At 22:19 +0000 1-01-06, Sebastian Rahtz wrote:
>at a minimum, you have to get a flex patched to handle Unicode; it
>exists, but its not standard.

Please give a reference (URL) to Unicode Flex.

-- I am myself thinking of patching up Bison so that it can do Unicode
parsing. The drawback with this approach is that it is more complicated to
write regular words, but one anyhow has access to the more general LR(1)
grammars that Bison can handle.

From the practical point of view, developing Flex and Bison synched
together proves difficult, because they belong to different organizations:
Bison is maintained by GNU with a public development policy; Flex by
somebody else, and I have not heard of any public development of Flex for
the last couple years.

>you *can* write an XML parser in flex,
>as Richard Goerwitz has demonstrated, but its not trivial.

I would not use Flex for doing the actual parsing, only to recognize
lexemes. Flex just recognizes regular words which can be turned on/off by
some context switches (called "start-conditions").These Flex context
switches are useful in connection with Bison, becuase one can use them to
handle language context dependencies.

  Hans Aberg