Section 3.7.1 page 82
pure parser
---- code ----
int status;
yypstate *ps = yypstate_new ();
do {
status = yypush_parse (ps, yylex (), NULL);
} while (status == YYPUSH_MORE);
82 Bison 3.0
yypstate_delete (ps);
---- end code ----
It was reported in a bug report that the yypush_parse() calling sequence
description differs from the example. Even if this is the correct interface it
can't be used. So, just as below, what is the interface with yylex() and how
does information between yylex() and yypush_parse() occur? Since this is a
'pure' parser,, yylval and yylloc can not be global, so what is the correct
interface? yypstate can not be a vehicle since it is not based to yylex(), and
it can't be global because that violates independence requirement for
reentrancy.
impure parser (no doubt can not interface with a unicorn).
---- code ----
extern int yychar;
int status;
yypstate *ps = yypstate_new ();
do {
yychar = yylex ();
status = yypush_parse (ps);
} while (status == YYPUSH_MORE);
yypstate_delete (ps);
---- end code ----
And yylex() returns an int. What happened to yylval and yylloc? Is the
assumption that the example is complete "That’s it." because this example
doesn't use this data? Shouldn't data tunneling be mentioned to explain how
information generated in the lexor is now passed to the parser? It is true that
this is mentioned elsewhere, but so in yyychar?
_______________________________________________
help-bison@gnu.org https://lists.gnu.org/mailman/listinfo/help-bison
