[antlr-interest] ANTLR v3.1.1 released

[Terence Parr]

Hi. Good thought, though it should not be doing threads by default.
Ter
On Oct 21, 2008, at 12:08 PM, Randall R Schulz wrote:

> On Tuesday 21 October 2008 11:52, Randall R Schulz wrote:
>> On Tuesday 21 October 2008 11:44, Terence Parr wrote:
>>> ...
>>>
>>> It's hard to see a code trace that could given an error (not
>>> warning) that wouldn't be trapped here.  Hmm...ok, let me know if
>>> you find a future issue.
>>>
>>> Ter
>>
>> From the 3.1.1 source:
>>
>> -==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==-
>>    public static ErrorState getErrorState() {
>>        ErrorState ec =  
>> (ErrorState)threadToErrorStateMap.get(Thread.currentThread()); if  
>> ( ec==null ) {
>>            ec = new ErrorState();
>>            threadToErrorStateMap.put(Thread.currentThread(), ec);
>>        }
>>        return ec;
>>    }
>>
>>    public static int getNumErrors() {
>>        return getErrorState().errors;
>>    }
>> -==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==-
>>
>>
>> Perhaps its some kind of thread issue / mix-up?
>
>
> I see something called "two-threaded DFA conversion" and a class  
> called
> NFAConversionThread. Maybe the errors are getting reported on a thread
> other than the one you check when wrapping up?
>
>
> Randall Schulz
>
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>