[Ada] ACATS progress between 3.3.3 and 3.4

[Arnaud Charlet]

> The # of failures went down from 24 to 8, which is a very good sign. All the
> difficult problems of tasking (chapter 9, c9) are solved now. The remaining
> 8 errors on 3.4 are all regressions from 3.3, though, which is a bit
> disappointing. 

Right, these are most likely back-end bugs that are not specific to Ada.
The c34 failures for instance have regressed after non Ada changes made
around 2003-10-15 (a few days before the big merge). Since the ACATS were
not integrated at the time, there was no easy way to get people's attention
on these regressions.

> I suppose that the Ada maintainers (ACT) are well aware of the current
> situation and I see no use in reporting the regressions as bugs.

Note that the Ada maintainers at the FSF are a group of individuals.
AdaCore is not the FSF Ada maintainer. AdaCore is the maintainer of the
GNAT Pro technology, and contributes major Ada changes to the FSF.

> What is the intended politics for Ada on the 3.4 branch? I haven't seen any
> check-ins on the 3.4 branch since its creation on Jan 16. 

Same as for the other languages: only critical fixes go into the branch,
knowing that the critical fixes are usually not Ada related (but rather
back-end or middle-end) anyway, even if they fix Ada test cases. There
have actually been a bunch of non Ada patches submitted on the 3.4 branch
that fix problems shown by Ada code.

Arno