[sldev] Endian bug in 1.15.0.2 message template reader?

John Hurliman jhurliman at wsu.edu
Tue May 15 15:18:43 PDT 2007


Paul TBBle Hampson wrote:
> On Mon, May 14, 2007 at 11:18:01AM -0700, Phoenix wrote:
>   
>> On 2007 May 13, at 17:14, Paul TBBle Hampson wrote:
>>     
>
>   
>> Since 1.15.0.2 works fine on my G5, I suspect a misconfiguration in
>> your build environment, but if you find anything wrong in linden code
>> in that area please let us know so it can be corrected.
>>     
>
> I don't suppose you'd be willing to send me a binary, or try rebuilding
> my source (just updated to 1.15.1.3 with the same problem, available at
> http://www.tbble.net/debian/slviewer/) although the latter option only
> really makes sense if you happent to be running Debian or Ubuntu or
> another dpkg-based distro. (If you want to save some traffic, the
> .orig.tar.gz is the pristine Linden source archive)
>
> Before I can diagnose this much further, I need to modify the packet
> dump routine to dump the un-zerocoded version of the package, because
> every time I try to compare the output to the packet dump to the message
> template file, I seem to end up looking at completely the wrong bytes.
>
> Either that, or I've misunderstood something... A variably-repeated
> block has one byte indicating number of repeats, right?
>   

Unless it is a "Variable 2", then it has two bytes indicating the number 
of blocks. Variable 2 isn't as common for blocks as it is for fields 
though, but I think it does happen in some places. It's quite common for 
fields that are just raw binary data though.

John Hurliman


More information about the SLDev mailing list