This is the mail archive of the gcc-patches@gcc.gnu.org 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]
Other format: [Raw text]

Re: PR 32272: Error out if building separately but srcdir has build remnants.


* DJ Delorie wrote on Mon, Sep 21, 2009 at 11:57:07PM CEST:
> 
> > If the user built in-tree at some point in time, and things failed to
> > clean up right,
> 
> "Failed to clean up right"?  Can we get a confirmation that "make
> clean" or "make distclean" *does* clean up right?

No (I don't know).  But I can work on it if you like.  :-)

Looking at toplevel makefile though, distclean should remove this
directory.

> Also, if we've configured in the source tree, there would be other
> signs - config.log, Makefile, etc.  We should check for all those
> things (and this one) at the same time.  If you're going to be a
> little paranoid, might as well be a lot paranoid.

Well, existence of config.status will cause configure to complain
already.  The point of this patch is that nothing will complain about
host-$host but the GCC_TOPLEV_SUBDIRS logic will turn out fatally wrong
which causes grief for even not so inexperienced users.  This patch
would make PR 32272 trivial to diagnose.

Anyway, I'm not going to insist.  I just figured if there are five
people complaining about this, it might happen again, and this would
save us (me at least) time analyzing the bug.

Cheers,
Ralf


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