<html><head><meta http-equiv="Content-Type" content="text/html charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div><div>On May 11, 2012, at 9:59 AM, Sergei Larin <<a href="mailto:slarin@codeaurora.org">slarin@codeaurora.org</a>> wrote:</div><blockquote type="cite"> Here is the latest patch version, addressing your comments. I still call<br>getUnderlyingObject() for reasons mentioned in comments.<br>It does look cleaner that way, so hopefully we are getting closer to being<br>done.<br><br>Sergei<br><br><span><patch_aa_in_dag_v3.patch></span></blockquote></div><br><div>Sergei, I think this is committable now. Two minor comments.</div><div><br></div><div><pre style="line-height: 16px; background-color: rgb(248, 248, 248); "><span class="gi" style="color: rgb(0, 160, 0); ">+ // Is this pointer refers to</span></pre><div>Grammar</div></div><div><br></div><div><pre style="line-height: 16px; background-color: rgb(248, 248, 248); position: static; z-index: auto; "><span class="gi" style="color: rgb(0, 160, 0); ">+ // TODO: Need to handle multiple memory operands.</span>
<span class="gi" style="color: rgb(0, 160, 0); ">+ assert(MIa->hasOneMemOperand() && "Multiple memory operands.");</span></pre>I think assert should only be for things that can't happen in theory. For unimplemented pieces I suggest something like llvm_unreachable that triggers in a release build.</div><div><br></div><div>-Andy</div></body></html>