gtristan commented on issue #1335: URL: https://github.com/apache/buildstream/issues/1335#issuecomment-1635253087
> Okay. Then yes, it fixes the symptom. I still don't like the idea of downloading gigabytes of unnecessary failure buildtree into CI from CAS but seem I am having hard time convicing everyone the problem is not just the lack of retry but the downloading of failure artifacts. This seems like a nice optimization to do but I would consider that as orthogonal to this issue, currently artifacts are downloaded as a unit, before buildstream knows that it is a failure or not. That said, it may be possible to have the buildstream client start by downloading the artifact metadata and conditionally downloading the artifact content in the case that the client doesn't want the data (which _might_ be determined by whether we have `--retry-failed` specified). It sounds like a good enhancement, it would be good to file/handle that in a separate issue. -- 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]
