Re: Readline patch

[Marco Gerards]

Jeroen Dekkers <address@hidden> writes:

> On Sun, Nov 30, 2003 at 10:13:29PM +0100, Yoshinori K. Okuji wrote:
> > On Sunday 30 November 2003 19:36, Jeroen Dekkers wrote:
> > > I think that checking the version of gcc is the best way.
> > 
> > Generally speaking, using a version number for checking software is a 
> > bad idea. If one uses a unreleased version (e.g. by checking it out 
> > from a CVS repository), what happens?
> 
> Generally speaking, you are right. You should test for features and/or
> bugs. But I'm not sure if it's possible to write an easy autoconf test
> for this. If it's possible, we should do that of course. But if it
> isn't, we should check the version number. AFAIK the gcc CVS version
> is just the version of the next release of that branch.

Well, I'm quite sure it is possible to write such test.  Why do we
have to bother which version can do what anyway?  (I won't test every
version).

And what about future releases of gcc?  AFAIK gcc still has this bug,
how can we be compatible with future gcc releases if we do not know
when it will be fixed?

--
Marco