Re: Precedence Rules for '$' and '^'

> If an earlier rule starts with '$', does that mean that a subsequent
> rule starting with '^' has the '^' ignored?


A related question, perhaps the one you were asking, is whether once $
has matched, ^ can still match the same position in the case of an
empty line/string.


For example, ^$ clearly matches the empty string, but does $^ match it
too? What about ^$^, $^$, and ^^$$?


The answers you expect depend on whether you think about ^ and $ as
matching positions in a line or imaginary characters at the beginning
and end of the line. In the position model, you'd expect that for an
empty string, both match position 0, so /$^/ and all the others would
work just fine. In the imaginary character model, "abc" is really
"(imaginary-^)abc(imaginary-$)" and the empty string is really
"(imaginary-^)(imaginary-$)". Then you'd expect only /^$/ to match,
of the examples given above.


Most implementations use the position model, but regexp parsers
sometimes get in the way when trying to find out. In Perl, for
example, '' =~ '$$' is true but '' =~ /$$/ is not, because in the
latter, $$ is some interpolated variable. In most greps, ^ and $ are
only recognized as special at the beginning and end of the pattern, so
^$$ is a regexp matching a line containing just a dollar sign, while
in egrep, ^ and $ are always special, so ^$$ is usually a regexp
matching an empty line.


Russ

Return to the comp.compilers page.
Search the comp.compilers archives again.