Altered get_bucket to use HEAD.
#2082
Conversation
This is mostly backward-compatible (& less expensive). However, if you were parsing exception messages from this method, you should refer to the documentation on how to change your code.
| the bucket exists. | ||
|
|
||
| If the default ``validate=True`` is passed, a request is made to the | ||
| service to ensure the bucket exists. Prior to Boto v2.25.0, this fetched |
There was a problem hiding this comment.
maybe use .. versionchanged:: for this?
|
I would still like to hear what others think about the change, and see if anyone has any other concerns. The code itself, provided everyone is ok with the changes, looks good to me, so |
to the code.keys weren't being fetched before.
|
I updated the wording on the docs & added the |
|
I did confirm by experimentation that using |
In reaction to #2078 (& the mentioned blog post), this pull alters the behavior of
S3Connection.get_bucket. This switches from the old approach of trying to fetch keys in the bucket to utilizingHeadBucket.Some further notes. I back-ported the old method to
GSConnection, since it doesn't appear Google Storage supportsHEADon a bucket - https://developers.google.com/storage/docs/reference-methods. I'd love to be corrected on this, but for now, I copied over the old behavior. I don't have a GS account to test with, so I'm really hoping I didn't break it. 馃槙Beyond that, this is mostly backward-compatible (& less expensive). The only thing that hurts is that
HEADrequests (rightfully) don't include a body, but that makes the creation ofS3ResponseErrorworse because there are no details. I faked in the common cases, but if anyone was parsing this before, they're going to have to alter their code. I've documented how they should approach that.Review please? @jamesls @danielgtaylor @garnaat @mfschwartz