<div dir="ltr"><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Sep 5, 2014 at 6:27 PM, Paul Davis <span dir="ltr"><<a href="mailto:paul.joseph.davis@gmail.com" target="_blank">paul.joseph.davis@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 id=":10x" class="a3s" style="overflow:hidden">Is that enough of a problem you'd want to see an option to disable the<br>
sub-binary and instead make out of the source binary?</div></blockquote></div><br>Yes and no. If you end up copying the data to another process, the problem goes away. It is a hard call, because the safest solution is to copy. But at times, you want the other option and then you pin a couple binaries with binary:copy/1 in order to achieve speed. I don't know how many have been hit by this in practice.</div><div class="gmail_extra"><br></div><div class="gmail_extra">My call is that if it is a lot, the default should probably be the safe bet. And then you have an option for the faster code path.<br><br clear="all"><div><br></div>-- <br>J.
</div></div>