[antlr-interest] Bug in exception initialization code?

[Michael Bedward]

Hi Søren,

You're right - it's not the most informative error message :)

Michael

On 29 March 2011 01:35, Søren Boisen <sboisen at mail.dk> wrote:
> Hi Michael,
>
> Thanks for the reply. I agree that it was due to an error in my tree walker
> grammar (that I've now found and fixed), but I think it is pretty bad that
> ANTLR generates an internal error with no useful information, whenever I
> make a mistake in my grammar. It certainly doesn't make it easy to learn
> ANTLR and how to create tree parsers :-)
>
> I was therefore hoping for Ter or someone else with access, to create an
> issue in the bug tracker, so it could be solved eventually.
>
> Regards
> Søren
>
> 2011/3/28 Michael Bedward <michael.bedward at gmail.com>
>>
>> Hi Søren,
>>
>> I see that error every so often and, so far al least, it has always
>> been due to an error in my tree walker grammar - specifically an
>> alternative that arises in the AST but that I forgot to consider in
>> the grammar.
>>
>> Michael
>>
>> On 28 March 2011 23:29, Søren Boisen <sboisen at mail.dk> wrote:
>> > Greetings,
>> > I have a encountered a problem while trying to write a tree parser for
>> > my
>> > grammar. Sometimes when trying to walk the AST and the tree parser fails
>> > to
>> > match some input, I get the following exception:
>> >
>> > can't look backwards more than one token in this stream
>> >    org.antlr.runtime.misc.LookaheadStream.LB(LookaheadStream.java:159)
>> >    org.antlr.runtime.misc.LookaheadStream.LT(LookaheadStream.java:120)
>> >
>> >
>> > org.antlr.runtime.RecognitionException.extractInformationFromTreeNodeStream(RecognitionException.java:144)
>> >
>> >  org.antlr.runtime.RecognitionException.(RecognitionException.java:111)
>> >
>> >
>> > org.antlr.runtime.MismatchedTreeNodeException.(MismatchedTreeNodeException.java:42)
>> >
>> >
>> > org.antlr.runtime.tree.TreeParser.recoverFromMismatchedToken(TreeParser.java:135)
>> >    org.antlr.runtime.BaseRecognizer.match(BaseRecognizer.java:115)
>> >
>> >
>> > com.brunata.ws.rest.odata.query.NearestTelegramQueryWalker.queryOp(NearestTelegramQueryWalker.java:243)
>> >
>> >
>> > com.brunata.ws.rest.odata.query.NearestTelegramQueryWalker.stmt(NearestTelegramQueryWalker.java:128)
>> >
>> > I have looked through the FAQ, the issue tracker and Google, but I
>> > couldn't
>> > find anything that looked related.
>> >
>> > From the stack trace it looks like an error in the
>> > extractInformationFromTreeNodeStream, since I guess it shouldn't expect
>> > to
>> > always be able to look back more than one token in a stream.
>> > Is there a workaround for this? This is the code I use for creating
>> > parsers
>> > and streams:
>> >
>> >        String query = ...;  // this is a parameter to the method
>> >        Map<String, String> propertyFilters = ...; // this is a field on
>> > the
>> > class
>> >
>> >        ANTLRStringStream input = new ANTLRStringStream(query);
>> >        QueryParserLexer lexer = new QueryParserLexer(input);
>> >
>> >        CommonTokenStream tokens = new CommonTokenStream(lexer);
>> >        QueryParserParser parser = new QueryParserParser(tokens);
>> >
>> >        QueryParserParser.stmt_return result = tryParse(parser, query);
>> >        ast = (CommonTree) result.getTree();
>> >
>> >        CommonTreeNodeStream nodes = new CommonTreeNodeStream(ast);
>> >        nodes.setTokenStream(tokens);
>> >
>> >        NearestTelegramQueryWalker walker = new
>> > NearestTelegramQueryWalker(nodes);
>> >        propertyFilters = tryWalk(walker, query);
>> >
>> > Am I using the wrong kind of token/node streams or something?
>> >
>> > Regards,
>> > Søren Boisen
>> >
>> > List: http://www.antlr.org/mailman/listinfo/antlr-interest
>> > Unsubscribe:
>> > http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>> >
>
>