main, swarm, http, prod, api, client: repair globally pinned content through CLI publisher

[mortelli]

This PR attempts to solve issue #2193, which is about manifests being garbage collected, causing publishers data to be inaccessible and (as a result) the content repair process to fail.

The alternative instead, is to always attempt content repair through a fixed publisher provided through a CLI flag, instead of the derived from the manifest.

This solution:

• rolls back most changes from #2175 (this code is put into the global-pinning-manifest-publisher branch should we need it later)
• adapts code from #2183 due to implementation similarities
  • said PR is now on hold since this solution excludes the publisher fallback implementation

Implementation details

• adds a Recovery Publisher flag on boot-up
  • without this flag set, no content repair will be attempted
  • the flag will be passed to the NewRecoveryHook func and used thereafter
• only this sole publisher will be use for all content repair processes (this can later be expanded to 1 feed per root hash as specified in #2193)
• functions queryRecoveryFeed, getFeedTopicAndUser, getFeedContent and publisherToAddress were added from #2183 to make the code in the prod package more readable
• all the code related to the publisher field in manifests is rolled back
  • prod tests are adapted as a consequence

closes #2193

[mortelli]

putting this on hold since related discussions are taking place

[mortelli]

Closed in favor of #2202.