Documentation

It can sometimes be handy to have a canonical representation of a given patch. We achieve this by defining a canonical form for each patch type, and a function canonize which takes a patch and puts it into canonical form. This routine is used by the diff function to create an optimal patch (based on an LCS algorithm) from a simple hunk describing the old and new version of a file.

tryToShrink ps simplifies ps by getting rid of self-cancellations or coalescing patches

Question (Eric Kow): what properties should this have? For example, the prim1 implementation only gets rid of the first self-cancellation it finds (as far as I can tell). Is that OK? Can we try harder?

sortCoalesceFL ps coalesces as many patches in ps as possible, sorting the results in some standard order.

canonizeFL ps puts a sequence of primitive patches into canonical form. Even if the patches are just hunk patches, this is not necessarily the same set of results as you would get if you applied the sequence to a specific tree and recalculated a diff.

Note that this process does not preserve the commutation behaviour of the patches and is therefore not appropriate for use when working with already recorded patches (unless doing amend-record or the like).

tryShrinkingInverse ps deletes the first subsequence of primitive patches that is followed by the inverse subsequence, if one exists. If not, it returns Nothing