-
Notifications
You must be signed in to change notification settings - Fork 505
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
Running cmake -LA
changes library configuration
#1847
Comments
This seems to me like something we should fix before the next release. Thoughts @dstebila @praveksharma @bhess? |
Seems peculiar, and worth fixing. I notice from the cmake documentation that there's a command-line option
So I'm cautiously optimistic about the |
As a long-term solution, I feel like we should refactor our CMake files so that this issue doesn't occur, |
Following up on this, I've confirmed that the unwanted side effects occur with |
Can we close this since #1848 has landed? |
Seems reasonable to me. |
Describe the bug
If
cmake -LA ..
is run between the library configuration (cmake -GNinja <config options> ..
) and build (ninja
) commands, the library configuration may be changed.To Reproduce
From the
liboqs
directory:As expected, the breakpoint is hit: the AVX2 code is enabled in the build.
Here is the exact same command sequence with
cmake -LA ..
added between config and build:The breakpoint cannot be set: the AVX2 code is disabled. Also note that the number of targets built is greater in the first case (127) than in the second (101).
Expected behavior
The
cmake -LA ..
call should have no side effects.Screenshots
See above
Environment (please complete the following information):
main
, at commit 5670edfAdditional context
We use this pattern basically everywhere in CI to print config variables, so the bug's impact could have been huge. Thankfully, it looks like there were no failures hidden by the bug. This branch removes all occurrences of
cmake -LA ..
from CI, and everything passes. Based on some older investigation, I think we can infer that the weekly tests are unaffected as well.I believe that the issue lies in our filtering logic when an option like
OQS_MINIMAL_BUILD
orOQS_ALGS_ENABLED
is set. I'll investigate further.The text was updated successfully, but these errors were encountered: