I'd like to make a slight change to the is_digit, is_wordchar,
and other is_* ops. Currently calling these ops at the offset
following the last codepoint results in a "get_byte past the end of the buffer (1 of 1)" error,
it would be nicer if they simply returned false (0) at
this one position. (Going any further than that could generate
the error message.)
Sounds reasonable. We can change that along with the proposed C<is_cclass> implementation.
As it is now, I'm having to put in extra checks to watch for the end of the string when I'm just wanting to run along a string of digits. Of course, if we had an op to find the offset of the first non-digit/non-word/non-whitespace/etc. codepoint then I could use that. :-)
Well, then working on the general find and find_not classifying opcodes and implementation seems to be more effective, then ...
I'll be glad to write the patch for is_digit and friends if it's appropriate.
... reparing doomed code :-)
Pm
leo
