What's with all the Cisco stuff?

J.T. Conklin jtc@redback.com
Fri Aug 13 12:35:00 GMT 1999


>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:
>> What's the deal with all the Cisco-specific stuff ending up in GDB?

Andrew> I'll ignore the politics :-)

But since you brought it up... :-(

It is not my intent to use my participation/contributions in GDB to
thwart competitors, but I will not hesitate to complain when I think
poor technical judgement is being exercised when code is integrated
into GDB. Note that I did the same thing with the HP TUI code, and
I have no relationship whatsoever with HP.

It was said that GDB should have a facility to examine kernel objects
for various RTOS.  I do not disagree with this.  But after using GDB
on many RTOS, including several at Cisco, I will argue that GDB *has*
a facility to examine kernel objects --- One that doesn't need more C
code or a mechanism for generic target "queries".

I also have to ask, why the rush to integrate code that isn't ready?
The TUI doesn't work on non-HP machines.  The mutant remote protocol
could have waited until the infrastructure was in place to properly
stack target vectors, and KOD could have waited until it had the
benefit of review of a larger pool of embedded systems users.

The latter is particularly bothersome.  I think it is all to easy for
Cygnus folks to keep what should be public discussion inside when it
should be inclusive.  I didn't see it myself when I was at Cygnus, 
but it is particularly obvious from the outside.  

I'm familiar with the concept of getting things in, and letting them
mature.  But I've seen oh so many cases where long term costs far
overshadow any short term gains.  

        --jtc

-- 
J.T. Conklin
RedBack Networks


More information about the Gdb mailing list