Sat, 1 Jul 1995 16:55:53 GMT

Related articles |
---|

LL(1) Grammar for Expressions with Assignment bergmann@elvis.rowan.edu (1995-06-27) |

Re: LL(1) Grammar for Expressions with Assignment jos@and.nl (1995-06-30) |

Re: LL(1) Grammar for Expressions with Assignment parrt@lonewolf.parr-research.com (1995-07-01) |

Newsgroups: | comp.compilers |

From: | parrt@lonewolf.parr-research.com (Terence John Parr) |

Keywords: | LL(1), parse |

Organization: | Parr Research Corporation |

References: | 95-06-075 |

Date: | Sat, 1 Jul 1995 16:55:53 GMT |

Seth Bergmann (bergmann@elvis.rowan.edu) wrote:

*: I've been trying to find an LL(1) grammar for the language of*

*: expressions such as:*

...

*: but excluded would be things like:*

*: 3+a=4 and 2 = a*

*: analysis. I suspect this language of expressions is not LL(1).*

I believe that it is non-LL(k) for any finite k; for example,

I can write this

(((((((((((3+a))))))))))) = 4

You would never be able to see past the '('s with finite lookahead.

I believe a semantic analysis of the resulting AST would be the best

route.

I suspect that any grammar that could describe this lvalue problem

(purely) syntactically would be very very large.

Regards,

Terence

--

Post a followup to this message

Return to the
comp.compilers page.

Search the
comp.compilers archives again.