🚀 go-pugleaf

In <hjbD3GvD2.6z2@netcom.com> hjb@netcom.com (Hwa-Jin Bae) writes:

>it's one thing to alpha-release something.  it's quite another to
>purposefully do what cygnus does.  i have confirmation of this from cygnus
>personnel, their current workers.  cygnus isn't simply delaying
>distribution of some of their software for alpha-testing purpose. they're
>doing it for marketing, promotion, and money.  i think they need not resort
>to doing so, but they do.

Disclaimer: I am a very happy Cygnus customer.

The reality is the Cygnus does a huge amount of work on the 
development tools.  If you're really hell bent on having access
to yesterday's version of the tools and can convince the right
people at Cygnus that you won't beat them up if it doesn't work
(or better yet, fix it if it doesn't work), there is precedence
for that happening.    

It's wildly unrealistic to think that each day the tree is tested
on the dozens (hundreds?) of host/target combinations.  When they
do have enough combinations that have had suitable air-time, they
release them.  Yes, as a customer, I'm sure that I serve as a test
ground for the public releases.  I have a very controlled system,
and they have a controlled number of customers and platforms.  After
we've knocked on it for a while, it goes to the public at large.

Once upon a time, when I was (in an admittedly small way) helping 
make the tools go on my embedded system, I needed the chain-of-the-
[day|week].   Now that the stuff works (and my schedule doesn't say
"work on GNU tools" anymore), I just want to be able to send a 
bugreport in and get a patch with a day or so.   That's exactly
the model I needed.

Note, I'm not speaking for Cygnus, and anything they say certainly
gets precedence over what I just said.    



--
Robert Lipe    Sr. Software Engr.    Arnet Corp.     robertl@arnet.com