linguini1 commented on PR #18500:
URL: https://github.com/apache/nuttx/pull/18500#issuecomment-4012958589

   > The main problem I see with this tool is that CI runs builds in parallel, 
and this is configured via YAML and handled by github actions. All this happens 
automatically and I don't know if it's possible to fine-tune the built targets 
without completely reworking the CI.
   
   Yes, that would be an important requirement!
   
   > EDIT: probably we could inject your script here:
   > 
   > 
https://github.com/apache/nuttx/blob/f4a6e62f7f72c47df8d0146a19b4060c48f508aa/.github/workflows/build.yml#L222-L228
   > 
   > or integrate it with `cibuild.sh` tool
   
   My (perhaps naive) idea was to use this tool to either generate the `.dat` 
files used by the CI for selecting builds, or ideally just pass the output 
directly to the build workflow through the GitHub workflow output variables.
   
   > integration with select.py means we add python dependency
   
   I figured this would be alright for CI, since Python is used in a few other 
places for CI already. It also means the script is cross-platform, so no `.sh`, 
`.bat` copies.
   
   > another idea that is probably the least risky is split tools/ci/testlist 
into "chips" (stm32, ntf52 etc), and then just use github labels to select what 
should be build.
   
   That works, but then we have to maintain quite a few more labels and `.dat` 
files. We'd also have to update them every time a chip is added. The additional 
benefit of the script is being able to select past the granularity of chips all 
the way to boards and even specific configurations.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to