Open Fuco1 opened 8 years ago
This change breaks legit tests (yay testing!). I will think about refactoring the code around this. It's pretty hard for me to parse the code 1+ years on though.
Well it probably breaks the examples which tested the wrong case originally. It seems only those where the property is violated fail.
This is the valid test I'm talking about - https://github.com/lewang/flx/blob/master/tests/flx-test.el#L296 .
I haven't forgotten about this. There are some other optimizations coming in and I'll revisit this after that PR is merged.
Thanks for the update.
The variable
word-count
is only determined onflx-word-p
which doesn't break on camelCase changes. OTOH,words-length
properly reflects the number of words (based onflx-boundary-p
.Further, we might probably want to remove the
word-count
variable alltogether, it doesn't serve any purpose as it just mirrorswords-length
but in an inconsistent way.