gcc newsletter #14

[Mathieu Lacage <Mathieu dot Lacage at sophia dot inria dot fr>]

hi all,

As usual, http://gccnews.chatta.us or bellow for a text-only version.

regards,
Mathieu

gcc release status
------------------

Of course, while I was away trying to fix the roof of my new apartment, 
Mark Mitchell sent a new release status[0], froze[1] the 3.4 branch and 
released the first release candidate[2] for gcc 3.4.3.

Giovanni Bajo would like to see a clearer release criateria[3] for gcc 4.x.

gcc development
---------------

Sometimes, specific versions of gcc show major speed regressions. This 
3x slowdown[4] is one such great example. I found the discussion[5] on 
where it comes from pretty interesting.

What kind of langage extensions can be accepted for inclusion in gcc 
and how they can be accepted is a recuring topic which was, once more, 
discussed[6] on the gcc mailing list after a request[7] for c++ nested 
functions. Everyone agrees that only serious extensions should be 
integrated: this means that they need to be designed such that 
interaction with other langage features are clearly defined and such 
that they are compatible with the existing langage standard.

Despite a lot of rather depressing test results a few months ago, -O0 
seems to be getting faster as shown by these results from Karel Gardas[8] 
and from Richard Guenther[9].

The lengthy discussion (some would call that a flamewar) about the way 
the Ada frontend is maintained has had some influence on the way the 
Ada maintainers work since they have announced[10] that they will try 
to follow more closely the gcc contribution rules.

An ambitious proposal on how to change the architecture of gcc has been 
posted[11] by Geoff Keating. Of course, one could wonder[12] why LLVM 
would not be used as a basis for such a work. Licensing might not be an 
issue[13] so only the technical problems might need discussion but so far, 
nothing constructive has been said about them.

The tree-ssa integration seems to draw[14] a few contributors interested 
in new optimization passes.

[0] http://gcc.gnu.org/ml/gcc/2004-10/msg01005.html
[1] http://gcc.gnu.org/ml/gcc/2004-10/msg01257.html
[2] http://gcc.gnu.org/ml/gcc/2004-11/msg00006.html
[3] http://gcc.gnu.org/ml/gcc/2004-10/msg01229.html
[4] http://gcc.gnu.org/ml/gcc/2004-10/msg00755.html
[5] http://gcc.gnu.org/ml/gcc/2004-10/msg01034.html
[6] http://gcc.gnu.org/ml/gcc/2004-10/msg00892.html
[7] http://gcc.gnu.org/ml/gcc/2004-10/msg00861.html
[8] http://gcc.gnu.org/ml/gcc/2004-10/msg00952.html
[9] http://gcc.gnu.org/ml/gcc/2004-10/msg00955.html
[10] http://gcc.gnu.org/ml/gcc/2004-10/msg00953.html
[11] http://gcc.gnu.org/ml/gcc/2004-10/msg01073.html
[12] http://gcc.gnu.org/ml/gcc/2004-10/msg01121.html
[13] http://gcc.gnu.org/ml/gcc/2004-10/msg01124.html
[14] http://gcc.gnu.org/ml/gcc/2004-10/msg01193.html