Tim,
On 2004 Jul 10 , at 01.02, Tim Jenness wrote:
> It's a bit annoying that link stages fail but the make continues. It's
> also annoying that make install is allowed independently of a failed
> make.
> Shouldn't the install target depend on the binary actually being there
> to
> install?
That would seem sensible. But is this behaviour different from
standard? That is, are any of my changes behaving differently from the
targets which unpatched automake generates? If not, then I agree they
probably ought to fail differently, but I'd be very reluctant to change
anything, since in the first place it could potentially be a very large
task, and in the second place it could get in the way of tracking
updates in the main automake sources.
Having said that, it's more than probable that I've added rules which
should fail when links fail, but which don't, in which case I can fix
those promptly. Do you have to hand a dump of this (not) failing?
Norman
--
----------------------------------------------------------------------
Norman Gray : Physics & Astronomy, Glasgow University, UK
http://www.astro.gla.ac.uk/users/norman/ : www.starlink.ac.uk
|