bug-gnu-libiconv
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [bug-gnu-libiconv] iconv and symbol font


From: Bruno Haible
Subject: Re: [bug-gnu-libiconv] iconv and symbol font
Date: Thu, 28 Aug 2008 01:20:38 +0200
User-agent: KMail/1.5.4

Dave Davey wrote:
> We have been working on the GNU unrtf package to support all codepages
> supported by iconv rather that just having specific codepages hard
> coded within unrtf.  This is working well, except for the fact that
> many unrtf files use the symbol font, and conversion from it's encoding
> to unicode does not appear to be supported by iconv.
> 
> Would it be possible to have this table added?  (We have a file in the
> iconv format which has the unicodes.)  Or alternatively is there a
> way to specify an encoding file additional to the ones in the iconv directory?

Unfortunately, this is not possible for two reasons:

  1) You would want portability. That is, you probably don't want to
     extend only GNU libiconv's iconv_open() function, but also the one
     from glibc, and the one from other major iconv() implementations (Solaris
     in particular). Extensions from a single vendor won't help you.
     For discussing such extensions in general, you'd better ask on the
     Austin Group's mailing list; this is the group which is responsible
     for maintaining the POSIX standard.

  2) Encodings from iconv are text encodings. Font encodings are a different
     thing. For example, JIS X 0208 is a font encoding, but it is not supported
     by iconv (in GNU libiconv) since it is not a text encoding. EUC-JP on the
     other hand is a text encoding but not a font encodings. There is a strong
     relation between a JIS X 0208 converter and an EUC-JP converter, but the
     fact remains that they are not the same.

     The difference between text encodings and font encodings is that
     text encodings support all possible characters and also control
     characters, whereas font encodings are often limited to a single script
     and never contain control characters.

Bruno





reply via email to

[Prev in Thread] Current Thread [Next in Thread]