This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: egcs-1.1.2 status
- To: law at cygnus dot com
- Subject: Re: egcs-1.1.2 status
- From: Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>
- Date: Wed, 10 Mar 1999 10:35:55 +0100
- Cc: manfred at s-direktnet dot de,mh at exept dot de,egcs at egcs dot cygnus dot com
- References: <Your message of Tue, 09 Mar 1999 14:33:41 MST. <4.2.0.25.19990309135509.03adb1a0@mail.lauterbach.com>
At 07:08 10.03.99 , Jeffrey A Law wrote:
> In message <4.2.0.25.19990309135509.03adb1a0@mail.lauterbach.com>you write:
> > Not really :-). Currently enable-cpp=/lib will break if you try to set
> the
> > installroot via "make prefix=/var/tmp/egcs-root/usr install" (the
> usual way
>The pathname provided by enable-cpp is an absolute pathname -- ie, it is
>not relative to $prefix. That's how it is supposed to work.
Yeah, but the last time I asked for $install_root I was told to use $prefix
for this purpose :-). This is now no longer true.
> > I would suggest to make it at least possible to pass cpp_install_dir from
> > the toplevel make, that would solve it for now.
>I don't think we need this right now.
Hmm, it would make it possible to use the following construct:
make prefix=$RPM_BUILD_ROOT/usr cpp_install_dir=$RPM_BUILD_ROOT/lib install
I think this makes it quite clear, what I'm trying to do. Sure I can copy
cpp in the spec file, but I like the above solution more.
> > Take this as a vote to implement a $install_root variable during the
> > configure/Makefile overhaul.
>We can put it on the list of things we want. This adds quite a bit of
>complexity though.
True, but it will also make things clearer for the user.
Franz.