<div dir="ltr">That's really good point, we probably should use ulittle32_t instead. Regarding alignment, don't we guarantee that the in-memory object we are building aligned properly?</div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Sep 14, 2015 at 5:34 PM, Rafael Ávila de 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">rafael accepted this revision.<br>
rafael added a comment.<br>
This revision is now accepted and ready to land.<br>
<br>
LGTM.<br>
<br>
The new version is writing directly to mmaped memory. It is also using a type that in unaligned and may have non-native endian.<br>
<br>
But this is code that will likely be revisited once we can link something large enough to benchmark, so yes, lets keep the simpler version for now.<br>
<br>
<br>
<a href="http://reviews.llvm.org/D12867" rel="noreferrer" target="_blank">http://reviews.llvm.org/D12867</a><br>
<br>
<br>
<br>
</blockquote></div><br></div>