Environment
The text was updated successfully, but these errors were encountered: