Code review comment for lp:~frankban/juju-core/get-charm-api

On second though and after discussion with the team on today's standup,
some changes are needed before this can land. Please ping me to take a
look again when ready.

Instead of downloading AND extracting the charm, we got a general
agreement you should cache the charm archive, and then use something
like charm.zipOpen to read a single file from it, giving a path. You can
also simplify file path related operations, because the zip reader
allows you to range over all files, including their paths and do simple
matching urlFileArg == zipFile.Name(). Look into findArchiveRootDir and
fixPath inside the charms handler in apiserver.

https://codereview.appspot.com/67750045/