Re: PHP Unicode strings impl proposal

On Aug 23, 2005, at 11:31 PM, Rolland Santimano wrote:

I meant that 'start'/'length' would be UChar counts and probably fall in the middle of a lead/trail sequence. But as you mentioned below, the iteration macros should take care of that.

start/length parameters should always be interpreted in codepoint context, so they should never fall in the middle of a surrogate pair.

That's true about the table - sorry, I wasn't clear the first time around. I'm thinking of an approach similar to that used in php_u_trim(), ie. iterating over both strings. Of course, its O(mn) rather than O(m+n).

ICU has its own u_strtok_r() function, but its limitation is that it does not support surrogate pairs (which we should).

Err, the ICU docs & code seem to suggest that they DO handle surrogates.

So, why not use u_strtok_r() for that, instead of messing with a table?

Ok, do we accept combining characters by themselves as delimiters ?

I don't see why not. -Andrei