Codecov Report

❗ No coverage uploaded for pull request base (develop@b79b599). Click here to learn what that means.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             develop     #708   +/-   ##
==========================================
  Coverage           ?   95.33%           
==========================================
  Files              ?       19           
  Lines              ?     3665           
  Branches           ?        0           
==========================================
  Hits               ?     3494           
  Misses             ?      171           
  Partials           ?        0           

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b79b599...510d09f. Read the comment docs.

Read concern is supported since 0.19.1 #628

@pcorpet would it be possible for you to find time to review and merge this change?

Hi @m-aciek and thanks for your contribution. If we're going to go that way, please make it possible to use a with_options (both in collection and database).

However two things to do before removing those checks:

• what should change in the behavior with a different read concern? From a quick glimpse to the doc, it seems that with a in-memory implementation with only one replica, all the read concerns will have the same effect. Please confirm.
• let's still make sure that setting a bad type as a read_concern param in any of those functions would return the same error that pymongo would raise.

Thanks for looking into this.

Read concern works already in with_options of collection.

I think read concern shouldn't make a difference when using one replica, which we are mocking.

I wonder if I should bring a duck typing like in collection.with_options() (with using _WITH_OPTIONS_KWARGS) or just checking the type (first trying to import it from PyMongo, then from mongomock) is enough?

You're right, I think we're fine like this.

Pease cover the new lines with some new tests, and I'll merge it.

I've added missing test. I've added read concern type check support in Collection constructor (the constructor may be called in Database().get_collection() method). Also removed read concern type check from Collection().with_options() as it is now handled in the constructor.

Pylint failure is not connected with changes here. I've added a new issue to track that: #709.

Is there a way to retrigger Codecov check?

Added test for database.read_concern attribute, added setting default ReadConcern() if no value provided.

Thanks for all the cleanups and tests.