<div dir="ltr"><div class="gmail_extra">Hum, that's not good indeed... I'm not sure how to fix that, at least not within the context of this patch. </div><div class="gmail_extra"><br></div><div class="gmail_extra">
Is there a technical reason why we don't parse the inline asm when outputting asm? It seems like we should, at least to make sure it's well formed, or it contains the same instruction set, to give users a better error...</div>
<div class="gmail_extra"><br><br><div class="gmail_quote">On 12 December 2013 21:41, Jim Grosbach <span dir="ltr"><<a href="mailto:grosbach@apple.com" target="_blank">grosbach@apple.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> While commenting on PR18231, I had a horrible realization. This patch will cause diagnostics to be generated when compiling with "-c" that will not be generated when compiling with "-S". That's a problem. There shouldn't be any diagnostics that depend on whether we're doing direct-to-object or going through a .s file first. When generating a .s file, the inline assembly is just passed through as a textual blob with no interpretation done.<br>
</blockquote><div><br></div></div></div></div>