Schmidt
Schmidt
Hi, thats good input, thanks! I absolutely agree that we need better documentation, especially for the keychain usage. I would happily review a PR :smile:
Yes, while I am generally in favor of more testing rather than less, I think we have sufficient test coverage here. The exception is also not new, it is already...
yes very true.
Oh I did and it's... mysterious. Have also restarted them multiple times, and its reproducible. I have no clue what it is yet. Could you also have a look?
`test.retry.js` reliably fails in some scenarios. I suspect that the test fails because they mock replication internals. I have [adjusted it](https://github.com/pouchdb/pouchdb/pull/8499/files#diff-aa4ef569ea92a5feccb10854696ae783ee268b5c2e1419f70abd67e9ee17d33e) so that they no longer depend on `_all_docs` but...
Update on the test failures: https://github.com/pouchdb/pouchdb/pull/8507 fixes a race condition when cancelling replications. The [commit is included](https://github.com/pouchdb/pouchdb/pull/8499/commits/42165eec072956c57aed8f03c200159e09750987) in this pull. This also fixes the failing test here.
also very nice for `_changes` and `_bulk_get`
You should use Array.pack and string.unpack "m" Greetings, Johannes
I meant couchrest could use Array.pack("m*") instead of using Base64.encode64 to obviate that version dependency. http://www.ruby-forum.com/topic/132076 Greetings Johannes