[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: [PATCH] Add select encoding in Control Center (for CVS HEAD)
From:       Ilya Konstantinov <kde-core-devel () future ! shiny ! co ! il>
Date:       2001-09-28 2:57:41
[Download RAW message or body]

On Thu, Sep 27, 2001 at 02:08:26PM -0700, Neil Stevens wrote:
> On Thursday September 27, 2001 01:21, David Faure wrote:
> > > UTF-16 should be supported in Qt 3 as well as ISO-8859-6.
> >
> > But we decided against using UTF-16 and other multi-byte encodings
> > as the global encoding, since many programs (KDE and non-KDE)
> > don't support this at all.
> > UTF-16 as a font charset is fine, but as the encoding for saving files,
> > is too dangerous. It breaks much, far too much.
> 
> But if it's not provided as an option, won't that mean that it will never 
> be even possible to make it safe?

It's not very important to make it safe. Text and HTML files written in
UTF-16 usually begin with two bytes FF FE or FE FE, denoting the
endianess, which help QTextStream device determine what it deals with.
Thus, QTextStream can recognize UTF-16 data even if the locale encoding
is set otherwise. This is important with HTML files (we need to know
the page's encoding to process the META tag which denotes the encoding
- a sort of a chicken and egg problem :). 

Other than that, I don't see much use for it. What about the Far-Asian
encodings? Any of them not null / slash safe (slash is important for
Unix file name restrictions) and still in widespread use?

[prev in list] [next in list] [prev in thread] [next in thread]