Closed goekce closed 12 months ago
Hi @goekce,
Thanks for this PR.
remove -f because the listed directories will always exist \ This may help to catch future errors
I am not sure about this part. When I test locally on my laptop (not in Docker), running rm -f <non-existing directory>
, a success status (0) is still returned. Also, I'm always a bit worries about permission issues with Docker, when there's a cross-over between host and container permissions, like in this case. Would you mind removing that commit?
rm -f <non-existing directory>
, a success status (0) is still returned I believe there is a misunderstanding Marc.rm -f
always succeeds andrm
will output an error if a directory/file does not exist. This way you can catch future errors.when there's a cross-over between host and container permissions
I don't understand this point, can you rephrase?
@goekce sorry - I made a mistake trying to test this locally during the weekend. You are right, rm -r
will end with an error status if one or more of the folders do not exist, making it possible to catch a potential future name mismatch re-occurrence.
when there's a cross-over between host and container permissions
I don't understand this point, can you rephrase?
It was out of an abundance of caution, but it's easy to introduce file ownership permission issues when bringing host files into a container. e.g. if the uid
of the user on the host is different than in the container. Then, depending on the umask
the files had on the host, a different user/uid
might be prompted before the files/folders can be removed, if rm is used without "f" (or they could be impossible to remove altogether without running chmod
first. However in this case here, I believe the folders are removed by user root
, and so it should not matter :)
What it does
Fixes wrong paths and removes an unnecessary option.
How to test
Rebuilding the Docker image
Review checklist
I don't know what thoroughly means however I could successfully build the Docker image.
closes #309