Re: Current statu of ASIS for GNAT 2012, please ?

[Ludovic Brenta <ludovic@ludovic-brenta.org>]

"Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr> writes:
> Le Sat, 01 Oct 2011 11:11:35 +0200, Ludovic Brenta
> <ludovic@ludovic-brenta.org> a écrit:
>> Again, why do you keep trying to find other, more difficult solutions to
>> your problem?
> I may have to give a summary of the history so far: as pointed in this
> thread, every thing was fine, except for AUnit, which could be
> compiled, but which gave me errors as soon as I wanted to build an
> AUnit application. I did not want a more difficult solution, I just
> wanted AUnit working.

Ah, you said "just".  This is the sure sign of a problem.  No, you don't
"just" want AUnit, you want GCC 4.6 plus ASIS 2010 plus SPARK plus AUnit
working together.  Good news: Debian unstable provides the first three
today.  Bad news: AUnit has not yet migrated to gnat-4.6 in the same
Debian unstable.  Conclusion: the easiest solution to your problem is to
install Debian unstable and recompile AUnit (an only AUnit) with
gnat-4.6.  You do not need to recompile GCC, ASIS or SPARK.

> I tried, as you suggested, to install a Debian Sid within a chroot
> environement, which was not better: there is either libaunit3, which
> does not come with required AUnit files such as aunit.gpr, or either
> libaunit1-dev, which provides a aunit.gpr among others required files,
> but which unfortunately depends on GNAT 4.4... and bahm, conflict.
>
> As this thread seems to drive you a bit nervous (a guess from your
> wordings), I will not tell more here, about future evolution of the
> story, to help keep the climate here, peaceful.

My words may be a bit strong.  This is because you do not seem to be
accepting the solutions I propose for your problem.  Instead, you invent
new problems like "how do I recompile GCC 4.6".

> Except just a last side note: I've submitted today a bug report to
> Debian, about some wrong things with a GCC source package from Debian
> Testing.  I straight away get a reply asserting I did not applied the
> required patches (but I did), and the report was invalidated (did they
> really tested ?).  Interestingly, a similar report was submitted to
> FSF too, by someone else, few months ago, the exact same. Was rejected
> too and not taken into account for some other excuses (that's probably
> why the bug is still there months after), pretending the bug does not
> exist (while multiple people on the web have testified of the same
> bug). If really the Open Source world want people to submit bug
> reports, one good idea may be to acknowledge bug reports, may be as
> long time pending bug if there is no other way. Always the same
> stories, either excuses to not take the bug into account, either
> excuse like "it's not a bug, it's a feature", and so on. Finished with
> bug reports for me. As that seems to serve nothing in any places, I
> will not bother any more now. Just saying this here as I know you are
> a Debian maintainer, if ever that can help you and your colleagues
> understand why many people don't bother submitting bug reports.

The goal of Debian is not to get more bug reports.  It is not my
personal goal either to get more bug reports.

Also, if you think you have found a bug, the burden of proof is on you,
the bug submitter.  You must prove the existence of the bug by showing
evidence that you did everything correctly and followed all the
instructions.  Just saying "I followed the instructions" is not enough;
you must prove you did so by showing the exact commands you ran and the
exact output you got.  Furthermore, you have to state your expectations
explicitly and prove they were reasonable, because sometimes the bug is
in your expectations and not in the software.  Only then can you file a
valid bug report.  A lot of people are "driven" away from free software
simply because they do not accept this burden of proof.  I am not
worried about the "loss" of bug reports from such people.

There is a funny tutorial about how properly to report bugs here:

http://www.chiark.greenend.org.uk/~sgtatham/bugs.html

-- 
Ludovic Brenta.