Re: Binarycloud and UTF

Manuel Holtgrewe wrote:
> Though mbstrings provides more multi byte safe functions it is not very 
> common and if binarycloud depends on it we set the entry level pretty 
> high (own server).

We can have a branch for multilanguage, and a third one for unicode 
multilanguage...

>  - mb_ereg_*  Ereg seems not to be multi byte safe. PREG is not UTF safe 
> seemingly, too.

False, it's since a long time (pcre version 3 I read)

> However, there seems to be a "/u" flag for unicode 
> support. We need to investigaget there [2].

 From manual:
u (PCRE_UTF8)
     This modifier turns on additional functionality of PCRE that is 
incompatible with Perl. Pattern strings are treated as UTF-8. This 
modifier is available from PHP 4.1.0 or greater on Unix and from PHP 
4.2.3 on win32.

So u flag is for Pattern only, and available.

>  - mb_parse  We need no global parsing from GET/POST/COOKIE, I guess
>  - mb_split  Splitting by regex can be done with preg and "/u" flag
>  - mb_strcut There is iconv_substr that does the same
>  - mb_strimwidth The same can be done with iconv_substr and iconv_strlen
>  - mb_strlen, mb_strpos, mb_strrpos, mb_substr There are equivalent 
> methods with iconv
>  - mb_strtolower, mb_strtoupper There is no equivalent, but I do not 
> know if we need this for base code. Base configuration names, for 
> example, are in English always and that fits into 7 bytes.
>  - mb_strwidth No equivalent - required in base code?
>  - mb_substitute_character, mb_substr_count No direct equivalent - 
> required in base code?



> I think we should evaluate all this and then look on binarycloud's base 
> code (vortex packages can use mbstring IMO or we provide two versions) 
> and get sure that unicode works. In many cases we will not need unicode 
> safeness since the string that is operated are event ids and those come 
> from configuration (See next point).

And what if Indian or Chinese people want to have their events ids or 
strings id not set in english ?

> C.) Most configuration values should be English and use 7bit only. This 
> means we do not need multi byte safe methods in many places.

No. If we make a multibyte system, it must be safe, not guess it will 
not receive any mb cahr :(

> D.) We use PEAR/I18Nv2 to transcode from UTF-8 to the charset requested 
> by the user (or set in code). See [3] for the documentation of this 
> feature of I18Nv2

I'll read this pear file this night.

-- 
Jean-Christophe Michel
_______________________________________________
dev mailing list
dev-PnctHDZWAvB/Cz2I37pSEPZ4XP/[email protected]
http://lists.binarycloud.com/mailman/listinfo/dev