Remove 'Bucket.__iter__' and 'Bucket.__contains__'.

[tseaver]

Both are "sugar" over other methods, but do not allow passing an explicit connection.

In addition, Bucket.__iter__ has scaling issues, which cannot be mitigated using the filtering parameters accepted by Bucket.list_blobs.

Moved to a new PR from #876.

[dhermes]

Big time test failure here. Do you know why?

[tseaver]

I ran the tests in a different shell, which I closed accidentally before committing. :(

I will rebase / fix after merging #876.

[tseaver]

@dhermes rebased.

[dhermes]

So the reason I asked you to peel this off from #876 was really so we could discuss if this should happen.

I have never really liked the methods, but some may like them. So let's discuss?

[tseaver]

To clarify what we are discussiing: first, we are removing:

if blob_name in bucket:
   do_something()

which is a syntactic alias for:

blob = Blob(blob_name, bucket)
if blob.exists():
    do_something()

Note that do_something() likely needs to construct the Blob anyway, which makes the second a lot more natural. The second form also allows us to override the connection used by the exists() call, if desired.

Second, we are removing:

for blob in bucket:
    do_something(blob)

which is a syntactic alias for:

for blob in bucket.list_blobs():
    do_something(blob)

Here, the list_blobs() form allows us to override the connection used, if desired, as well as providing filtering parameters to minimize the set of returned blobs: the latter could be crucial for dealing with very large buckets.

[dhermes]

@jgeewax Are you OK with us removing this?

I am on-board with this change.

[jgeewax]

I'm down with removing the methods, but we need to update the docs in the same PR, right? (https://googlecloudplatform.github.io/gcloud-python/latest/storage-buckets.html refers to for blob in bucket and print 'blob' in bucket.)

[dhermes]

+1

[tseaver]

@jgeewax 3a3edb4 removes the docstring examples which referred to the sugar.