Security decorator should log client IP address

[danballance]

Fixes #410

Changes proposed in this pull request:

• trusted_ips list optionally passed in at application start. If passed then we'll only trust X-Forwarded-For if remote_addr is in the list of trusted IPs (i.e. known proxies under our control)
• "for client IP 'x.x.x.x" format string is added to log lines
• testfixtures added to setup.py in order to make use of LogCapture. I wouldn't normally recommend testing log lines (it does feel fragile) but in this case we need to be sure the correct IP is logged.

[coveralls]

Coverage remained the same at 100.0% when pulling 5f8a71b on danballance:log_client_ips into e9db7fe on zalando:master.

[coveralls]

Coverage remained the same at 100.0% when pulling 5f8a71b on danballance:log_client_ips into e9db7fe on zalando:master.

[rafaelcaricio]

Could you please rebase this PR? I am seeing a lot of changes related to the PR #414.

[rafaelcaricio]

Better than doing all this mocking is to use test_request_context. It is safer because we relay on how Flask actually works and not how the mocks are set up.

api_client.application.test_request_context('/').push()

Reference:

• http://flask.pocoo.org/docs/0.12/reqcontext/
• http://flask.pocoo.org/docs/0.12/api/#flask.Flask.test_request_context

[danballance]

Okay, no problem, I will review this code. A few things to clean up. Thanks for the feedback.

[rafaelcaricio]

@danballance Thank you for the PR. Would be nice if we could have some documentation on what it does and how this works. What do you think? Do not need to be long. Basically a summary of your comments from #410.

[rafaelcaricio]

@danballance Please review the changes made in this PR. It needs rebasing.

[danballance]

Hi,

I've just been looking for a gap here where I can look again at this. Tomorrow may be possible - certainly this week.

[danballance]

I am looking at this, but 1.1.7 appears to be a major re-factor and lots of things have broken. I still hope to resolve everything this week, but it's not going to be tomorrow now.

Will review again after refactor.

[rafaelcaricio]

@danballance Yes, sorry about that. It was an expected internal refactor to move the project forward (see #380).

[danballance]

No problem - it makes complete sense to decouple Flask. It's a big change though! I can see you guys are still changing quite a lot of code so I think I'll wait until the dust settles before I return to this :)

[rafaelcaricio]

@danballance Sure, no problem. Yeah, now we are fixing some things since the decoupling started.

[jmcs]

I'm closing this PR since it wasn't updated in more than 1 year. If you are still interested in adding this change please reopen it.