Hello all, 

New on the list, here's my first posting. (No doubt there will be more...
:-)) 

I was looking for a way to make entity substituion in texts more efficient
and started digging. As a result of my digging, I stumbled over a number of
errors in the entites[] array in swish.h which will lead to incorrect or
incomplete substitution of character entity references:

- code point  204 (Igrave) has an incorrect translation as "E"; this must
be "I" ;

- code point 182 is incorrectly named "mu"; this must be "para" (I have no
idea where this one came from);

- code point 168 has an incorrect name (die): must be uml;
- code point 175 has an incorrect name (macron): must be macr;
- code point 176 has an incorrect name (degree): must be deg;
- code point 184 has an incorrect name (Cedilla): must be cedil;
I do know where these last four came from: they were the names proposed in
the (never-to-be) HTML+! (Time-wise, that fits with the origin of the code
in 1994-1995, of course). HTML 2.0 did not define code points 160 ... 197,
as of  HTML 3.2 the whole set 160 ... 255 is defined, including the
(standard) names I indicated above.

The first error would lead to an incorrect translation, of the others
"para" and "uml" would not be recognized in a document and left
unconverted; the last three will probably still be translated correctly
(though I didn't test this) since the code seems to match by the  start of
a string.

Quite apart from efficiency -these errors are of course easy to fix.

Cheers,


Marjolein Katsma      webmaster@javawoman.com
Java Woman - http://javawoman.com/