Finally, I managed to fix up libffi and run testing on libjava under
Darwin/i386.
Here is the libffi testsuite output:
Test Run By sandro on Sun Mar 26 10:49:37 2006
Native configuration is i386-apple-darwin8.5.2
=== libffi tests ===
Schedule of variations:
unix
Running
Summary:
GCC 4.1 itself appears to be very stable, both on MIPS and AMD64.
There are, however, a large number of packages using code (especially
C++) which GCC 4.1 treats as errors. Fortunately, most of them are
trivial to fix. By compiling about 6200 packages, over 500 new
bugs have been discov
FLOAT_EXPR and FIX_TRUNC_EXPR are both expensive so they really
should be added to this list.
Take the following code:
int f(float a, int *b)
{
int i;
for(i = 0; i<1000;i++)
*b = a;
}
Compile with -O1 and we get on PPC:
_f:
fctiwz f0,f1
addi r2,r1,-4
stfiwx f0,0,r
config.guess:
i686-pc-linux-gnu
gcc -v:
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../gcc-4.1.0/configure
Thread model: posix
gcc version 4.1.0
/etc/issue:
Welcome to SUSE LINUX 10.0 (i586) - Kernel \r (\l).
uname -a:
Linux bischof 2.6.13-15.8-default #1 Tue Feb 7 11:07:24
I am looking at the source code of gcc.c, and I found the driver function
main.c calls do_spec(). It seems that do_spce() is the function, called by
main(), which invokes the preprocessor, compiler, assembler and linker
individually at different step according to the input to the driver; and
e
> Can someone give a brief intro of what do_spec(), do_spec_x() and
> execute() do, so that I can follow the procedure?
The comments seem generally adequate, but that could just be me :-)
For do_spec ():
/* Process the spec SPEC and run the commands specified therein.
Returns 0 if the spe
On Mar 26, 2006, at 3:56 PM, Ben Elliston wrote:
A seperate but sort of related question is: After looking at
Makefile, I think toplev_main () in toplev.c (which is wrapped in
main() in main.c)is compiled to cc1(i.e, $BACKEND), is it correct?
But why can't I find cc1 in the mygcc/bin director
"cvs-commit at developer dot classpath dot org" <[EMAIL PROTECTED]> writes:
| --- Comment #7 from cvs-commit at developer dot classpath dot org
2006-03-20 20:03 ---
| Subject: Bug 11070
|
| CVSROOT:/cvsroot/classpath
| Module name:classpath
| Branch:
| Changes by:
On Mar 26, 2006, at 5:32 PM, Gabriel Dos Reis wrote:
Why fixing a classpath bug shows up as a commit to fix an unrealted
GCC/C++ bug?
Because he committed with the wrong number.
-- Pinski
On Mon, 2006-03-27 at 03:32 +0200, Gabriel Dos Reis wrote:
> "cvs-commit at developer dot classpath dot org" <[EMAIL PROTECTED]> writes:
>
> | --- Comment #7 from cvs-commit at developer dot classpath dot org
> 2006-03-20 20:03 ---
> | Subject: Bug 11070
> |
> | CVSROOT:/cvsroot
> does not care about cost really and always pulls out the calculation
> so maybe
> LIM should be doing that instead.
>
Zdenek will argue (or at least, has in the past), that pulling out
things should only be done if you can prove that it won't increase
register pressure. I believe that is not
Thanks Joe.
Piyush Garyali
--
My Blog http://verypondycherry.blogspot.com
On 3/25/06, Joe Buck <[EMAIL PROTECTED]> wrote:
> On Fri, Mar 24, 2006 at 03:15:58PM +0530, Piyush Garyali wrote:
> > Is there any solution to fix the troubles which result from the change
> > to the name mangling algorith
12 matches
Mail list logo
