Boris Kolpackov wrote:

Noel Yap <[EMAIL PROTECTED]> writes:
How is "shortest stem" defined?

Literally. Here is the algorithm I have in mind:


1. For each otherwise equally applicable implicit rule extract the
   stem (i.e. what % matched).

2. Find the shortest stem and use corresponding rule.

This would mean that the most "specialized" or "restricted" rule will be used which is what you usually want. For example:

/foo/%:

/foo/bar/%:

When you try to match "/foo/bar/baz" against those rules the second
one is chosen.

I just found this in my Inbox so I apologize for the late response.


The above algorithm sounds really good. May I add that, if two patterns somehow have equal weighting with the above algorithm, that the one with the smallest stem-length-to-pattern-length-ratio should take precedence?

Even without this addition, though, I'm guessing your proposal will catch 99+% of the cases.

Thanks,
Noel



_______________________________________________
Help-make mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/help-make

Reply via email to