Re: More on that ^@^@ bug (Was: DOCBOOK: db150.zip bug?)

[Norman Walsh <ndw at nwalsh dot com>]

/ Simos Xenitellis <S.Xenitellis@rhbnc.ac.uk> was heard to say:
| The code that looks suspicious is (i think):
| 
| ======
|      70 (define (dsssl-country-code #!optional (node (current-node)))
|      71   (let* ((lang     ($lang$))
|      72          (ctrycode (if (> (string-index lang "_") 0)
|      73                        (substring lang
|      74                                   (+ (string-index lang "_") 1)
|      75                                   (string-length lang))
|      76                        #f)))
|      77     (if ctrycode
|      78         (string->symbol (case-fold-up ctrycode))
|      79         #f)))
| ======
[...]
| Eventually, the value of line 79 (#f) is the output of the function.
| 
| Question: Is the Country Code derived from the Language Code?
| 	  I don't think so.
| 
| Solution: If we accept that lang-code=country-code, we should then
| 	  change line 76 to 'lang)))'.

We read the DSSSL spec differently. The country characteristic is
defined as:

  country: is #f or a symbol specifying the ISO 3166 country
  code in upper-case. This affects line composition in a
  system-dependent way. The initial value is #f.

I interpret this to mean that country should be #f unless the
language has a country-specific localization. In all cases except
Brazilian Portuguese, the value should be #f.

What appears to be a bug to me is that an explicit value of #f
causes JadeTeX to barf.

                                        Cheers,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com>      | He would like to start from
http://www.oasis-open.org/docbook/ | scratch. Where is scratch?--Elias
Member, DocBook Editorial Board    | Canetti