Message ID | 1396875879-22199-1-git-send-email-gustavo@zacarias.com.ar |
---|---|
State | Changes Requested |
Headers | show |
Dear Gustavo Zacarias, On Mon, 7 Apr 2014 10:04:39 -0300, Gustavo Zacarias wrote: > Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> > --- > .../2.18-svnr23787/glibc-0001-accept-make4.patch | 33 ---------------------- > package/glibc/glibc.mk | 2 +- > 2 files changed, 1 insertion(+), 34 deletions(-) > delete mode 100644 package/glibc/2.18-svnr23787/glibc-0001-accept-make4.patch So we don't do like glibc where we support both 2.18 and 2.19 ? This is really a question, maybe it makes sense to only support one version at a time. In which case we could remove 2.18 from glibc. Thomas
On 04/07/2014 10:29 AM, Thomas Petazzoni wrote: > So we don't do like glibc where we support both 2.18 and 2.19 ? This is > really a question, maybe it makes sense to only support one version at > a time. In which case we could remove 2.18 from glibc. My opinion is that we should just keep it as simple as possible. If there's a special need for some architecture (like microblaze) let it be so, but really unless we know there's some issue or exception i don't think there's much value in keeping multiple versions, specially since *glibc is backwards ABI compatible. Regards.
Dear Gustavo Zacarias, On Mon, 07 Apr 2014 10:36:43 -0300, Gustavo Zacarias wrote: > > So we don't do like glibc where we support both 2.18 and 2.19 ? This is > > really a question, maybe it makes sense to only support one version at > > a time. In which case we could remove 2.18 from glibc. > > My opinion is that we should just keep it as simple as possible. > If there's a special need for some architecture (like microblaze) let it > be so, but really unless we know there's some issue or exception i don't > think there's much value in keeping multiple versions, specially since > *glibc is backwards ABI compatible. Fine with me. My impression so far was that for critical components such as gcc/binutils/gdb/libc, we were keeping older versions a little bit longer to give it some time to test the newer version before forcing everyone to upgrade. Surely we need to continue to do this with gcc, where most of the 4.x.0 versions are typically broken on at least one architecture. Thomas
On 04/07/2014 10:41 AM, Thomas Petazzoni wrote: >> My opinion is that we should just keep it as simple as possible. >> If there's a special need for some architecture (like microblaze) let it >> be so, but really unless we know there's some issue or exception i don't >> think there's much value in keeping multiple versions, specially since >> *glibc is backwards ABI compatible. > > Fine with me. My impression so far was that for critical components > such as gcc/binutils/gdb/libc, we were keeping older versions a little > bit longer to give it some time to test the newer version before > forcing everyone to upgrade. Surely we need to continue to do this with > gcc, where most of the 4.x.0 versions are typically broken on at least > one architecture. D'oh, eglibc 2.19 is broken for PowerPC SPE ABI. So it seems we'll do two versions after all. I'll send a revised patch shortly. Regards.
>>>>> "Gustavo" == Gustavo Zacarias <gustavo@zacarias.com.ar> writes: Hi, >> Fine with me. My impression so far was that for critical components >> such as gcc/binutils/gdb/libc, we were keeping older versions a little >> bit longer to give it some time to test the newer version before >> forcing everyone to upgrade. Surely we need to continue to do this with >> gcc, where most of the 4.x.0 versions are typically broken on at least >> one architecture. > D'oh, eglibc 2.19 is broken for PowerPC SPE ABI. > So it seems we'll do two versions after all. > I'll send a revised patch shortly. Ok, thanks. It anyway seems like the safest option (and matches our other toolchain-related packages)
diff --git a/package/glibc/2.18-svnr23787/glibc-0001-accept-make4.patch b/package/glibc/2.18-svnr23787/glibc-0001-accept-make4.patch deleted file mode 100644 index 4f426f2..0000000 --- a/package/glibc/2.18-svnr23787/glibc-0001-accept-make4.patch +++ /dev/null @@ -1,33 +0,0 @@ -Backport upstream patch (28d708c44bc47b56f6551ff285f78edcf61c208a) to accept -make-4.0 or newer. -We patch both configure and configure.in files so if we ever have to run -autoreconf in the glibc source, then the fix will be propagated properly. - -Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> - -Index: glibc-2.18-svnr23787/libc/configure -=================================================================== ---- glibc-2.18-svnr23787.orig/libc/configure -+++ glibc-2.18-svnr23787/libc/configure -@@ -4772,7 +4772,7 @@ $as_echo_n "checking version of $MAKE... - ac_prog_version=`$MAKE --version 2>&1 | sed -n 's/^.*GNU Make[^0-9]*\([0-9][0-9.]*\).*$/\1/p'` - case $ac_prog_version in - '') ac_prog_version="v. ?.??, bad"; ac_verc_fail=yes;; -- 3.79* | 3.[89]*) -+ 3.79* | 3.[89]* | [4-9].* | [1-9][0-9]*) - ac_prog_version="$ac_prog_version, ok"; ac_verc_fail=no;; - *) ac_prog_version="$ac_prog_version, bad"; ac_verc_fail=yes;; - -Index: glibc-2.18-svnr23787/libc/configure.in -=================================================================== ---- glibc-2.18-svnr23787.orig/libc/configure.in -+++ glibc-2.18-svnr23787/libc/configure.in -@@ -989,7 +989,7 @@ AC_CHECK_PROG_VER(CC, ${ac_tool_prefix}g - critic_missing="$critic_missing gcc") - AC_CHECK_PROG_VER(MAKE, gnumake gmake make, --version, - [GNU Make[^0-9]*\([0-9][0-9.]*\)], -- [3.79* | 3.[89]*], critic_missing="$critic_missing make") -+ [3.79* | 3.[89]* | [4-9].* | [1-9][0-9]*], critic_missing="$critic_missing make") - - AC_CHECK_PROG_VER(MSGFMT, gnumsgfmt gmsgfmt msgfmt, --version, - [GNU gettext.* \([0-9]*\.[0-9.]*\)], diff --git a/package/glibc/glibc.mk b/package/glibc/glibc.mk index 6579f46..dbffdb7 100644 --- a/package/glibc/glibc.mk +++ b/package/glibc/glibc.mk @@ -16,7 +16,7 @@ GLIBC_SRC_SUBDIR = . endif else ifeq ($(BR2_TOOLCHAIN_BUILDROOT_EGLIBC),y) -GLIBC_VERSION = 2.18-svnr23787 +GLIBC_VERSION = 2.19-svnr25243 GLIBC_SITE = http://downloads.yoctoproject.org/releases/eglibc/ GLIBC_SOURCE = eglibc-$(GLIBC_VERSION).tar.bz2 GLIBC_SRC_SUBDIR = libc
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar> --- .../2.18-svnr23787/glibc-0001-accept-make4.patch | 33 ---------------------- package/glibc/glibc.mk | 2 +- 2 files changed, 1 insertion(+), 34 deletions(-) delete mode 100644 package/glibc/2.18-svnr23787/glibc-0001-accept-make4.patch