The idea would be to host the code coverage results for each pull request. It is possible to retain an artifact per build in github, but they must be zip archives which cannot be directly browsed.
It would be ideal to hook into github events to set up actions to create a new coverage report for a pull request and store it in a matching folder (ie. PR123/..). Then, when a pull request is resolved, this coverage folder will be deleted.
Github pages could be used to host a pull request, but a consideration is that if every coverage report is added to the repository, it will grow much faster than the repository it follows. This could potentially be remedied by removing coverage directories from git history:
https://stackoverflow.com/questions/10067848/remove-folder-and-its-contents-from-git-githubs-history/32886427#32886427
Another possibility would be to utilize an S3 bucket with a CloudFront CDN hosting it. This would allow each coverage report to be uploaded to the bucket as a folder. Note, in this case, cleanup when the pull request is resolved would actually be optional. It would be possible to simply set a retention period for the bucket such that uploaded files eventually disappear.
The idea would be to host the code coverage results for each pull request. It is possible to retain an artifact per build in github, but they must be zip archives which cannot be directly browsed.
It would be ideal to hook into github events to set up actions to create a new coverage report for a pull request and store it in a matching folder (ie. PR123/..). Then, when a pull request is resolved, this coverage folder will be deleted.
Github pages could be used to host a pull request, but a consideration is that if every coverage report is added to the repository, it will grow much faster than the repository it follows. This could potentially be remedied by removing coverage directories from git history:
https://stackoverflow.com/questions/10067848/remove-folder-and-its-contents-from-git-githubs-history/32886427#32886427
Another possibility would be to utilize an S3 bucket with a CloudFront CDN hosting it. This would allow each coverage report to be uploaded to the bucket as a folder. Note, in this case, cleanup when the pull request is resolved would actually be optional. It would be possible to simply set a retention period for the bucket such that uploaded files eventually disappear.