<p dir="ltr">I was thinking more along the lines of a unit test instead of relying on system stack limits, but am not too strongly tied to that.</p>
<p dir="ltr">~Aaron</p>
<div class="gmail_quote">On Jul 24, 2013 10:13 PM, "Rafael EspĂndola" <<a href="mailto:rafael.espindola@gmail.com">rafael.espindola@gmail.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On 24 July 2013 15:34, Aaron Ballman <<a href="mailto:aaron@aaronballman.com">aaron@aaronballman.com</a>> wrote:<br>
> Patch LGTM; is there a way that we could add a test to ensure<br>
> diagnoseOverflow doesn't break again (since this is fixing a<br>
> regression)?<br>
<br>
Thanks, I committed it as r187097.<br>
<br>
I found this by enabling pipefail and running the tests on a windows<br>
vm. I hope to be able to enable it really soon now. Once that is in,<br>
parser_overflow.c should fail on any systems that goes out of stack<br>
space. I guess we could explicitly run ulimit in the test, but I am<br>
not sure that adds a lot of value.<br>
<br>
> ~Aaron<br>
<br>
Cheers,<br>
Rafael<br>
</blockquote></div>