You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm a new user of this library, so apologies in advance if I've missed something obvious!
I was hoping to use it as a drop-in replacement for the native open() function, but my existing code uses mode='x' to prevent accidental overwriting in case of a filename collision. I couldn't see any indication that's supported here, but I did find that S3 has recently added support for conditional requests, and Cloudflare R2 (which I'm using right now) supports the same headers, so it looks like it'd now be possible to add in a way that's not subject to race conditions.
Is this something that's planned or would be of interest? Seems like it'd be a useful addition that fits nicely within the scope of matching the existing python filesystem API.
The text was updated successfully, but these errors were encountered:
Specifically: it will work locally, but the s3 PR is not yet done, because the mocker library doesn't implement it :|. I believe it does work, and I intend to implement it for GCS amd azure and maybe others. None of this is released.
Got it, and thanks for the quick follow up - I'll be fine running it directly from the branch for now, it should still save me a good amount of time being able to use with my code as-is.
I'm a new user of this library, so apologies in advance if I've missed something obvious!
I was hoping to use it as a drop-in replacement for the native
open()
function, but my existing code usesmode='x'
to prevent accidental overwriting in case of a filename collision. I couldn't see any indication that's supported here, but I did find that S3 has recently added support for conditional requests, and Cloudflare R2 (which I'm using right now) supports the same headers, so it looks like it'd now be possible to add in a way that's not subject to race conditions.Is this something that's planned or would be of interest? Seems like it'd be a useful addition that fits nicely within the scope of matching the existing python filesystem API.
The text was updated successfully, but these errors were encountered: