-----BEGIN PGP SIGNATURE-----
Version: 2.6.3ia+4K

iQEVAgUBP4bCvVuYv80UoVTdAQHJfAgApTXZbJt0Xqp13PYp2TkJhNHOapLT+45e
hlof1aQ70ymsxNJ10804229LSWGfKecLvAPanEdvXFUZHrDvmq2y64PRAYNRPWdn
d5WyAlUHGOEdgVXgQbmasPHwVTtzss5jwuAAbxyjmwxxPv3lqDmunTUD0WTAD9c7
p4esiYK0CBHAbGZHDoFqeAzRRlNkZm9xVJtVeIkJG4Guoo69tg3E9BCDWlwYHe4O
sKEAtRJ45/DGiVVhNK8/V45ck0BME+feQfxpdlMpYcY6h9Q2ewqU/ERxYzoHZQpB
oaipkA0/6llmuClxpvzmuivpVhEm+P9NrWIH7WmuLcyNOCvEPQMR2w==
=vkpp
-----END PGP SIGNATURE-----

On Fri, Oct 10, 2003, Michael Schloh von Bennewitz wrote:

> [...] the real problem is that FSL currently
> doesn't review its l2 sections for duplicate entries. Some additional
> possible improvements include:
> 
>   1. Add logic to detect duplicate l2 sections (just compare the 'ident'
>      tokens), and FSL will complain this to the debug log and possibly the
>      channel corresponding to the l2 section that FSL choses to use.
> 
>   2. Change the config process logic to use all the duplicate l2 sections
>      equally, writing to their conflicting channels sequentially.
> 
Option 2 describes how fsl was designed. If it behaves differently it's
a bug.

>   3. Write a new fsl-checkconf(8) that reads through its valid (according to
>      cnfdir/FSL_PREFIX.*) config files, parses them but does nothing else.
>      It returns detailed errors for problems such as duplicate l2 sections.
> 
We should have a fsllogger(1) and a linter or config checker in the
future.

--
Thomas.Lotterer@cw.com, Cable & Wireless