HI Lukas,<br><br><div class="gmail_quote">On 11 October 2010 08:32, Lukas Renggli <span dir="ltr">&lt;<a href="mailto:renggli@gmail.com">renggli@gmail.com</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Wow, that&#39;s a very nasty bug. Amazing that this hasn&#39;t hit anybody<br>
before. Looks like this could happen quite common in systems like<br>
Pier, not only with MABooleanDescription instances.<br>
<br>
  <a href="http://code.google.com/p/magritte-metamodel/issues/detail?id=7" target="_blank">http://code.google.com/p/magritte-metamodel/issues/detail?id=7</a><br>
<br>
The validation code is really ugly. And likely this bug was introduced<br>
when trying to support the validation for recursive structures. I<br>
would be happy if someone could rewrite that code? (also get rid of<br>
these nested, resumable exceptions).<br></blockquote><div><br></div><div>I&#39;m happy to help, especially as I&#39;d like a fix for the code I&#39;m developing. </div><div>However I&#39;m struggling to understand the intent of the some of the code for example I don&#39;t understand the intent behind the use of object and previous in MAGraphVisitor&gt;&gt;#use:during</div>
<div><br></div><div>MAGraphVisitor&gt;&gt;#use: anObject during: aBlock</div><div><span class="Apple-tab-span" style="white-space:pre">	</span>| previous |</div><div><span class="Apple-tab-span" style="white-space:pre">	</span>(seen includes: anObject)</div>
<div><span class="Apple-tab-span" style="white-space:pre">		</span>ifTrue: [ ^ self ].</div><div><span class="Apple-tab-span" style="white-space:pre">	</span>anObject isNil</div><div><span class="Apple-tab-span" style="white-space:pre">		</span>ifFalse: [ seen add: anObject ].</div>
<div><span class="Apple-tab-span" style="white-space:pre">	</span>previous := object. object := anObject.</div><div><span class="Apple-tab-span" style="white-space:pre">	</span>aBlock ensure: [ object := previous ]   </div>
</div>