Collect should not attempt to redownload media file if it has the same hash as the last download

[seadowg]

Collect currently redownloads a media file if the hash in the response does not match the calculated MD5 hash of the file it already has. This doesn't quite match the spec:

If the hash value identified in the manifest differs from the hash value for a previously-downloaded media file, then the file should be re-fetched from the server.

Really, Collect should compare the hash in the manifest response against the manifest response form the previous time it downloaded the file rather than use the file itself as the spec does not explicitly state that the hash (with an md5 prefix) is originated from the file contents (it just happens to be the case most of the time). This is specifically a problem for Central's entity list media files which return a hash that will reflect whether the entity list has changed or not, but is not derived from the CSV file.

The actual impact of this problem is lowered by the fact that we use HTTP caching and the redownload of the same file will often be caught there.

Acceptance (using Central)

• Given I'm using a project with an entity follow up form
  And I've previously downloaded the form
  And I've cleared the app's cache (in app settings)
  When I update forms
  Then the entity list is not redownloaded

• Given I'm using a project with an entity follow up form
  And I've previously downloaded the form
  And I've cleared the app's cache (in app settings)
  When the entity list is updated
  And I update forms
  Then the entity list is redownloaded

• Given I'm using a project with a form that has a media file
  And I copy the form and media file onto device instead of downloading them
  When I update the form
  Then the media file is redownloaded

• Given I'm using a project with a form that has a media file
  And I've downloaded the form
  And I've cleared the app's cache (in app settings)
  When I update the form
  Then the media file is not redownloaded