No, this is not the bug of GCC. Alhough config/pa/linux-unwind.h
looks incorrect at first glance (TRAMP_SIZE is 9), it runs correctly
(because of padding).
In fact, following program runc correctly on hppa too, we see
"clean it up".
$ gcc -fexceptions -save-temps func.c test-pthread-cancel-3.c -
looks big, but only
inherent change is the change of FFI_CLOSURES, others are caused
by source change between 3.4.1 and 3.4.4.
* debian/patches/m32r-libffi.dpatch: Updated to 3.4.4.
-- NIIBE Yutaka <[EMAIL PROTECTED]> Sat, 24 Sep 2005 17:53:46 +0900
--- gcc-3.4-3.4.4/debian/patches/m3
The patch I sent yesterday was wrong.
Here is update. The intention of this patch is that supporting binary
only NMU by porters.
--
gcc-binary-only-nmu-20050924.patch
Description: Binary data
1:4.0.1-8 GCC support library
Versions of packages gcc-3.4 recommends:
ii libc6-dev 2.3.5-6GNU C Library: Development Librari
* debian/rules.conf (DEB_VERSION): Support binary-only NMU.
* debian/rules.patch (DEB_VERSION): Likewise.
-- NIIBE Yutaka <[EM
otoff,libffi,limits,stack}.dpatch: New files.
* debian/patches/autoreconf.dpatch: Enable again.
-- NIIBE Yutaka <[EMAIL PROTECTED]> Wed, 24 Nov 2004 18:10:14 +0900
diff -u gcc-3.4-3.4.3/debian/patches/autoreconf.dpatch
gcc-3.4-3.4.3/debian/patches/autoreconf.dpatch
--- gcc-3.4-3.4.3/deb
YAEGASHI Takeshi wrote:
> > Is sh a new or a revived port? Is there a reason not go with
> > gcc-3.1/gcc-3.2?
> >
> > No Fortran, no Java?
>
> Sorry I'm not certain about the current status of GNU toolchain. Does
> the ABI change between GCC 3.0 and 3.1 or later?
[...]
> So, it might be
6 matches
Mail list logo