[klibc] Question about islower() in ctype.h (ANSI compat headers test)

H. Peter Anvin hpa at zytor.com
Mon Mar 7 19:38:13 PST 2005

Arnd Bergmann wrote:
> On Maandag 07 März 2005 22:55, H. Peter Anvin wrote:
>>It doesn't seem worth it dropping the iso-8859-1 default for high bytes 
>>(it makes the table 128 bytes larger but the code for each invocation 
>>gets smaller since it doesn't have to bounds-check); I don't expect to 
>>have any kind of wide char/Unicode support any time soon, as much as I 
>>like Unicode.
> Just nitpicking, but is there any reason to use ISO-8859-1 instead of
> ISO-8859-15, which contains the € (Euro) sign? Unfortunately, they are
> not treated identical because e.g. islower() should be true for 'ž' (184),
> but not the respective character '¸' in ISO-8859-1. (Hmm, I wonder if this
> mail comes across correctly in UTF-8).

Yes, ISO-8859-1 is more widely used (because most people who were able 
to change charsets at all just switched straight to UTF-8) and 
toupper/tolower are simpler.

It's there mostly as "a bonus"; that really matters for klibc is that 
ASCII works.


More information about the klibc mailing list