>I've had a PR open since 1.3H about a nasty resolver bug that breaks
>one of my applications.  The PR was filed with a clear description of
>the problem, a trivial set of steps to artificially reproduce the
>failure mode, and a patch to correct the bug.
>
>Each time I've mentioned this bug here I've been asked to update my
>patch for the latest version of res_send.c and resubmit it.  So far
>there are three versions of the patch in the PR database, and no
>obvious sign of anything being done about it.

	first of all, sorry for putting the PR on hold for a long time.

>This bug was also reported to the BIND development team at the same
>time.  They quickly acknowledged the problem and fixed it immediately
>in their distribution.

	actually, the patch on the PR is little bit aggressive than BIND 4.9.7
	when we have less-than-0 value in "seconds" variable.  Not sure
	if it makes large difference or not.  I think I will put a code
	equivalent to BIND 4.9.7 to the tree.

	Your patch: set seconds to 0 when seconds < 0, 
	BIND 4.9.7 and 8.2.2: set seconds to 1 when seconds <= 0, 

itojun