Re: regress and -m64

2006-08-29 Thread Jack Howarth 
I've posted the results of last night's build of gcc trunk
on my dual G5 using Geoff's TImode patch and changes to
prune.exp to suppress the failures from the ld64 linker
warnings.
 Jack

linkage gcc.misc-tests

2006-08-29 Thread Jack Howarth 
Eric,
   Does the following test fail for you under your x86_64 patch set for
Macintel?

Running 
/sw/src/fink.build/gcc4-4.1.999-20060828/gcc-4.2-20060828/gcc/testsuite/gcc.misc-tests/linkage.exp
 ...
Executing on host: 
/sw/src/fink.build/gcc4-4.1.999-20060828/darwin_objdir/gcc/xgcc 
-B/sw/src/fink.build/gcc4-4.1.999-20060828/darwin_objdir/gcc/  "-w" 
-fno-show-column -c  -m64 -o linkage-x.o 
/sw/src/fink.build/gcc4-4.1.999-20060828/gcc-4.2-20060828/gcc/testsuite/gcc.misc-tests/linkage-x.c
(timeout
 = 300)
cc -c  
/sw/src/fink.build/gcc4-4.1.999-20060828/gcc-4.2-20060828/gcc/testsuite/gcc.misc-tests/linkage-y.c
 >&/dev/null
Executing on host: 
/sw/src/fink.build/gcc4-4.1.999-20060828/darwin_objdir/gcc/xgcc 
-B/sw/src/fink.build/gcc4-4.1.999-20060828/darwin_objdir/gcc/ linkag
e-y.o linkage-x.o  -fno-show-column  -lm   -m64 -o linkage.exe(timeout = 
300)
ld64 warning: in linkage-y.o, file is not of required architecture
can't resolve symbols:
  _main, referenced from:
  __start in crt1.o
ld64 failed: symbol(s) not found
collect2: ld returned 1 exit status
compiler exited with status 1
output is:
ld64 warning: in linkage-y.o, file is not of required architecture
can't resolve symbols:
  _main, referenced from:
  __start in crt1.o
ld64 failed: symbol(s) not found
collect2: ld returned 1 exit status

FAIL: gcc.misc-tests/linkage.c link
testcase 
/sw/src/fink.build/gcc4-4.1.999-20060828/gcc-4.2-20060828/gcc/testsuite/gcc.misc-tests/linkage.exp
 completed in 0 seconds

This is under the Xcode 2.3 cctools. I'll try the Xcode 2.4 cctools and see if 
that improves.
  Jack

RE: linkage gcc.misc-tests

2006-08-29 Thread Jack Howarth 
Geoff,
   Isn't the gcc.misc-tests/linkage.c failing just because we don't have
an entry in linkage.exp that defines the native flags for Darwin? Also,
this test look pretty dicey in that it uses...

catch { exec cc -c $native_cflags "$srcdir/$subdir/linkage-y.c" 
>&/dev/null }

and...

catch { exec c89 -c $native_cflags "$srcdir/$subdir/linkage-y.c" 
>&/dev/null }

which makes me wonder how in the world you can be sure it will be executing the
compiler you just built for testing (instead of whatever is in /usr/bin)?
Jack

Re: linkage gcc.misc-tests

2006-08-29 Thread Geoffrey Keating 


On 29/08/2006, at 5:27 PM, Jack Howarth wrote:


Geoff,
   Isn't the gcc.misc-tests/linkage.c failing just because we don't  
have

an entry in linkage.exp that defines the native flags for Darwin?


Yes.


Also,
this test look pretty dicey in that it uses...

catch { exec cc -c $native_cflags "$srcdir/$subdir/linkage- 
y.c" >&/dev/null }


and...

catch { exec c89 -c $native_cflags "$srcdir/$subdir/ 
linkage-y.c" >&/dev/null }


which makes me wonder how in the world you can be sure it will be  
executing the
compiler you just built for testing (instead of whatever is in /usr/ 
bin)?


If you think about the purpose of this test, you'll realise that this  
is correct.




smime.p7s
Description: S/MIME cryptographic signature

Re: linkage gcc.misc-tests

2006-08-29 Thread Eric Christopher 

Jack Howarth wrote:

Eric,
   Does the following test fail for you under your x86_64 patch set for
Macintel?



No, but that's because I have a patch to fix it :)

-eric