On Fri, Apr 30, 2010 at 9:39 PM, Wesley W. Terpstra <span dir="ltr"><<a href="mailto:wesley@terpstra.ca">wesley@terpstra.ca</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
amd64-debian.<br><br>I thought it might be bad RAM, so I reran and saw exactly the same effect. Perhaps something to do with debian/sid? The machine also has 64GB of RAM if that is relevant. I'll try a debootstrap of sid chroot on my 4GB debian box this weekend to see what happens.<br>
</blockquote></div><br>I've tried debian/lenny on that same 64GB machine, and it built clean. I then used the regression free mlton built by debian/lenny in debian/sid, and it passed. On a hunch, I then rebuilt ONLY the runtime using debian/sid and immediately got regressions.<br>
<br>My conclusion? gcc 4.4.3-9 is miscompiling our runtime, while 4.3.2-1.1 (lenny) and 4.4.3-2 (last sid upload) do not. I'll try to narrow down which object file in particular causes the problem.<br>