Re: (no subject)

True,
that is definitely a bug, and should be fixed

but like i said, do not expect kannel to urldecode headers cause http 
headers are not supposed to be urlencoded...

[email protected] wrote:

>
> Hi,
>
> It is not the existence of the '+' sign which causes Kannel to fail, 
> but if I exceed 21 numbers in the
> address.
>
> This seems like a prime DOS attack to me.  All I have to do is call 
> sendsms with 22 character
> 'to' address, and Kannel takes a nose dive.
>
> This is a different issue to the URLEncoding one, which is my 
> misunderstanding of Http Header
> standard.
>
> Regards
> Jason
>
> Konstantin Vayner <[email protected]> wrote on 14/04/2003 
> 06:27:46 PM:
>
> > Why do you assume that HTTP headers should be urlencoded? that is wrong
> > So why would kannel go for wrong behavior in parsing them? :)
> >
> > [email protected] wrote:
> >
> > >
> > > Ki,
> > >
> > > Just to experiment, I tried the following header:
> > >         X-Kannel-To: 413131315325424234+556666724243247
> > >
> > > And it causes the server to crash with the following message:
> > >         2003-04-14 18:21:44 [6] PANIC: gw/smsc/smpp_pdu.def:207:
> > > smpp_pdu_pack: Assertion `octstr_len(p->destination_addr) < 21' 
> failed.
> > >
> > > I was trying to get it to send to multiple recipients, and URLEncoded
> > > the address, (thus the + sign), but apparently
> > > X-Kannel-To does not get UNUrlEncoded?
> > >
> > > I tried it again with the following header
> > >         X-Kannel-To: 413131315325424234 556666724243247
> > >
> > > And all worked as expected.
> > >
> > > Kannel 1.3.1 on Solaris
> > >
> > > Regards
> > > Jason
> >
> >
> >