[arch-dev-public] binutils-2.20 toolchain rebuild in [testing]
Hi The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes: kernel-headers - bump to latest binutils - bump to latest glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build. I will call for a signoff a few days after getting the x86_64 builds done. Allan
On Mon, Oct 19, 2009 at 10:07 AM, Allan McRae <allan@archlinux.org> wrote:
Hi
The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes:
kernel-headers - bump to latest
binutils - bump to latest
glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils
gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build.
I will call for a signoff a few days after getting the x86_64 builds done.
Allan
FYI, kernel-headers would be a candidate for the any arch. However, if you do the switch, then you'll need to release the toolchain for both arches at the same time otherwise it might break some stuff.
On Mon, 2009-10-19 at 15:45 -0400, Eric Bélanger wrote:
On Mon, Oct 19, 2009 at 10:07 AM, Allan McRae <allan@archlinux.org> wrote:
Hi
The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes:
kernel-headers - bump to latest
binutils - bump to latest
glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils
gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build.
I will call for a signoff a few days after getting the x86_64 builds done.
Allan
FYI, kernel-headers would be a candidate for the any arch. However, if you do the switch, then you'll need to release the toolchain for both arches at the same time otherwise it might break some stuff.
I wouldn't be too sure about that. Have you diffed extracted tarballs of kernel-headers for both architectures and did it return only .PKGINFO related differences? Reason for asking this is the /usr/include/asm directory. Though these files are just stupid header files, the asm directory is usually taken from architecture-specific includes.
On Mon, Oct 19, 2009 at 4:11 PM, Jan de Groot <jan@jgc.homeip.net> wrote:
On Mon, 2009-10-19 at 15:45 -0400, Eric Bélanger wrote:
On Mon, Oct 19, 2009 at 10:07 AM, Allan McRae <allan@archlinux.org> wrote:
Hi
The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes:
kernel-headers - bump to latest
binutils - bump to latest
glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils
gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build.
I will call for a signoff a few days after getting the x86_64 builds done.
Allan
FYI, kernel-headers would be a candidate for the any arch. However, if you do the switch, then you'll need to release the toolchain for both arches at the same time otherwise it might break some stuff.
I wouldn't be too sure about that. Have you diffed extracted tarballs of kernel-headers for both architectures and did it return only .PKGINFO related differences? Reason for asking this is the /usr/include/asm directory. Though these files are just stupid header files, the asm directory is usually taken from architecture-specific includes.
I had the same thought and had done a diff for kernel-headers-2.6.30.5-1. Only the .PKGINFO were different. It's possible that this will be changed for future updates though. Maybe it would be safer to keep it arch dependent.
Eric Bélanger wrote:
On Mon, Oct 19, 2009 at 4:11 PM, Jan de Groot <jan@jgc.homeip.net> wrote:
On Mon, 2009-10-19 at 15:45 -0400, Eric Bélanger wrote:
On Mon, Oct 19, 2009 at 10:07 AM, Allan McRae <allan@archlinux.org> wrote:
Hi
The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes:
kernel-headers - bump to latest
binutils - bump to latest
glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils
gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build.
I will call for a signoff a few days after getting the x86_64 builds done.
Allan
FYI, kernel-headers would be a candidate for the any arch. However, if you do the switch, then you'll need to release the toolchain for both arches at the same time otherwise it might break some stuff.
I wouldn't be too sure about that. Have you diffed extracted tarballs of kernel-headers for both architectures and did it return only .PKGINFO related differences? Reason for asking this is the /usr/include/asm directory. Though these files are just stupid header files, the asm directory is usually taken from architecture-specific includes.
I had the same thought and had done a diff for kernel-headers-2.6.30.5-1. Only the .PKGINFO were different. It's possible that this will be changed for future updates though. Maybe it would be safer to keep it arch dependent.
I have considered this and did a diff on the packages and came to same conclusion as you. But I did not want to assume that this will always be the case... screwing up the toolchain is not something I want to do! Also, it is not a real "any" package. I believe the package is different on e.g. ppc (asm stuff that Jan mentioned), although I have not checked... So, I think I will leave this package as is. Allan
I'd wanted to help out with the x86_64 builds but your glibc patchset is missing (glibc-2.10.1_20091018.tar.bz2). -Andy
Andreas Radke wrote:
I'd wanted to help out with the x86_64 builds but your glibc patchset is missing (glibc-2.10.1_20091018.tar.bz2).
x86_64 builds and all sources are now uploaded. Allan
On Tue, 2009-10-20 at 00:07 +1000, Allan McRae wrote:
Hi
The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes:
kernel-headers - bump to latest
binutils - bump to latest
glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils
gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build.
I will call for a signoff a few days after getting the x86_64 builds done.
Allan
Maybe I'm a bit late with this, but could you also split the fortran and objc libs to their own package or leave them in gcc-libs? The compilers for these languages take up 9MB and 18MB of the total package, where the libs are small.
Jan de Groot wrote:
On Tue, 2009-10-20 at 00:07 +1000, Allan McRae wrote:
Hi
The binutils-2.20 toolchain rebuild is in [testing] for i686. I will do the x86_64 rebuilds in the next day or two. A summary of changes:
kernel-headers - bump to latest
binutils - bump to latest
glibc - grab upstream patchset for proposed future stable release, fix overflow bug (FS#16253), patch to build against latest binutils
gcc - bump to latest, use package spitting (gcc-libs, gcc, gcc-fortran, gcc-objc), move static libraries from gcc-libs to gcc, add gcc-ada package, do not run fixincludes during build.
I will call for a signoff a few days after getting the x86_64 builds done.
Allan
Maybe I'm a bit late with this, but could you also split the fortran and objc libs to their own package or leave them in gcc-libs? The compilers for these languages take up 9MB and 18MB of the total package, where the libs are small.
I agree here and had been thinking the same thing. Not only are the compiler packages relatively large, they also require gcc (67MB). I will look into moving these when I get a chance (before I call for a signoff). I will leave the ada libraries where they are and reassess when we get apps that actually depend on them. Allan
participants (4)
-
Allan McRae
-
Andreas Radke
-
Eric Bélanger
-
Jan de Groot