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
Add mipsel to Github Actions CI setup #1120
Conversation
Hmmm, seems like "allow failure" isn't a feature of Github Actions We could add @briansmith how do you want me to proceed? |
I factored out the first commit into its own PR; I retained you as the author of the commit: PR actions/toolkit#1121. |
I see that CI/CD fails in the build.rs step. In this PR think we should fix that, and make whatever other changes we need to make until we get to the point where the build fails at the linking step. This will help us identify which features we need non-assembly fallback implementations for. |
PTAL at PR actions/toolkit#1122 which is an attempt to adapt build.rs to support platforms that don't have any assembly code in ring. |
May I help with this issue somehow? Is it ok if I create another branch with both this changes and the changes of actions/toolkit#1122 in attempt to make both work together? |
The best way to help in the immediate term would be to review the PRs that are already posted related to this work, starting with PR actions/toolkit#1122. If you can help with the reviews, then I'll update the other PRs that are required to get MIPS working and ping you for review on them too. |
With this new commit, built on top of actions/toolkit#1174 I've been able to run Running Linking errors on mips32 test
and more... |
Yes, that's to be expected, because we don't have platform-agnostic fallbacks for every primitive yet. PRs like actions/toolkit#1011 will help with that. I think ChaCha20-Poly1305 will be the last missing one and then we'd be good to go, IIRC. |
actions/toolkit#943 is looking a bit stale, but could it be used to help with ChaCha20-Poly1305? |
Now that actions/toolkit#1174 is merged, I think if we rebase this then it will work? |
Codecov Report
@@ Coverage Diff @@
## main actions/toolkit#1120 +/- ##
==========================================
- Coverage 92.88% 92.87% -0.01%
==========================================
Files 116 116
Lines 18270 18270
==========================================
- Hits 16970 16969 -1
- Misses 1300 1301 +1
Continue to review full report at Codecov.
|
@briansmith I've rebased but something seems up with codecov 🤔 |
Codecov sometimes reports the wrong results when only a subset of the jobs have reported coverage to it. It seems like it's fine now? Anyway, it does seem like the mips jobs are failing due to warnings. It seems those failures kill the entire What do y'all think? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This PR will need to be rebased on top of PR #1274 once that is merged. That should fix the warnings/errors in rand.rs.
@@ -245,6 +246,9 @@ jobs: | |||
- target: i686-unknown-linux-musl | |||
host_os: ubuntu-18.04 | |||
|
|||
- target: mipsel-unknown-linux-gnu | |||
host_os: ubuntu-18.04 | |||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In addition to adding an entry to the test:
matrix, we should add a similar entry to the coverage:
matrix. (One of my primary interests in this PR is increasing the test coverage of the fallback implementations.)
As I mentioned above, I think PR actions/toolkit#1274 will probably fix the build errors/warnings, or at least some of them.
When reviewing PR actions/toolkit#1181, and in the ensuing discussion, I realized that PR actions/toolkit#1181 is probably almost completely unneeded now. So I think we should just extend this PR with whatever is needed to get the MIPS job to turn green. |
@briansmith I'm trying to port the changes but I'm getting quite a few It seems to be related to the recent changes on prefixes. Any tips on how to troubleshoot why |
In build.rs:
Replace |
I've done a similar change already, but it seems to be only happening on test mode. |
replaces actions/toolkit#804