Swarming to learn about Isolate server

A few things are expected:

Part A:
- TaskProperties has new properties:
  - isolated_compression_algo
  - isolated_hash
  - isolated_hashing_algo
  - isolated_namespace
  - isolated_server
- task_runner.py knows how to run an isolated files, via issue 173, in addition 
to running normal tasks.
- swarming.py trigger pass this new information instead of archiving repeatedly 
run_isolated.zip (!)
- Web UI /user/task/<id> prints out the command to download files.
- Web UI /user/task/<id> links to /browse page on isolate server.

Part B:
- During task deduplication, Swarming server triggers a taskqueue that will 
"warm" all the files referenced by the isolated files on the isolate server. 
This ensures that a deduped task is not referring to files not present anymore 
on the isolate server. See issue chromium:445293.
- Swarming server knows about isolate server default expiration dates so it 
doesn't even try for older tasks.

Original issue reported on code.google.com by maruel@chromium.org on 9 Jan 2015 at 8:08

Blocked on: #173

madecoste / swarming

Swarming to learn about Isolate server #197