This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Hopeful patch for configure breakage

On Thu, Mar 15, 2001 at 08:29:52PM -0800, Mark Mitchell wrote:
> >>>>> "Zack" == Zack Weinberg <zackw@Stanford.EDU> writes:
>     Zack> Yes, but does it cause trouble in practice?  The
>     Zack> alternatives we have are (a) back out the patch (something I
>     Zack> would prefer to avoid, but it may be necessary), (b) figure
>     Zack> out a way to compile and run a program on the build machine
>     Zack> (which can be done, but would be a pain).
> Is that hard?  That's what configure does all the time, right?  We
> already know what the host C compiler is, so I can't imagine this
> would be that hard...

It turns out that we only calculate the host C compiler in the
Makefile (specifically, in build-make).  configure doesn't know what
it is.  We *could* do it in configure but that would be a major change
on top of the existing problems.

>     Zack> Or - it just occurred to me - we could run the object file
>     Zack> through od(1).
> That's a good idea, if it exists on most machines.

The manpage archive at implies that it was in V7, which is
good enough for me.  I'll work up a patch.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]