Page MenuHomePhabricator

copies: avoid unwanted side effect from one branch to another
ClosedPublic

Authored by Alphare on Nov 27 2020, 11:11 AM.

Details

Summary

Without this copy, change in a one descendant branch (With "remove" change
only) could affect computation on another descendant branches.

This was not caugh by the test because the test graph are "too simple". I
started writing more test in that regards, but I a submitting this changes
earlier because I want to get more code landed to allow other optimisation work
to happens.

Diff Detail

Repository
rHG Mercurial
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.

Event Timeline

Alphare created this revision.Nov 27 2020, 11:11 AM
marmoute updated this revision to Diff 23814.Nov 29 2020, 6:13 AM
pulkit added a subscriber: pulkit.Nov 29 2020, 2:12 PM

I started writing more test in that regards, but I a submitting this changes earlier because I want to get more code landed to allow other optimisation work to happens.

Shall we expect tests for this future in series?

I started writing more test in that regards, but I a submitting this changes earlier because I want to get more code landed to allow other optimisation work to happens.

Shall we expect tests for this future in series?

Yep, I have a list of new cases to get tested that I'll implement and submit later. I just wanted the large code movement to be out of the way first to help other to start working on performance. I also plan to get some rust unnittest for some of the lower lever bits, but mu Rust foo was a bit too low at the time of writing this. Now that Raphaël Gomès is no longer in vacation and Simon Sapin is around, I'll write follow up about that.

pulkit accepted this revision.Nov 30 2020, 11:39 AM
This revision is now accepted and ready to land.Nov 30 2020, 11:39 AM