[bitc-dev] GLR parsing hazards?
Jonathan S. Shapiro
shap at eros-os.com
Thu Feb 26 10:06:03 EST 2009
It seems that GLR parsing has become popular. It seems to me, naively
and ignorantly, that one effect of GLR parsing is to admit
progressively more ambiguous languages. Here is my question: are GLR
grammars ultimately unambiguous?
I have no problem with languages that require large finite lookahead,
and I have no problem with resolving reduce/reduce and shift/reduce
conflicts by hand-specified disambiguation (as some GLR grammars are
My concern, which comes from my ignorance, is that GLR grammars seem
to make unresolvable grammar ambiguities progressively harder to
discover at grammar compile time. At the end of the day, what I want
to be confident about is that every valid input program has a unique
Does GLR ensure this?
More information about the bitc-dev