Brad,
On Thu, 26 Feb 2004, Brad Cavanagh wrote:
> Okay, I've patched dts_setsig.c so that this step proceeds fine. I'll
> email the patch to Norman privately. The documentation for sigaction
> strongly suggests that the patch should work for both Linux and OS X (it
> looks like BSD systems are a little more anal about prototyping than Linux
> systems.
Having looked at this, I think you should just commit this fix
yourself. If you're feeling strong and know anything about autoconf,
you could probably put in a #ifdef HAVE_SYS_TYPES_H at the top with
appropriate changes to configure.ac, but we'd probably be OK without
it.
I did have a look at this file and start autoconfing it, by doing
things like checking for HAVE_SYS_TYPES_H and similar things, but
there's one or two some slightly odd things in there, and it didn't
really seem necessary, so I gave up on it.
Rather oddly, it maintains both a one-argument and a three-argument
(with siginfo_t) signal handler. That seems heroically tolerant of
ancient unixes (Brian is too good-natured, I think), but looks fragile
rather than broken, so is probably best left alone right now.
> However, now make fails on HLP due to a dynamic library problem (the root
> cause of which is the libtool issue). Tim feels that getting rid of
> dynamic libraries will do the trick, as lsthlp built just fine when we
> removed the dynamic libraries in libraries/hlp/.libs/. Is it possible to
> switch to static libraries? This would definitely fix all of the libtool
> issues I've been having the past few days.
I fear this is necessary, in the short term. I'll sort out how best to
configure this this afternoon.
Norman
--
---------------------------------------------------------------------------
Norman Gray http://www.astro.gla.ac.uk/users/norman/
Physics and Astronomy, University of Glasgow, UK [log in to unmask]
|