[antlr-interest] Failure on OpenJDK on Debian

Ola Bini ola.bini at gmail.com
Wed Apr 1 01:16:18 PDT 2009


Sam Barnett-Cormack wrote:
> Ola Bini wrote:
>> Gavin Lambert wrote:
>>> At 02:44 1/04/2009, Ola Bini wrote:
>>>> I have a grammar that works well with several different Java 5 and
>>>> Java 6 environments on both Windows, Linux and MacOS X (including
>>>> SoyLatte).  But I've just got a bug report that it blows up on
>>>> OpenJDK build 1.6.0_0-b11, that can be installed in Debian.
>>>> Has anyone else seen this? The errors look like this:
>>>>
>>>> generate.antlr:
>>> Looks like your lexer rules are ambiguous (which probably isn't 
>>> helped by the fact that you're also using literals in your parser 
>>> rules).  You should definitely consider correcting your rules so 
>>> that the ambiguity warnings go away; doing that should make the 
>>> error go away as well.
>>>
>>>
>> I don't have any ambiguity warnings on any other Java implementation.
>
> I've had similar issues with ANTLR sporadically claiming that my lexer 
> rules are ambiguous (they aren't unless k=1), and I would hope that my 
> development machine isn't considered underpowered (1.6GHz Atom, 1GB 
> RAM, running up-to-date Ubuntu and the Sun JDK 1.6). Maybe the default 
> timeout should be a bit bigger?
>
Yep, something like that might be better. I did try to explicitly set K 
to 2, since the problems were all related to K=1 ambiguities, but that 
didn't change anything in this case.

Cheers

-- 
 Ola Bini (http://olabini.com) 
 Ioke creator (http://ioke.org)
 JRuby Core Developer (http://jruby.org)
 Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)
 Practical JRuby on Rails (http://apress.com/book/view/9781590598818)

 "Yields falsehood when quined" yields falsehood when quined.




More information about the antlr-interest mailing list