autoconf-patches
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: C++ fix and test


From: Akim Demaille
Subject: Re: C++ fix and test
Date: 20 Dec 2000 16:41:31 +0100
User-agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Channel Islands)

>>>>> "Alexandre" == Alexandre Oliva <address@hidden> writes:

Alexandre> On Dec 20, 2000, Akim Demaille <address@hidden> wrote:
>>>>>>> "Alexandre" == Alexandre Oliva <address@hidden>
>>>>>>> writes:
Alexandre> On Dec 20, 2000, Akim Demaille <address@hidden> wrote:
>>>> Why do you want to try without stdlib.h?

Alexandre> Because stdlib.h may not be safe for multiple inclusion,
Alexandre> and the user test may already #include it.

>> Then it means we have to decide if the user must or must not, but
>> `can' and `cannot' are not usable.  I can't see how you'd guarantee
>> that it will work if some test include it and some others don't.

Alexandre> If we provide a declaration of exit(int) compatible with
Alexandre> that of stdlib.h, we'll be fine.

Why do we try to prototype exit?  I'm lost.  There are really some C++
compilers with includes such as stdlib.h which have a chance of
compiling something serious, but don't prototype exit?

I mean, sure there is the std:: stuff, but we can test for `using
std;' and just do that, but if exit is not prototyped in stdlib.h, or
something else goes wrong, how could the package we're configuring
could be completely compiled?  This system is too broken, end of the
story.

I'm honestly lost.  I don't understand the point of all this, and I
can't see how those complications might make it more robust.  Let's
just include stdlib.h, and we will see.  Let's not ask for troubles,
let them come to us.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]