From: ts <decoux@...>
Date: 2008-01-16T00:55:31+09:00
Subject: Re: Fwd: 1.9 significantly slower than 1.8 on Mac

>>>>> "D" == Derek Chesterfield <groups@dezuk.com> writes:

D> So unless there is some internal optimisation that occurs when the  
D> block is removed, it looks like each_line is the problem, not Regexp???

 Well, some part of #each_line for 1.8.6

    for (s = p, p += rslen; p < pend; p++) {
        if (rslen == 0 && *p == '\n') {
            if (*++p != '\n') continue;
            while (*p == '\n') p++;
        }
        

 easy : increment p and test 

 the same for 1.9

    while (p < pend) {
        int c = rb_enc_codepoint(p, pend, enc);
        int n = rb_enc_codelen(c, enc);

        if (rslen == 0 && c == newline) {
            while (p < pend && rb_enc_codepoint(p, pend, enc) == newline) {
                p += n;
            }
            p -= n;
        }

 a little more complex :

  retrieve the code point
  retrieve its length
  etc,


Guy Decoux