gai_cancel()

[Carlos O'Donell]

On 06/26/2017 11:41 AM, Phil Blundell wrote:
> On Mon, 2017-06-26 at 17:35 +0200, Florian Weimer wrote:
>> How hard would it be to write a test case for this?  We also need a
>> ChangeLog entry.
> 
> It's a race, so a testcase is always going to be a bit hit-and-miss. 
> But something similar to the example from the original bug report would
> probably show it up fairly reliably.  I'll see what I can do.

Even if it is unreliably the result is that we have a false-positive and
that IMO is OK from a regression testing standpoint. If we ever see the
test fail it would indicate an incomplete fix and that's important
information.

-- 
Cheers,
Carlos.