[gme-users] Too many passes while parsing XML file error

[James Edmondson]

A reference to a reference of the same type? I hadn't even thought of
that. It seems amazingly dangerous. I'll give it a shot.

Thanks,
James


On Wed, Jun 29, 2011 at 12:57 PM, Kevin Smyth
<ksmyth at isis.vanderbilt.edu> wrote:
> Yes, those are all valid combinations. Try references R1 and R2 where R1->R2 and R2->R1. Or try Subtype B' which contains its Archetype B.
>
>
> -----Original Message-----
> From: gme-users-bounces at list.isis.vanderbilt.edu [mailto:gme-users-bounces at list.isis.vanderbilt.edu] On Behalf Of James Edmondson
> Sent: Wednesday, June 29, 2011 12:54 PM
> To: gme-users
> Subject: Re: [gme-users] Too many passes while parsing XML file error
>
> As a followup to this, I can't seem to replicate this problem with a simple metamodel in GME 11.3.23. I constructed a metamodel which allows for a Model A to contain both references to Model A and objects of Model A, and attempted the following model derivations:
>
> 1.) A model B of type A which contains a reference R to model B
> 2.) A model B of type A which contains a subtype S of model B
> 3.) A model B of type A which contains an instance I of model B
> 4.) A model B of type A which contains a copy C of model B
>
> Neither of these seemed to go through the 100 passes. To download the paradigm I created and the XME files I have exported, you can look
> here:
> http://sites.google.com/site/distributedreasoner/Cycles_Attempts.zip
>
> Either GME can only have this problem in larger models, or I am missing something. Any thoughts on how I can replicate this issue?
>
> Thanks,
> James
>
>
> On Wed, Jun 29, 2011 at 10:11 AM, James Edmondson <jedmondson at gmail.com> wrote:
>> Hi Kevin,
>>
>> I would think that a good debug message could be constructed using a
>> coloring scheme or by keeping a list of generated models so far in the
>> importation process and having a simple check for cycles in the XME
>> structure that looks at the previously concretized models from the XME
>> source. I'm not sure if this is a common bug with model developers or
>> not, but if it is, a cycle detection algorithm in the XME import would
>> be very useful, I would think, in tracking down the culprit and
>> constructing detailed debugging information.
>>
>> Cycle and deadlock detection have a lot of attached literature. Does
>> anyone already have a deadlock detection interpreter for GME that
>> might be modified to look for this in an XME file (before it is
>> imported). This might help James Hill find the cycle in his client's
>> large XME model. He probably doesn't have the MGA file, so this is
>> sort of like finding a needle in an XML haystack. Does anyone have
>> experience in looking for these cycles in XME files so we all might
>> know specifically what to look for when this happens in our own
>> models?
>>
>> Thanks,
>> James
>>
>>
>>
>> On Wed, Jun 29, 2011 at 9:41 AM, Kevin Smyth <ksmyth at isis.vanderbilt.edu> wrote:
>>> Yes, this only happens when there is a circular reference. For
>>> example, if you have a subtype, then move the archetype into the
>>> subtype, there is a circular reference. Or if a set has a reference
>>> member that references the set. Or a simple circular reference chain.
>>>
>>> Unfortunately, due to the design of the xme parser, the only way to
>>> determine what is happening is with a debugger.
>>>
>>> Kevin
>>>
>>> -----Original Message-----
>>> From: gme-users-bounces at list.isis.vanderbilt.edu
>>> [mailto:gme-users-bounces at list.isis.vanderbilt.edu] On Behalf Of
>>> James H. Hill
>>> Sent: Tuesday, June 28, 2011 4:08 PM
>>> To: gme-users
>>> Subject: Re: [gme-users] Too many passes while parsing XML file error
>>>
>>> Hi James,
>>>
>>> On Jun 28, 2011, at 4:52 PM, James Edmondson wrote:
>>>
>>>> Hi James,
>>>>
>>>> Do you have a link to the XME available for download?
>>>
>>> First, let me check with sponsors before attaching the model. ;-)
>>>
>>>> The only thing I
>>>> can think of is a circular ref of a recursive model.
>>>
>>>> Is that something
>>>> that is possible with your paradigm?
>>>
>>>> Like a model ref to itself within
>>>> the model?
>>>>
>>>>
>>>
>>> That could be possible. It would be great if GME would report why it
>>> is going though 100+ phases in the parsing. I have never seen it go
>>> over 3, even for large models. :-)
>>>
>>> Thanks,
>>>
>>> James
>>>
>>>> Thanks,
>>>> James
>>>>
>>>>
>>>> On Tue, Jun 28, 2011 at 12:39 PM, James H. Hill <hillj at cs.iupui.edu>
>>>> wrote:
>>>>> Hi,
>>>>>
>>>>> I have a model that when I try to import it, goes through at least
>>>>> 100+ passes before displaying the following error message:
>>>>>
>>>>>   Too many passes while parsing the XML file.
>>>>>
>>>>> Has anyone experienced this problem?! If so, how did you resolve
>>>>> this
>>>
>>>>> problem? This is the first time I have encountered this problem,
>>>>> and the .xme files in only 181 KB, which is small compared to
>>>>> others that
>>>
>>>>> I am able to successfully import.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> James
>>>>>
>>>>> _______________________________________________
>>>>> gme-users mailing list
>>>>> gme-users at list.isis.vanderbilt.edu
>>>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>>>>
>>>> _______________________________________________
>>>> gme-users mailing list
>>>> gme-users at list.isis.vanderbilt.edu
>>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>>
>>> _______________________________________________
>>> gme-users mailing list
>>> gme-users at list.isis.vanderbilt.edu
>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>> _______________________________________________
>>> gme-users mailing list
>>> gme-users at list.isis.vanderbilt.edu
>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>>
>>
> _______________________________________________
> gme-users mailing list
> gme-users at list.isis.vanderbilt.edu
> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
> _______________________________________________
> gme-users mailing list
> gme-users at list.isis.vanderbilt.edu
> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>