string_cmp_value missing UTF8 code

[Peter-Jacob]

As I look at non-strict comparisons, I found possibly an inconsistency in the string_cmp_value function:

static RX_INLINE int string_cmp_value(value *v1, value *v2) {
    return string_cmp(v1->string_value, v1->string_length,
                      v2->string_value, v2->string_length);

Shouldn't we distinguish between ASCII and UTF8?

My proposal would be:

#ifndef NUTF8
  static RX_INLINE int string_cmp_value(value *v1, value *v2) {
    return string_cmp(v1->string_value, v1->string_chars,
                      v2->string_value, v2->string_chars);
#else
   static RX_INLINE int string_cmp_value(value *v1, value *v2) {
    return string_cmp(v1->string_value, v1->string_length,
                      v2->string_value, v2->string_length);
#endif

looking further down in rxvmvars.h I found some also string_cmp_const. Is there a misunderstanding on my side?

[rvjansen]

This was the question that gave me the impression that there was work done on the non-strict comparisons. Is it resolved?

[Peter-Jacob]

you are right, we need the distinction. I am still baffled, that I should have done something in this area. I can't remember! If true, I'd doubt my state of mind.

[Peter-Jacob]

Code changed and committed: 6d015c9. The results are now:

random(1,999,8142) 653
random(1,999,8142) 653
random(1,999) 715
random(1,999) 185
random(1,999) 72
random(1,999) 742
random(1,999,8142) 653
random(1,999) 715
random(1,999) 185
random(1,999) 72
random(1,999) 742

[rvjansen]

looking great!