Extract links from HTML5 media tags

[tsdko]

Port of mirror/wget@6a2d67b.

I remembered to:

• Update or add unit tests if needed
• Update or add documentation/comments if needed
• Made sure stray files or whitespace didn't get committed
• If significant changes, branch from develop and set to merge into develop
• Read the guidelines for contributing

Changes:

• Adds poster to the list of recognized link attributes and video, audio and source to the list of recognized tags to extract links from.

A few days after making these changes I noticed some of these are actually extracted without being defined there explicitly, not sure if adding them all is a problem. My main reason for making the PR was wpull not downloading files linked in poster attributes.

[JustAnotherArchivist]

Thanks, looks good to me. The test suite is currently failing miserably though for a variety of reasons, and we need to fix that first (see comments on #393).

[TheTechRobo]

LGTM!