This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: GCC Release Status (2004-05-23)


Mark Mitchell wrote:
[...]
> GCC 3.5
> =======

I feel that this release should really be named 4.0. I do know
that it is a Steering Committee decision and not just yours, but
I wanted to voice it publicly nevertheless.

The internals of GCC have changed quite a bit, gfortran
will be provided instead of g77, (likely) a new binary-compatibility
API for Java, Ada might not initially be available, etc. - quite
a few user-visible changes to justify a major version bump, IMHO.


> Target stage 2 date: July 1st.
> 
> The tree-ssa merge is complete, and starting to settle.
> 
> I plan to close Stage 1 on or about July 1st.  The tree-ssa changes
> are plenty of change for one release cycle, and going through the
> release process will force us to knock a lot of the bugs out of
> tree-ssa.

May I dare suggest that we give a bit of more time for
Stage 1 for this release? Some of the reasons:

1. Tree-SSA was merged only recently and exposed to a much
   wider set of hackers. Give them some time to get familiar
   with it and possibly contribute new optimisation passes,
   etc.

2. Perhaps some of the redundant RTL passes can be removed
   now.

(The point of #1 and #2 is that this is *the* time when we
can clean up the internals of GCC to a large extent. This
does not seem possible in just a month. It will also
possibly attract many more talented hackers towards GCC.)

3. Many of you will be out of action for almost a week due
   to the 2004 GCC Summit.

4. Perhaps Ada (and Treelang) will get
   function-as-trees/gimplification. Or the effort might
   expose shortcomings in the GENERIC representation that
   might be useful to incorporate before essentially being
   standardised.

5. Perhaps LNO can also be merged in.

etc.

I know that as a Release Manager you'd say that we'd
all like to have some more time and will never get
around to actually pushing out a release at this rate,
but even you must agree that the situation this time
around is a bit different from other GCC releases of
the near past.

I also suspect that 3.4 will have a much longer active
life (among users) than any other 3.x release series.

Ranjit.

-- 
Ranjit Mathew          Email: rmathew AT gmail DOT com

Bangalore, INDIA.      Web: http://ranjitmathew.tripod.com/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]