Re: [Nmh-workers] Patches for FreeBSD port of nmh-1.2

[Scott Blachowicz]

On Mon, Feb 27, 2006 at 04:34:51PM +0000, address@hidden wrote:
> Scott Blachowicz wrote:
> >I've recently updated the "port" in the FreeBSD ports collection to
> >nmh-1.2 and there were a couple of patches I needed to do in order to
> >get it to build cleanly. I figured I would send them here to see if
> >they're already dealt with in the source tree or if others have already
> >seen them.
> 
> Yes, we found out just after releasing 1.2 that it didn't configure
> properly on the BSDs. Fixes are in CVS; if you felt like confirming
> that current head of CVS works on BSD that would be nice. I have
> tested it but only as a check that it builds...

I'll try...but I just got back to work after a week off and am drowning
in unread email and a release due out this week...:)

> >etc/Makefile.in - use $(INSTALL_SCRIPT) to install the scripts instead
> >    of $(INSTALL_PROGRAM) so it doesn't try to run 'strip' on the
> >    installed scripts (which fails, btw).
> 
> ...which would explain why I didn't notice this. However, I don't
> understand why your suggested patch would make any difference.
> According to the documentation of AC_PROG_INSTALL, configure should
> set both INSTALL_SCRIPT and INSTALL_PROGRAM to the same thing
> (ie /usr/bin/install or similar). Checking the configure script
> itself confirms this. So it shouldn't be trying to run strip on
> either the binaries or the scripts.
> 
> Indeed, testing on my Linux system, we run install (without any
> '-s' argument) and it doesn't try to run strip on anything.

In my case, I was seeing a '-s' argument in INSTALL_PROGRAM, but
not in INSTALL_SCRIPT...so that's why I made the patch I did (and,
I assumed, the reason for the distinction between INSTALL_PROGRAM &
INSTALL_SCRIPT).

> So I think this needs more investigation before we consider
> applying that bit of the patch. (The other parts are already
> covered by existing changes in CVS.)

Sure...I'll try to do a little more checking.

Thanx,
-- 
Scott