[hunchentoot-devel] A potential Hunchentoot/Lispworks(?) fix

[Edi Weitz]

[Cc to mailing list.]

Thanks, that makes sense.  I've applied your patch to the dev version.

Edi.


On Mon, Feb 15, 2010 at 11:58 PM, Matt Lamari <matt.lamari at gmail.com> wrote:
>
> (Thanks to Martin Simmons of Lispworks for this).
>
>
> I'm still testing; but
> in Acceptor.lisp
> process-connection
>
> At the end - change it so, even if the force-output fails, the stream
> will still be closed.
>
> (I'm still testing at my end - this took the leaking problem WAY down)
>
>
> (when *hunchentoot-stream*
>        ;; as we are at the end of the request here, we ignore all
>        ;; errors that may occur while flushing and/or closing the
>        ;; stream.
>        (ignore-errors*
>          (force-output *hunchentoot-stream*))
>        (ignore-errors*
>          (close *hunchentoot-stream* :abort t)))
>
> Edi Weitz wrote:
>> On Thu, Feb 11, 2010 at 8:35 AM, Matt Lamari <matt.lamari at gmail.com> wrote:
>>
>>
>>> Sorry to pester you privately; but I was wondering if you could verify
>>> the leak on your Win32 Lispworks before I approach Lispworks.
>>>
>>> I don't know anyone else with a Win32 lispworks.
>>>
>>
>> I'll be out of town for a week and I'm pretty busy right now.  I
>> wouldn't be able to test anything before end of next week.  Maybe you
>> should ask on the LW mailing list if you're unsure?  There are lots of
>> people there using Win32 LW... :)
>>
>
>