<p><span class="h-card" translate="no"><a href="https://mastodon.social/@regehr" class="u-url mention">@<span>regehr</span></a></span> it is, it&#39;s just a pain to lower our undef semantics into SMT queries</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> at the SMT query level, it should be pretty reasonable, you&#39;re basically saying &quot;give me a model where the target displays a behavior not shown by the source&quot; and then if this is unsat you&#39;re good</p>

<p><span class="h-card" translate="no"><a href="https://not.acu.lt/@ignaloidas" class="u-url mention">@<span>ignaloidas</span></a></span> </p><p>and X 0 -&gt; 0<br />or X 1 -&gt; 1<br />mux X 0 0 -&gt; 0<br />mux X 1 1 -&gt; 1</p><p>most of the rest is representable via the mux</p>

<p><span class="h-card" translate="no"><a href="https://not.acu.lt/@ignaloidas" class="u-url mention">@<span>ignaloidas</span></a></span> i don&#39;t think i can describe it in a mastodon post? it&#39;s relatively standard x-propagation but the devil is in the details</p>

@whitequark@mastodon.social what semantics for undef did you choose?

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@regehr" class="u-url mention">@<span>regehr</span></a></span> refinement is an _incredible pain_</p>

<p>adding refinement (verification of transformations with undefs)</p>

<p>Note that when I say &quot;arm trans kids&quot; I am not advocating violence. The weapons are strictly for deterrence. This is why, if a kid comes out as trans, they should be issued one (1) UGM-133A Trident II intercontinental ballistic missile</p>

<p><span class="h-card" translate="no"><a href="https://donotsta.re/users/mei" class="u-url mention">@<span>mei</span></a></span> coarse DCE is the only thing that is allowed to actually delete cells; any pass can unalive a cell but it&#39;s just scheduled for deletion if it does</p><p>this is because it also compacts the numbering</p>