<div class="gmail_quote">On Thu, Aug 8, 2013 at 5:10 AM, Rafael Espíndola <span dir="ltr"><<a href="mailto:rafael.espindola@gmail.com" target="_blank">rafael.espindola@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On 8 August 2013 01:36, Richard Smith <<a href="mailto:richard@metafoo.co.uk">richard@metafoo.co.uk</a>> wrote:<br>
> Typos 'conpute', 'allignof', and 'comput' in the comment.<br>
<br>
</div>Fixed.</blockquote><div><br></div><div>Looks like you're also missing an update to ASTContext.h to change the parameter name there?</div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">
> I'm not really a big fan of ASTContext knowing about the semantics of the<br>
> GNU alignof extension like this, but the change looks correct. Changing the<br>
> comment to talk about semantic alignment versus ABI alignment would help<br>
> (but that doesn't cover the following-references / not-following-references<br>
> half).<br>
<br>
</div>AST suffers from being the place we have for sharing code between Sema<br>
and Codegen. As you point out, "semantic" would be a bit vague as what<br>
we compute is specific to alignof.<br></blockquote><div><br></div><div>I suspect we'll want to revisit this at some point, but LGTM for now.</div></div>