Re: [PATCH] BUG in gw/wapbox.c:dispatch_datagram

I made current fix, so I probably should not vote ;)

Main problem here is: does all current configuration files run without 
changes ? That why I used hard-coded value 10
(Configuration variable max-messages could be later used to drop too 
long (too expensive) messages).

Aarno

On 24.10.2003, at 15:45, David White wrote:

> Whoops.
>
> A tiny little screwup made the patch included in the quoted message
> unusable.
>
> Attached is another try.
>
> Thanks!
>
> Dave WHITE
> ONE GmbH
>
> On Fri, 2003-10-24 at 15:27, David White wrote:
>> Good Hello!
>>
>> We've identified a bug in the STABLE version (1.2.1) involving the 
>> PPG.
>> When messages with length 140-[0-6] arrive at dispatch_datagram in
>> wapbox.c, the length of the existing udhdata (UDH port addressing IE) 
>> is
>> not counted with the length of the msgdata (the WSP Push PDU).
>>
>> This results in malformed Push OTA messages being sent out.
>>
>> Our results with the one-line patch needed to fix this in 1.2.1 have
>> been satisfactory.
>>
>> This bug is fixed in CVS HEAD by hard-wiring the maximum number of
>> messages to 10. The enclosed patch reverts to a corrected version of 
>> the
>> old logic to calculate the number of SMs needed for Push OTA.
>>
>> David WHITE
>> ONE GmbH
>>
>>
> <wapbox-dispatch_datagram.patch>