For issue 29550. Preliminary submission.
Comment From: philwebb
Thanks for the pull-request @polarbear567. When raising #29550 I was originally thinking we'd offer a way to do this programmatically. The properties are an interesting alternative that didn't occur to me and I'm now not sure which is best. I'm going to flag this for team discussion to see what other think.
Comment From: wilkinsona
I quite like the property-based approach. They do feel quite low-level but not so low that I think they ought to only be available programatically.
Comment From: polarbear567
Sure, I'll take a look at it when I'm free. I'll modify PR according to comments. I don't have any good ideas for testing. Do you have any suggestions? @snicoll
Comment From: snicoll
I don't have any good ideas for testing. Do you have any suggestions?
Yeah good point, I didn't look far enough. The mapping is hidden behind the predicate so it's hard to test this way. We can look ourselves as part of merging this, thanks for the update!
Comment From: snicoll
Thanks again @polarbear567.