Re: [avr-libc-dev] [bugs #10329] crtm168.o is not being built

[Theodore A. Roth]

I don't see the binutils patch in your attachments. Are you applying
that too? If not, then that might be causing the checks to fail.

On Wed, 8 Sep 2004, E. Weddington wrote:

> Which config.* file is that again?

What you sent was what I wanted to see.

>
> Below is the beginning of my make output (is this what you wanted?). It
> shows that avr-gcc does not support any of the new devices.
>
> This still could be a build problem on my part. My apologies if it turns
> out to be this way.
>
> -----------------------------------------------------

<snip>

> checking if avr-gcc   has support for attiny13... no
> checking if avr-gcc   has support for attiny2313... no
> checking if avr-gcc   has support for atmega48... no
> checking if avr-gcc   has support for atmega88... no
> creating ./config.status

<snip>

> checking if avr-gcc   has support for atmega168... no
> checking if avr-gcc   has support for at90can128... no
> creating ./config.status
> creating Makefile

That was specifically what I wanted to see. If you look into the
config.log file in the avr5 build dir, I suspect that you will see
somthing like this:

configure:1662: checking if avr-gcc   has support for at90can128
configure:1671: avr-gcc   -c -mmcu=at90can128  conftest.c 1>&5
unknown MCU `at90can128' specified
Known MCU names:
   avr2
   at90s2313
   at90s2323
   at90s2333
   at90s2343
   attiny22
   attiny26
   at90s4414
   at90s4433
   at90s4434
   at90s8515
   at90c8534
   at90s8535
   at86rf401
   avr3
   atmega103
   atmega603
   at43usb320
   at43usb355
   at76c711
   avr4
   atmega8
   atmega8515
   atmega8535
   avr5
   atmega16
   atmega161
   atmega162
   atmega163
   atmega169
   atmega32
   atmega323
   atmega64
   atmega128
   at94k
   avr1
   at90s1200
   attiny11
   attiny12
   attiny15
   attiny28
conftest.c:0: error: MCU `at90can128' supported for assembler only
configure: failed program was:
#line 1664 "configure"
#include "confdefs.h"

int main() {
int main (void) {}
; return 0; }



I just noticed that the failed program is a bit wrong. It should just be
this:

int main() {
; return 0; }

Here's the patch I just committed to HEAD to fix this:

@@ -287,7 +287,7 @@ AC_DEFUN(
     AC_MSG_CHECKING(if ${CC} has support for $1)
     AC_TRY_COMPILE(
       [],
-      [int main (void) {}],
+      [],
       [has_dev_support=yes],
       [has_dev_support=no]
     )


---
Ted Roth
PGP Key ID: 0x18F846E9
Jabber ID: address@hidden