acme hates software: Network connectivity may fail

• From: Leon Brocard
  Date: 19:31 on 15 Sep 2007
  Subject: Network connectivity may fail
  
  I have recently been playing with a Vista laptop.
  
  Yes, I know.
  
  Anyway, Skype doesn't work on it in my sister's flat. It does work in
  my flat. It's taken me days to figure out why:
  
    http://support.microsoft.com/kb/934430
  
  "Network connectivity may fail when you try to use Windows Vista
  behind a firewall device"
  
  Well it's a good thing that nobody uses firewalls or routers then, eh?
  
  The fix is to run the following as root^Wadministrator:
  
    netsh interface tcp set global autotuninglevel=normal
  
  Now Microsoft is blaming my router's TCP implementation. That's
  strange, because Linux copes fine with it. And Mac OS X. And Windows
  XP. And the Wii.
  
  See, my thinking is that the TCP stack should be ultra flexible in
  order to make things work. It shouldn't break. It should work. It
  shouldn't automatically tune itself for speed and have a failure mode
  of broken. Tune yourself if you want but failure mode should be
  working but slow. IT SHOULDN'T BREAK.
  
  Grrr hate, Leon
  
  
  

• From: Peter da Silva
  Date: 21:04 on 15 Sep 2007
  Subject: Re: Network connectivity may fail
  
  Man, this is like that one where windows Vista for no good reason  
  sets the broadcast flag on all DHCP requests ... a flag that should  
  only be used when the TCP stack is broken, OR when you're using it as  
  a workaround for networks that have addresses that don't fit in some  
  field in a normal DHCP response.
  
  And what's this "netsh interface" crap? What's wrong with "ipconfig"?
  
  Microsoft decided to re-implement their TCP stack from scratch for  
  the third time, didn't they?
  
  
  
  

• From: David Mackintosh
  Date: 03:37 on 16 Sep 2007
  Subject: Re: Network connectivity may fail
  
  
  --oLBj+sq0vYjzfsbl
  Content-Type: text/plain; charset=us-ascii
  Content-Disposition: inline
  Content-Transfer-Encoding: quoted-printable
  
  On Sat, Sep 15, 2007 at 03:04:48PM -0500, Peter da Silva wrote:
  =20
  > Microsoft decided to re-implement their TCP stack from scratch for =20
  > the third time, didn't they?
  
  This is what happens when you get caught using BSD code and then try
  to re-write it out of embarassment.
  
  --=20
   /\oo/\
  / /()\ \ David Mackintosh |=20
           dave@xxxxxx.xxx  | http://www.xdroop.com
  
  --oLBj+sq0vYjzfsbl
  Content-Type: application/pgp-signature
  Content-Disposition: inline
  
  -----BEGIN PGP SIGNATURE-----
  Version: GnuPG v1.2.6 (GNU/Linux)
  
  iD8DBQFG7JbicwUBd0wDJQQRAtuIAJ9pfyR2oF4QBre3ScGX2ePUc0Th0wCbBgNN
  lVPQMG7Jh2zE1hTCm3uK5iw=
  =y+i5
  -----END PGP SIGNATURE-----
  
  --oLBj+sq0vYjzfsbl--
  
  
  

• From: Peter da Silva
  Date: 04:01 on 16 Sep 2007
  Subject: Re: Network connectivity may fail
  
  On 15-Sep-2007, at 21:37, David Mackintosh wrote:
  > This is what happens when you get caught using BSD code and then try
  > to re-write it out of embarassment.
  
  I wish. The only BSD code in Microsoft TCP by the year 2000 was user- 
  side. They were already producing lots of juicy hateful bugs in their  
  reimplementation by then.
  
  
  

• From: Michael G Schwern
  Date: 05:06 on 16 Sep 2007
  Subject: Re: Network connectivity may fail
  
  Peter da Silva wrote:
  > Man, this is like that one where windows Vista for no good reason sets
  > the broadcast flag on all DHCP requests ... a flag that should only be
  > used when the TCP stack is broken
  
  Sounds like a fair assumption, this is Windows we're talking about.
  
  
  
  

Generated at 10:26 on 16 Apr 2008 by mariachi