Re: LYNX-DEV Character set support 9

[Klaus Weide]

On Sat, 12 Jul 1997, Michael Sokolov wrote:

>    Now that you are back, I can repeat my question about supporting
> multiple MIME names for a character set. Here is the relevant posting
> repeated:
>    
>    I have downloaded the most recent list of MIME character set names
> registered with IANA. What I have seen there is that there is a pressing
> need for Lynx to support multiple MIME names for a single character set.

Michael, the IANA list of registered charsets alone certainly doesn't
establish any "pressing need".  Please cut down the rhetoric a bit...

Of all those names and alternative names listed there, lynx will,
hopefully!, only encounter a small number in real life.  It is not
necessary and only adds to confusion to implement names which never
get used.  Note the "(preferred MIME name)" designations which have
recently appeared in the registry.

> The official list defines several names for almost every character set. For
> example, for IBM code pages, cp<n> and ibm<n>, where <n> is a number, are
> equally valid.

Equally valid, and equally undesirable ;) (for general use on the
Internet, as opposed to internal use by e.g. Lynx)
Anyway, I had to make a choice between the two, so I chose the cp<n>
variant for those tables packaged with the development code.
If you argue that both occur in real life and are being sent by servers
(as opposed to just: they are in the official list), okay.  There is 
already some special-case recognition of aliases (some of them just hacks)
in function UCGetLYhndl_byMIME of UCdomap.c.  Recognition of "ibm<n>" as
"cp<n>" could be added there.

>    I think that the *.tbl file format should be changed to allow multiple
> MIME names to be defined. Or maybe it already does? (If yes, how?)
>    Please comment on my suggestion.

The general mechanism you propose might be nice to have, but if you really
want it, you may have to write it yourself...


    Klaus

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;