Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Sorry this commit is kind of monolithic! It's difficult to
separate out any of the changes since they're all related to
track sections' basic functionality.
|
|
|
|
|
|
An existing test for withAlbum (which is semi-newly based on
withUniqueReferencingThing, and thus withReverseReferenceList)
indicates that if we just can't *find* the album in a provided
data array, that's not an exit condition. This commit adapts
withReverseReferencingList to respect that when the data list
is empty.
(Treating an empty array as the same thing as an array that
isn't provided at all - the existing behavior - is really just
a bug!)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Instead of the flipwise direction.
|
|
|
|
|
|
|
|
Lets us build up symbolToStable simultaneously, which we use when
computing unstableSortIndices, officially making it "not magic".
|
|
It's still still magic though.
|
|
Still magic.
|
|
It's still magic though.
|
|
This code is terrifying.
|
|
This reverts commit ac7ceacc2a9c0cdf78e7c9f060d46c6d83ddb7b3.
Because the original behavior is mayhaps correct!?
|
|
(instead of nonsense)
repeated runs indicate this is probably deterministic, so still
more to investigate for withThingsSortedAlphabetically
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|