[antlr-interest] Problems with the ANTLRworks interpreter.
rkevinburton at charter.net
rkevinburton at charter.net
Wed Jul 30 18:35:27 PDT 2008
I don't know if this will help but I have found that the output changes based on the line ending selected (Unix(LF) or MS(CR LF).
Kevin
---- "Peter C. Chapin" <pcc482719 at gmail.com> wrote:
> Hello!
>
> I'm using ANTLRworks 1.1.7 and I'm building a grammar for a C-like
> language. I've created some unit tests for my grammar... specifically a
> collection of files containing text that is intended to be syntactically
> correct. My test program loops over these files, instantiates a parser
> for each (one at a time), and tries to parse them all. I use various
> start symbols to exercise parts of the grammar incrementally in these
> tests (helps to isolate errors).
>
> So, given a file containing just
>
> A = B
>
> my test program successfully parses this using a start symbol of
> 'expresion.' This is fine. However, when I use the ANTLRworks
> interpreter and enter the text 'A = B' I get a 'NoViableAltException.' I
> noticed that the interpreter is ignoring my WHITSPACE rule so I removed
> the white space from the text (using 'A=B') but that didn't help. In
> fact, the interpreter produces a NoViableAltException even when I just
> use 'A' as input text. However, a bare identifier is a valid expression
> in my target language (and it parses fine when using the test program I
> described above).
>
> I'm think I am missing something fundamental about how the interpreter
> is supposed to work. I'm hoping someone can set my feet upon the path of
> enlightenment! I did check the on-line ANTLRworks user's guide, but the
> section on the interpreter doesn't shed much light on my question.
>
> P.S. I do have backtracking mode turned on right now in my grammar's
> option. Could that have something to do with it?
>
> Peter
>
More information about the antlr-interest
mailing list