[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Problem is CC environment var. (was Re: Something screwed up in PROJECT_ROOT or PROJECT_TOPdir)
- Date: Mon, 21 Feb 2005 06:21:54 +0100
- From: Ralf Corsepius <ralf dot corsepius at rtems dot org>
- Subject: Re: Problem is CC environment var. (was Re: Something screwed up in PROJECT_ROOT or PROJECT_TOPdir)
On Sun, 2005-02-20 at 08:27 -0500, Peter Dufault wrote:
> RTEMS is now building (at least it started building and has been
> building for a while now) after I added CC_FOR_TARGET and
> CXX_FOR_TARGET to the configuration.
> My problem is that I wrap my configuration with "CC=gcc34" or
> "CC=gcc33", etc, so I can modify what gets built with what host
> compiler. For example, gdb 5.2 needs gcc33 while I'm building all
> other host stuff with gcc34.
> I just checked the RTEMS configure, and it specifies "CC_FOR_HOST", not
It uses CC_FOR_BUILD, CC_FOR_HOST and CC_FOR_TARGET in its toplevel
CC is used locally and is being set to one of CC_FOR_BUILD, CC_FOR_HOST
or CC_FOR_TARGET, depending on whether a configuration subdir is being
built for build, host or target.
If you set CC in the environment, this will override the CC being using
inside of a Makefile when invoking make in a subdirectory.
> Maybe someplace in multilib it must decide to use "CC" as the
> target compiler if it is set and "CC_FOR_TARGET" isn't?
Not quite. The toplevel configure script sets CC to one of CC_FOR_*.
> I'm pretty
> sure this is new behavior since I think I've always done this, but I
> won't swear to it.
Just do not set CC in the environment. Setting CC in the environment
means asking for trouble - In all configurations, not only RTEMS.