[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

Re: [xsl] Sorting substitution instructions by max. length of matches

---

---

David Carlisle wrote:

I'd probably use analyze-string rather than using an explicit recursion to do the lookup.

Thank you for this solution which is really elegant. Once I have seen
this, I come to think that I've exposed a rather trivial problem... :-)

(I think in both cases they may fail to find the longest match possible, as earlier matches may obscure a possible later longer match)

You are definitely right that I have to face this issue which I admittedly failed to notice, and that this problem is common to both approaches.

Can you think of an example of an input text and a regex with matches that would overlap in this way? (Each regex's matches will be checked in a pass of its own, so it must be matches of the same regex.) I tried but could not find an example within some time.

Yves

---