Changes between Initial Version and Version 2 of Ticket #32940


Ignore:
Timestamp:
Jul 18, 2021, 4:01:14 PM (3 years ago)
Author:
Nick Pope
Comment:

The squash argument was added in d3f00bd5706b35961390d3814dd7e322ead3a9a3. It seems to have been unused since it was introduced. (I didn't find squash=False anywhere.)

The if self.connector == conn_type and data in self.children: line had the first half of the expression added recently in b81c7562fc33f50166d5120138d6398dc42b13c3. Interestingly this was actually fixing a regression in d3f00bd5706b35961390d3814dd7e322ead3a9a3 as prior to that commit the condition was the same! The bug had originally been addressed in c127f0117d75220e346625cd8a009fb6a317ece1.

It'd probably be a good idea to see if Simon can cast an eye over this just to make sure we're not going to reintroduce some problem.

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #32940

    • Property Owner changed from nobody to Keryn Knight
    • Property Status new → assigned
    • Property Triage Stage Unreviewed → Accepted

  • Ticket #32940 – Description

    initial v2  
    3131- `1493` calls to `Q.add` of which `0` have a `data in children` match.
    3232- `2` calls to `Node.add` of which `1` is a `data in children` match.
    33 - `210862` calls to `WhereNode.add` of which `2` are a `data in children match.
     33- `210862` calls to `WhereNode.add` of which `2` are a `data in children` match.
    3434
    3535But of all of those matches, the connector is never the same, so it never does that return early expected optimisation.
Back to Top