GDB 6.4 schedule

[Andrew Cagney]

GDB 6.4, unlike some of our more recent releases, is faced with both a 
technical requirement and a very tight schedule:

- it must support full DW_OP_piece
GCC 4.0 is very aggressive at generating DW_OP_piece information.  If 
GDB is to to remain relevant as a debugger for that compiler we're going 
to need to fully support that feature.  If we don't we regress, just 
like we did in the past when we didn't properly support CFI.

- it must be out [well] before GCC 4.0
As a corollary, there is no point in releasing a GDB supporting 
DW_OP_piece after 4.0.  Nor is there a point to releasing just before. 
We need to make that GDB available as early as possible and that 
suggests trying to align ourselves with the 4.0 branch schedule.

The key point, and the one that differentiates this feature from others, 
is that if we fail to address this and stand still, we go backwards.

Looking across at http://gcc.gnu.org/develop.html 4.0 is due ``early 
2005'' and sniffing around suggests a preference for the branch being 
cut by end-of-year with early-to-mid '05 release (Daniel, with close 
ties to MarkM, probably has better insider information).  Even allowing 
for slip, that only drags things into January.

To ensure that GDB matches this schedule and is available at least in 
the 4.0 branch timeframe we're forced to take an even more aggressive 
position, setting our schedule to:

- branch January 2005
- release February 2005

and having a very strong preference for something even earlier.

(This of course also seriously limits changes others were making, as 
once DW_OP_piece is working, we're going to ship.)

Can we agree to this?
Andrew

PS: Hopefully people have been aware of this need for sometime, I 
certainly have.