<div dir="ltr">On Wed, Apr 3, 2013 at 7:00 AM, Carl Ritson <span dir="ltr"><<a href="mailto:C.G.Ritson@kent.ac.uk" target="_blank">C.G.Ritson@kent.ac.uk</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>Client, but didn't get round to finishing it. Also porting the TVM to Javascript is on my list of cool things to do that I'll never have time to do.</div>
</div></blockquote><div><br></div><div style>Yep. I have one of those lists, too. </div><div style><br></div><div style>Given that Emscripten was used to port the Unreal engine, though, I'm hoping that libtvm doesn't hide any complexity that would make this path break. I managed, in the in-between minutes yesterday, to get Emscripten installed on a VM; given how much time I have lately, it will probably take a week (at least) before I know if this path is viable. </div>
<div style><br></div><div style>I'll keep the list appraised, obviously.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>One thing that needs to be taken care with is appropriate sandboxing of the compiler. The occ21 compiler is relatively easy to crash (with malformed code), and hence could be exploited and used to hijack the host. Although I think the probability of this vector being exploited is low in reality.</div>
</div></blockquote><div><br></div><div style>Agreed on all counts. </div><div><br></div><div style>M </div></div></div></div>