[antlr-interest] Re: Intermitent bug with ANTLR.
Oliver Zeigermann
oliver at zeigermann.de
Fri Jan 30 13:44:27 PST 2004
I just checked this and I guess it actually does not apply. As far as I
know Daniel uses the XPA XML parser, not a SAX one to drive ANTLR, but
this one is single-threaded. Thus it can't be a race and I am clueless
as well...
Oliver
Oliver Zeigermann wrote:
> Hi guys,
>
> I really can't believe this is a VM bug (although I just found an
> obvious one the other day, so, hmmmm, maybe). What I rather suspect is
> there is some sort of race. At least this is indicated by this
> undeterministic behavior. To have a race you need at least two threads
> running. As far as I know ANTLR is strictly single-threaded, right? But,
> trara, XPA is *not*. It uses a producer/consumer thing to give both SAX
> parser an ANTLR the impression they are the master of it all. Although
> this code is very simple, there might still be a bug in it. If you can
> give me any hint upon which conditions this error occurs, I would highly
> appreciate this.
>
> Cheers,
> Oliver
>
>
> lachinois2002 wrote:
>
>>Hi Monty,
>>
>>Well... I'm outa luck, after adding 2 or 3 rules to my grammar it
>>seems like the bug doesnt show up anymore. This really looks like a
>>java VM bug. In anycase, if this happens again I'll trace it and
>>hope it behaves the same with the trace activated.
>>
>>Daniel Shane
>>
>>--- In antlr-interest at yahoogroups.com, mzukowski at y... wrote:
>>
>>
>>>To continue on the theme of the week, try building with -trace and
>>
>>then
>>
>>
>>>running. I'd be curious to see the difference between success &
>>
>>failure.
>>
>>
>>>Monty
>>>
>>>-----Original Message-----
>>>From: Daniel Shane [mailto:lachinois at h...]
>>>Sent: Wednesday, January 28, 2004 11:49 AM
>>>To: antlr-interest at yahoogroups.com
>>>Subject: [antlr-interest] Intermitent bug with ANTLR.
>>>
>>>Hi everyone, mostly Terence.
>>>
>>>I will greatly need your help on this one. I have a grammar (that
>>
>>uses XPA)
>>
>>
>>>and, listen to this, I sometimes get a grammar error, and
>>
>>sometimes not by
>>
>>
>>>executing the same command line with the same input.
>>>
>>>Here is what happens:
>>>
>>>$ java -cp
>>>"/usr/local/soft/antlr-2.7.2/lib/antlr.jar:/usr/local/soft/xpa-
>>
>>1.0.1/lib/xpa
>>
>>
>>>1_0_1-full.jar:.:Code"
>>>MainRegulationParser tmp/Fr/C.R.C.,_c._100.06-MTE.html > /tmp/g
>>>
>>>$ java -cp
>>>"/usr/local/soft/antlr-2.7.2/lib/antlr.jar:/usr/local/soft/xpa-
>>
>>1.0.1/lib/xpa
>>
>>
>>>1_0_1-full.jar:.:Code"
>>>MainRegulationParser tmp/Fr/C.R.C.,_c._100.06-MTE.html > /tmp/g
>>>
>>>$ java -cp
>>>"/usr/local/soft/antlr-2.7.2/lib/antlr.jar:/usr/local/soft/xpa-
>>
>>1.0.1/lib/xpa
>>
>>
>>>1_0_1-full.jar:.:Code"
>>>MainRegulationParser tmp/Fr/C.R.C.,_c._100.06-MTE.html > /tmp/g
>>>
>>>$ java -cp
>>>"/usr/local/soft/antlr-2.7.2/lib/antlr.jar:/usr/local/soft/xpa-
>>
>>1.0.1/lib/xpa
>>
>>
>>>1_0_1-full.jar:.:Code"
>>>MainRegulationParser tmp/Fr/C.R.C.,_c._100.06-MTE.html > /tmp/g
>>>line 3: expecting "<IdOfRegulation>", found 'TitleOfAct'
>>>
>>>$ java -cp
>>>"/usr/local/soft/antlr-2.7.2/lib/antlr.jar:/usr/local/soft/xpa-
>>
>>1.0.1/lib/xpa
>>
>>
>>>1_0_1-full.jar:.:Code"
>>>MainRegulationParser tmp/Fr/C.R.C.,_c._100.06-MTE.html > /tmp/g
>>>
>>>$
>>>
>>>These invocation were done in sequence and the input file has not
>>
>>changed
>>
>>
>>>between invocations. I printed some debug out of ANTLR to show the
>>
>>sequence
>>
>>
>>>of tokens that have been read and in both cases the token sequence
>>
>>is the
>>
>>
>>>same.
>>>
>>>Something is happening in ANTLR that is causing it to take a
>>
>>different path
>>
>>
>>>for what seems like a random reason.
>>>
>>>This is really causing me a headache, and I can trace the problem
>>
>>to one
>>
>>
>>>rule I added in the grammar. If I remove this rule, then I do not
>>
>>get this
>>
>>
>>>strange behavior (BTW, the rule is simple, so I'm not doing
>>
>>anything tricky
>>
>>
>>>here). I really cannot see why this additionnal rule would break
>>
>>anything
>>
>>
>>>and specially not like this.
>>>
>>>I really out of clues here, what would you like me to do to give
>>
>>you some
>>
>>
>>>additionnal information? Do you want the java files/grammar so
>>
>>that you can
>>
>>
>>>run it on your setup see if it does the same? Would you like an
>>
>>ssh access
>>
>>
>>>to this linux machine to see for yourself? Do you think the bug
>>
>>arrises from
>>
>>
>>>XPA?
>>>
>>>I'm running on Linux, RedHat 9 with the SUN JDK 1.4.2 for linux.
>>>
>>>Daniel Shane
>>>
>>>_________________________________________________________________
>>>MSN Search, le moteur de recherche qui pense comme vous !
>>>http://fr.ca.search.msn.com/
>>>
>>>
>>>
>>>
>>>Yahoo! Groups Links
>>>
>>>To visit your group on the web, go to:
>>>http://groups.yahoo.com/group/antlr-interest/
>>>
>>>To unsubscribe from this group, send an email to:
>>>antlr-interest-unsubscribe at yahoogroups.com
>>>
>>>Your use of Yahoo! Groups is subject to:
>>>http://docs.yahoo.com/info/terms/
>>
>>
>>
>>
>>
>>Yahoo! Groups Links
>>
>>To visit your group on the web, go to:
>> http://groups.yahoo.com/group/antlr-interest/
>>
>>To unsubscribe from this group, send an email to:
>> antlr-interest-unsubscribe at yahoogroups.com
>>
>>Your use of Yahoo! Groups is subject to:
>> http://docs.yahoo.com/info/terms/
>>
>>
>>
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
> To visit your group on the web, go to:
> http://groups.yahoo.com/group/antlr-interest/
>
> To unsubscribe from this group, send an email to:
> antlr-interest-unsubscribe at yahoogroups.com
>
> Your use of Yahoo! Groups is subject to:
> http://docs.yahoo.com/info/terms/
>
>
>
Yahoo! Groups Links
To visit your group on the web, go to:
http://groups.yahoo.com/group/antlr-interest/
To unsubscribe from this group, send an email to:
antlr-interest-unsubscribe at yahoogroups.com
Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
More information about the antlr-interest
mailing list