Experimental Physics and
Industrial Control System

Till Straumann <[email protected]> · Mon, 18 Nov 2002 12:19:55 -0800

M.C.Shao wrote:

----- Original Message -----
From: "Till Straumann" <[email protected]>

To: "M.C.Shao" <[email protected]>

Cc: "Andrew Johnson" <[email protected]>; "Tech-talk" <[email protected]>

Sent: Monday, November 18, 2002 6:29 PM

Subject: Re: building R3.14 problem for RTEMS

This is configured in epics/base/configure/os/CONFIG.Common.RTEMS
where RTEMS_BASE is set.
-- Till

I don't think so , I have cheked that configure file, and RTEMS_BASE is set to be "/opt/rtems" where the rtems is installed on my computer.The problem still exists.

Could be that you're fighting just one more head of the 'autotool' hydra:
CONFIG.Common.RTEMS expands the makefile variable CC_FOR_TARGET.
CC_FOR_TARGET is set in your RTEMS BSP's Makefile.inc (look in the RTEMS
installation tree, _not_ EPICS).
Makefile.inc is created when you build RTEMS. 'c/src/make/configure'
is responsible for the relevant expansion:
Makefile.inc will end up defining
CC_FOR_TARGET = <whatever the CC environment variable was set to when
you configured RTEMS>
I.e.:
you probably configured RTEMS in the following way:
configure CC=/opt/rtems..../
Instead, you should add the proper directory to your PATH and
configure CC=i386-rtems-gcc
(Note that it is _very_ advisable to use 'target'-gcc rather than
the 'gcc' alias in some 'target' subdir!!!)
You can chop that head off by simply editing 'Makefile.inc'
(correct the CC_FOR_TARGET & friends definitions) but remember
that the hydra will grow a new head next time you 'make' RTEMS
- unless you reconfigure.
HTH
-- Till

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System