Re: what DOES the GPL really say?

[David Kastrup <dak@mailhost.neuroinformatik.ruhr-uni-bochum.de>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 2301 bytes --]


kenner@lab.ultra.nyu.edu (Richard Kenner) writes:

> In article <slrn5rpshu.fcf.Olivier.Galibert@iria.mines.u-nancy.fr> Olivier.Galibert@mines.u-nancy.fr (Olivier Galibert) writes:
> >In article <5pim4l$5m3$1@news.nyu.edu>, Richard Kenner wrote:
> >>There's absolutely nothing wrong with pointing out that it's harmful
> >>to the GNAT community for pre-releases to be widely distributed.  The
> >>FSF itself (and Richard Stallman personally) has an extremely strong
> >>similar view about pre-releases of GCC.
> >
> >With the immediate side effect that most people won't even try to fix
> >bugs which may already have been fixed. And most of the time don't even
> >bother to report them because of the "I'll try again in 6 months with
> >the new release and then I'll see" effect.
> 
> This isn't a prudent course of action, since most bugs on a mature
> program such as GCC tend to be ones that have *not* already been
> discovered.

With all due respect, please look at the bug report data base.  More
than half of the bugs I reported since gcc 2.5.something are still
pending correction.  It might well be that they have been fixed for
years now, but the last major release of gcc (2.7.0) was made more
than two years ago.

Actually, for a few of them I got a message, I believe some time last
year, "this will be fixed in 2.8.0".

Considering the amount of bug reports piled up for gcc-2.7 already,
when gcc-2.8.0 is supposed to be vastly overhauled in quite a few
areas, I don't think it actually very useful anymore to write bug
reports for a software which has undergone major changes in the last
few years.

Note that this applies largely to the g++ frontend which has been
under the heaviest changes in the last few years.  The backend,
fortunately, is somewhat more stable.  I got it to produce illegal
code in some of my bug reports as well, but it might have got fed bad
information from the C++ frontend.

So if you manage to get bad code or error messages or so from
compiling ANSI C, or perhaps Ada, a bug report might not be as
hopeless as for g++.


-- 
David Kastrup                                     Phone: +49-234-700-5570
Email: dak@neuroinformatik.ruhr-uni-bochum.de       Fax: +49-234-709-4209
Institut f�r Neuroinformatik, Universit�tsstr. 150, 44780 Bochum, Germany