Make behaviour of cursor.count() on client reflect server

[carlevans719]

This PR changes the default behaviour of cursor.count() on the client. Previously, .count() on the client would return the number of matching documents taking into account skip and limit. Now, it will ignore those options unless it is passed true when invoked. (i.e. cursor.count(true)). This is consistent with the way it works on the server.

Issue: #1201
Introduction of feature on server-side: 258b14a & bce4df9
Previous PR (incomplete): #5023
Test app: meteor-issue-1201

[hwillson]

Hi @carlevans719 - thanks for submitting this PR! Everything here looks great! I made one small nitpick suggestion about variable naming, but functionality wise we're really close here. The approach you've implemented that sets applySkipLimit to false by default on the client side (to line up with what's happening on the server side) seems like the way to go, but this will break backwards compatibility for some people (especially those using Minimongo as a client side only cache who are expecting the current applySkipLimit true functionality). We'll review this further in our issue triage meeting today, and get back to you shortly with an update. Thanks!

[hwillson]

I think it might make sense to just carry the applySkipLimit variable naming throughout, instead of switching (and inverting) it to ignoreSkipLimit. This would help simplify the code a bit and makes the logic easier to follow. Also, since applySkipLimit is a Mongo count() supported option, most people reviewing the code will already be familiar with it.

[carlevans719]

Agreed! I'll get it updated

[carlevans719]

Hey @hwillson. I've updated ignoreSkipLimit to applySkipLimit in _getRawObjects.
I've left the default behaviour of _getRawObjects as it was (applying skip / limit) because .fetch(), .forEach() etc... expect it to do this.

Let me know if you have any other suggestions & feel free to point me towards changelogs / documentation if it needs updating to highlight this breaking change

[hwillson]

Thanks for making these changes @carlevans719! We talked about this a bit more, and have decided that maintaining backwards compatibility in this case is (unfortunately) more important than correctness. By giving people an option to change things we're at least providing a workaround to the broken behavior, and we can re-visit this in the future (potentially alongside other related breaking changes). Would you be able to flip things around so that applySkipLimit is true by default on the client side? This won't line up with the server side, but we'll make sure this is documented in the History.md file. Speaking of which, if you're able to add a small entry to the History.md explaining this change, that would be great!

[carlevans719]

Makes sense @hwillson! Is the entry in History.md okay to go under v1.5.2.2 or a different/new heading?

[hwillson]

If you could add a new heading called v.NEXT and add it under that, that would be great!

[carlevans719]

@hwillson all sorted. How does that look?

[hwillson]

Looks great @carlevans719 - thanks! As soon as the tests finish running, we should be all set here. LGTM!