<div dir="ltr"><div class="gmail_quote">On Wed, May 27, 2015 at 10:32 AM Duncan P. N. Exon Smith <<a href="mailto:dexonsmith@apple.com">dexonsmith@apple.com</a>> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
> On 2015 May 27, at 10:13, Reid Kleckner <<a href="mailto:rnk@google.com" target="_blank">rnk@google.com</a>> wrote:<br>
><br>
> On Tue, May 26, 2015 at 1:46 PM, Owen Anderson <<a href="mailto:resistor@mac.com" target="_blank">resistor@mac.com</a>> wrote:<br>
> I have no objections, though I think lib/MachineIR would make a prettier bikeshed ;-)<br>
><br>
> +1 for this. I think you probably want to limit the scope of your change to just splitting out MachineIR and not trying to rename CodeGen, AsmPrinter, etc. :)<br>
<br>
Agreed :). The rest seems out-of-scope for Alex's internship.<br></blockquote><div><br></div><div>I think even this is going to be out-of-scope unless the *existing* code happens to have a clean layering separation. I don't see one currently, but I might be missing it.</div><div><br></div><div>Specifically, I'm *strongly* opposed to splitting this out while there are circular references between any MI implementation and any CodeGen implementation. Not all of our linkers or platforms even allow this, but the errors tend to be brittle and sporadic rather than consistent. </div></div></div>