for the lexer side, one for the classical Lexing.lexbuf, one for
ULB.unicode_lexbuf. Depending on which is opened, Pxp_reader fills
one of these buffer implementations.
	There is also support for ulex in lex.src, and there is a
linker module for ulex.
	ulex passes the "reader" and "canonxml" regression tests.
	Regression tests use now a toploop for the PXP core. Set the
LEXER variable to select on of the toploops.


git-svn-id: https://godirepo.camlcity.org/svn/lib-pxp/trunk@675 dbe99aee-44db-0310-b2b3-d33182c8eb97