Support for internal endpoints in cloudfiles#229
Support for internal endpoints in cloudfiles#229jobelenus wants to merge 1 commit intoapache:trunkfrom
Conversation
There was a problem hiding this comment.
Passing use_internal_url=False shouldn't result in using it.
There was a problem hiding this comment.
Correct.
Please also explicitly declare use_internal_url argument in the method signature instead of "abusing" kwargs.
|
Adding a test case for this new functionality would also be good. |
|
A test case would be interesting, since the internal endpoint is only accessible from within the datacenter network. Is the test runner within that network? |
|
@jobelenus Sorry, I didn't actually meant doing the whole integration test. Sadly this would be too painful. I was just thinking of adding a simple unit test which tests that |
|
If I've correctly understood how you're doing the Driver/Connection kwargs, this testcase should meet your requirements. I may be wrong though |
|
@jobelenus Yeah, it's close enough :) |
|
Can you please fix lint (https://travis-ci.org/apache/libcloud/builds/17566981) and squash all commits? Once this is done, I'll go ahead and merge patch into trunk. |
|
Yea... that doesn't make sense from Travis. Over-indent then under-indent when I take away a tab? |
|
@jobelenus Do you happen to use tabs? If so, you should really use spaces. Looking at the file, you are missing one whitespace on line 110. |
|
I do use tabs. I don't see a missing whitespace? my line 110 matches line 147 (which is untouched) in behavior. |
|
It should match with a first argument, but currently it's matching with an open parenthesis. |
|
Ah. I didn't understand the Travis lint was expecting it to match to the first char of the first arg. Thank you. |
|
Np. Next time |
|
Also, please ping me when the commits are squashed. |
re-working `use_internal_url` in declaration, adding test case for endpoint variable lint white-space fix lint white-space fix match indent to the first character of first arg
|
Squashed up. Wow that was ugly. Shame on me for not being clean in the first place. |
|
Patch merged into trunk. This pull request can be closed now. Thanks! |
3 participants
Sorry for the ugliness of the commit history. I didn't do this cleanly from the start. The diff is clean, however.
Being able to use the internal network endpoint gives significant speed improvements (upwards of a whole second from my brief benchmarking) when your application server and cloudfiles are in the same region/datacenter.
RE: #186 (comment)