Re: utf8 stristr

[Owen Taylor <otaylor redhat com>]

On Mon, 2003-03-31 at 12:08, Michal Adamczak wrote:
> > strrstr() is not part of standard C, neither is strstr_len() or
> > strrstr_len(). However, your C library contains a perfectly good
> > strstr(), so there is no reason to tack a g_ in front of it.
>
> personally i think that they should appear just for consistency
> and there should appear replace and regex funcions as well

Regular expressions have been a long standing idea of something to add
to GLib, but we've had the problem that available Regex
implementations don't have good enough UTF-8 handling. Apparently
recent PCRE is distinctly better in that regard.

> is sizeof(char) always equals sizeof(utf8_char)?
> maybe i am wrong but i thought that soem characters beyond first chars 
> table
> are written with two bytes
>
> if not there's no need for strstr and other at all
> but my concern was that the size of one chracter can be 1 or 2 bytes

The strstr() operation doesn't need any UTF-8 support. The way
a string is encoded into UTF-8 is the same no-matter where
it appears in the string.

[ Though normalization may in some cases be necessary if you
  actually are doing searching based on user input -- see
  g_utf8_normalize() documentation ]

The g_strrstr() and so forth functions aren't there for Unicode
handling , but rather because they are useful in some cases.

Regards,
                                   Owen