From hans.huebner at gmail.com  Wed Jun 12 19:58:03 2013
From: hans.huebner at gmail.com (=?ISO-8859-1?Q?Hans_H=FCbner?=)
Date: Wed, 12 Jun 2013 21:58:03 +0200
Subject: Ignoring timeouts in Drakma
In-Reply-To: <5DC4A975-B4A0-49A6-9DE1-68F96B8159CC@mac.com>
References: <5DC4A975-B4A0-49A6-9DE1-68F96B8159CC@mac.com>
Message-ID: <CABj0bwKyHx4uSw6GFtVfyRAEAUTqpAXtfDqJaQM+P+pa9wd=mQ@mail.gmail.com>

This is not Drakma specific:

(defun try-request-from-multiple-urls (&rest urls)
  (loop
    (with-simple-restart (try-next "Try next URL")
      (handler-case
          (let ((url (first urls)))
            (format t "requesting ~A~%" url)
            (return (drakma:http-request url)))
        (usocket:timeout-error ()
          (setf urls (rest urls))
          (unless urls
            (format t "no more urls, returning~%")
            (return))
          (format t "timeout, trying next url~%")
          (invoke-restart 'try-next))))))

Tune as desired.  If you want to have the last timeout error percolate to
the caller, you'll want handler-bind instead.

-Hans


On Wed, Jun 12, 2013 at 9:42 PM, Patrick May <patrick.may at mac.com> wrote:

>         I have a list of IP addresses to poll and want to catch the
> USOCKET:TIMEOUT-ERROR from Drakma when a machine isn't available, log a
> message, and continue on mapping over the list.  Does anyone have an
> example of how to do this?
>
> Thanks,
>
> Patrick
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/drakma-devel/attachments/20130612/75cde449/attachment.html>

From patrick.may at mac.com  Thu Jun 13 01:33:30 2013
From: patrick.may at mac.com (Patrick May)
Date: Wed, 12 Jun 2013 21:33:30 -0400
Subject: Ignoring timeouts in Drakma
In-Reply-To: <CAB-HnLSKg9AeYGMsJzUpa6g-pyr_g8rLaUG==cuhUPrdDm_NrA@mail.gmail.com>
References: <5DC4A975-B4A0-49A6-9DE1-68F96B8159CC@mac.com>
 <CABj0bwKyHx4uSw6GFtVfyRAEAUTqpAXtfDqJaQM+P+pa9wd=mQ@mail.gmail.com>
 <73018FDD-5DFD-4011-9C90-90EF74A4B92F@mac.com>
 <CAB-HnLSKg9AeYGMsJzUpa6g-pyr_g8rLaUG==cuhUPrdDm_NrA@mail.gmail.com>
Message-ID: <EECD5550-7C2A-435C-8F4B-F1B6F7FF9E5D@mac.com>

On Jun 12, 2013, at 7:19 PM, Lu?s Oliveira <luismbo at gmail.com> wrote:
> On Wed, Jun 12, 2013 at 11:23 PM, Patrick May <patrick.may at mac.com> wrote:
>>> (defun try-request-from-multiple-urls (&rest urls)
>>>  (loop
>>>    (with-simple-restart (try-next "Try next URL")
>>>      (handler-case
>>>          (let ((url (first urls)))
>>>            (format t "requesting ~A~%" url)
>>>            (return (drakma:http-request url)))
>>>        (usocket:timeout-error ()
>>>          (setf urls (rest urls))
>>>          (unless urls
>>>            (format t "no more urls, returning~%")
>>>            (return))
>>>          (format t "timeout, trying next url~%")
>>>          (invoke-restart 'try-next))))))
>>> 
>>> Tune as desired.  If you want to have the last timeout error percolate to the caller, you'll want handler-bind instead.
>> 
>>        Thanks, Hans (and Drew C).  I figured out my problem on the train home (debugging by public embarrassment), but your use of with-simple-restart makes the code more understandable so I'll incorporate that.
> 
> Restarts are nice, but usually for communicating with other call
> sites. Seems a bit overkill here. If I understand your problem, all
> you need is handler-case really.
> 
> (defun poll-urls (urls)
>  (dolist (url urls)
>    (handler-case (drakma:http-request url)
>      (usocket:timeout-error ()
>        (format t "got timeout getting ~A~%" url)))))

	I ultimately went completely underkill and used ignore-error.  I need a shower now.

Thanks,

Patrick



From luismbo at gmail.com  Wed Jun 12 23:19:56 2013
From: luismbo at gmail.com (=?ISO-8859-1?Q?Lu=EDs_Oliveira?=)
Date: Thu, 13 Jun 2013 00:19:56 +0100
Subject: Ignoring timeouts in Drakma
In-Reply-To: <73018FDD-5DFD-4011-9C90-90EF74A4B92F@mac.com>
References: <5DC4A975-B4A0-49A6-9DE1-68F96B8159CC@mac.com> <CABj0bwKyHx4uSw6GFtVfyRAEAUTqpAXtfDqJaQM+P+pa9wd=mQ@mail.gmail.com>
 <73018FDD-5DFD-4011-9C90-90EF74A4B92F@mac.com>
Message-ID: <CAB-HnLSKg9AeYGMsJzUpa6g-pyr_g8rLaUG==cuhUPrdDm_NrA@mail.gmail.com>

On Wed, Jun 12, 2013 at 11:23 PM, Patrick May <patrick.may at mac.com> wrote:
>> (defun try-request-from-multiple-urls (&rest urls)
>>   (loop
>>     (with-simple-restart (try-next "Try next URL")
>>       (handler-case
>>           (let ((url (first urls)))
>>             (format t "requesting ~A~%" url)
>>             (return (drakma:http-request url)))
>>         (usocket:timeout-error ()
>>           (setf urls (rest urls))
>>           (unless urls
>>             (format t "no more urls, returning~%")
>>             (return))
>>           (format t "timeout, trying next url~%")
>>           (invoke-restart 'try-next))))))
>>
>> Tune as desired.  If you want to have the last timeout error percolate to the caller, you'll want handler-bind instead.
>
>         Thanks, Hans (and Drew C).  I figured out my problem on the train home (debugging by public embarrassment), but your use of with-simple-restart makes the code more understandable so I'll incorporate that.

Restarts are nice, but usually for communicating with other call
sites. Seems a bit overkill here. If I understand your problem, all
you need is handler-case really.

(defun poll-urls (urls)
  (dolist (url urls)
    (handler-case (drakma:http-request url)
      (usocket:timeout-error ()
        (format t "got timeout getting ~A~%" url)))))

Cheers,

--
Lu?s Oliveira
http://r42.eu/~luis/


From patrick.may at mac.com  Wed Jun 12 22:23:36 2013
From: patrick.may at mac.com (Patrick May)
Date: Wed, 12 Jun 2013 18:23:36 -0400
Subject: Ignoring timeouts in Drakma
In-Reply-To: <CABj0bwKyHx4uSw6GFtVfyRAEAUTqpAXtfDqJaQM+P+pa9wd=mQ@mail.gmail.com>
References: <5DC4A975-B4A0-49A6-9DE1-68F96B8159CC@mac.com>
 <CABj0bwKyHx4uSw6GFtVfyRAEAUTqpAXtfDqJaQM+P+pa9wd=mQ@mail.gmail.com>
Message-ID: <73018FDD-5DFD-4011-9C90-90EF74A4B92F@mac.com>

On Jun 12, 2013, at 3:58 PM, Hans H?bner <hans.huebner at gmail.com> wrote:
> This is not Drakma specific:
> 
> (defun try-request-from-multiple-urls (&rest urls)
>   (loop
>     (with-simple-restart (try-next "Try next URL")
>       (handler-case
>           (let ((url (first urls)))
>             (format t "requesting ~A~%" url)
>             (return (drakma:http-request url)))
>         (usocket:timeout-error ()
>           (setf urls (rest urls))
>           (unless urls
>             (format t "no more urls, returning~%")
>             (return))
>           (format t "timeout, trying next url~%")
>           (invoke-restart 'try-next))))))
> 
> Tune as desired.  If you want to have the last timeout error percolate to the caller, you'll want handler-bind instead.

	Thanks, Hans (and Drew C).  I figured out my problem on the train home (debugging by public embarrassment), but your use of with-simple-restart makes the code more understandable so I'll incorporate that.

	Sorry for being too quick to post.

Patrick




From patrick.may at mac.com  Wed Jun 12 19:42:54 2013
From: patrick.may at mac.com (Patrick May)
Date: Wed, 12 Jun 2013 15:42:54 -0400
Subject: Ignoring timeouts in Drakma
Message-ID: <5DC4A975-B4A0-49A6-9DE1-68F96B8159CC@mac.com>

	I have a list of IP addresses to poll and want to catch the USOCKET:TIMEOUT-ERROR from Drakma when a machine isn't available, log a message, and continue on mapping over the list.  Does anyone have an example of how to do this?

Thanks,

Patrick