>> 1) why is \r not considered whitespace?
> Perhaps because \r is really a arbitrary control character (Python
> uses UNIX' convention that \n is the line separator). Or perhaps it's
> an oversight. What set of characters does ANSI C define to be
> whitespace? Should \f be added too?
>> 2) even if I fix whitespace in string.py, it is not being picked up
>> because it has been hardcoded in module strop to be space, \t and
>> \n. (string.py imports strop.)
> Same reason, and I thought it wasn't ever going to matter. Do you
> think it is sufficient if I fix this by adding \r (and \f) to the
> hardcoded list, or do you think we'll need a more general routine?
> A possibility, if people think this is worth it, would be to somehow
> find out which characters the C library's "isspace()" function
> considers whitespace and use this definition throughout.
> Feedback, please!
In C source files, ANSI C defines whitespace to include tab (0x09),
newline (0x0a), vertical tab (0x0b), formfeed (0x0c), carriage return
(0x0d), space (0x20), and comments. I don't have doc on the ANSI
C library routines.
The SunOS (non-ANSI) isspace() says all of the above (except comments)
are whitespace.
Jon Vander Hill
jon@cas.org