On Wed, 25 Sep 2013 22:03:33 +0200 Lionel Cons wrote:
> On 25 September 2013 21:48, Glenn Fowler <[email protected]> wrote:
> >
> > On Wed, 25 Sep 2013 20:47:24 +0200 Lionel Cons wrote:
> >> Here's a request from one of my teams:
> >> Can the AST built in locale C.UTF-8 be renamed to a different name,
> >> for example builtin_C.UTF-8 to avoid collisions with a system's native
> >> C.UTF-8 locale (apparently they have the problem at least on AIX)
> >> which may have similar but different properties?
> >
> > I wan't aware of any system with a C.UTF-8 locale
> > is there a url describing what it means on AIX?

> Unfortunately I'm just the messenger, but I forward the question. What
> I can understand is that C.UTF-8 is a C locale with UTF-8 multibyte
> characters, but the differences start with the character classes and
> other details.

a question for the list:

for the C.UTF-8 local ast overrides the [:alpha:] class and wcwidth() print 
width function
this is to provide a test locale that is consistent across platforms

it was dgk's and mine impression that [:alpha:] and wcwidth() are language 
neutral
and thus one set of lookup tables for unicode [:alpha:] and wcwidth() should 
work

is there wiggle room in the uncode spec for an implementation to change what
[:alpha:] and wcwidth() mean?

if the answer is no then we can ask why ast and AIX differ on [:alpha:] and 
wcwidth()
(and any other differences too) and fix ast if its missing something

if the answer is yes then we'll have to use a name different from C.UTF-8 as 
Lionel suggests

_______________________________________________
ast-developers mailing list
[email protected]
http://lists.research.att.com/mailman/listinfo/ast-developers

Reply via email to