Re: Jade charsets question.

[James Clark <jjc@xxxxxxxxxx>]

You need to do

  jade -t fot -d test.dsl koi8r.dcl "<osfile bctf=identity>test.sgml"

If you set SP_CHARSET_FIXED=1, then you will get the same behaviour from
nsgmls as from jade.

Valeriy E. Ushakov wrote:
> 
> Some time ago, while learning SGML, I have written an SGML declaration
> for SGML documents that allows the use of russian alphabet (in koi8-r
> encoding, rfc1489) both in document content and markup (OminMark
> whitepaper was very helpful).  This declaration uses (in both CHARSET
> and SYNTAX)
> 
>     BASESET  "ISO 646-1983//CHARSET
>               International Reference Version (IRV)//ESC 2/5 4/0"
> 
> and then in DESCSET it maps chars from koi8-r to "a character number
> in the universal character set" (to quote the SP documentation).
> 
>                  -- ....... --
>                  192  1 1102 -- CYRILLIC SMALL LETTER YU --
>                  193  1 1072 -- CYRILLIC SMALL LETTER A --
>                  -- etc ... --
> 
> Running
> 
>     $ nsgmls koi8r.dcl test.sgml
> 
> parses the document without any errors.
> 
> Now I try to write a DSSSL style sheet for a document that uses my
> declaration and use both russian element names and russian content.
> 
> When I run jade
> 
>     $ jade -t fot -d test.dsl koi8r.dcl test.sgml
> 
> Jade emits an error messages like:
> 
>     jade:test.sgml:4:10:E: non SGML character number 225
> 
> for every russian letter in my document content.  It gives no errors
> for russian letters in element names in my document and corresponding
> element rules in my style sheet.  And it produces correct fot dump.
> 
> I reread clause 7.1 of DSSSL standard and Jade/SP documentation
> several times. I tried specifying baseset-encoding and standard-chars
> in my style sheet but that was of no avail.


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist