From: ······@gmail.com
Subject: http request with timeout
Date: 
Message-ID: <1ffdfbfd-10d0-4278-8190-cf7c2ed5973b@s8g2000prg.googlegroups.com>
Hi, is there something like drakma with connection-timeout option (in
drakma this only works on LispWorks)  that work in sbcl ?

thanks
happy new year everyone
-milan
From: Edi Weitz
Subject: Re: http request with timeout
Date: 
Message-ID: <umyrpyi1k.fsf@agharta.de>
On Tue, 1 Jan 2008 01:45:49 -0800 (PST), ·······@gmail.com" <······@gmail.com> wrote:

> Hi, is there something like drakma with connection-timeout option
> (in drakma this only works on LispWorks) that work in sbcl ?

  http://common-lisp.net/pipermail/drakma-devel/2007-December/thread.html

Edi.

-- 

European Common Lisp Meeting, Amsterdam, April 19/20, 2008

  http://weitz.de/eclm2008/

Real email: (replace (subseq ·········@agharta.de" 5) "edi")
From: Alex Mizrahi
Subject: Re: http request with timeout
Date: 
Message-ID: <477a66c5$0$90265$14726298@news.sunsite.dk>
 ??>> Hi, is there something like drakma with connection-timeout option
 ??>> (in drakma this only works on LispWorks) that work in sbcl ?

 EW> 
http://common-lisp.net/pipermail/drakma-devel/2007-December/thread.html

what's the conclusion? (it's a bit hard to follow thread with multiple 
poster each having own opinion).
people should use Stanislaw Halik's patch, otherwise operation might hang 
for a long time (infinite?) in some circumstances?

i'd like to have some robust solution for HTTP client, because there can be 
different weird sites out there, and some can even have malicious intent.
as i understand currently DoS attack against client using drakma is fairly 
easy -- either implement some kind of tarpit so socket will hang for a long 
time, or send very long chunked response and wait until client dies because 
of memory starvation, because read-body function does not have an option to 
limit amoun of I/O aborting input, and it's not possible to override this on 
user level.

so to have robust solution one has to patch drakma (copying part of 
read-body implementation is kinda equivalent to patching) and/or user some 
custom implementation-dependant solution to handle timeouts. am i right? 
From: Edi Weitz
Subject: Re: http request with timeout
Date: 
Message-ID: <ud4sligb0.fsf@agharta.de>
On Tue, 1 Jan 2008 18:13:51 +0200, "Alex Mizrahi" <········@users.sourceforge.net> wrote:

> http://common-lisp.net/pipermail/drakma-devel/2007-December/thread.html
>
> what's the conclusion? (it's a bit hard to follow thread with
> multiple poster each having own opinion).

Yes, I also was a bit irritated that people dared to have their own
opinions.  I hear the situation is a bit better over in Python-land.

> people should use Stanislaw Halik's patch, otherwise operation might
> hang for a long time (infinite?) in some circumstances?

They could use Drakma with LispWorks and utilize the keyword arguments
CONNECTION-TIMEOUT, READ-TIMEOUT, and WRITE-TIMEOUT.

Edi.

-- 

European Common Lisp Meeting, Amsterdam, April 19/20, 2008

  http://weitz.de/eclm2008/

Real email: (replace (subseq ·········@agharta.de" 5) "edi")
comp.lang.lisp / 2008 / jan