Bug#684005: unblock: busybox/1:1.20.0-6

2012-08-06 Thread Michael Tokarev
Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package busybox.

The version in unstable fixes a single bug - busybox basically was
unusable on s390(x) due to a programming error.  The single fix
merely removes __attribute__(aligned(1)) and similar qualifiers
for a few strings on s390(x) where it was problematic.

The other change in the package - reordering of patches in debian/series -
makes no effect whatsoever on the resulting source (after applying all
patches - I diffed the result to be sure), but makes it easy to group
upstream and non-upstream patches.

So the only change in this release is to remove forced-alignment of
static strings on s390(x).

Thanks!

/mjt

unblock busybox/1:1.20.0-6

diff -Nru busybox-1.20.0/debian/changelog busybox-1.20.0/debian/changelog
--- busybox-1.20.0/debian/changelog 2012-07-08 01:37:33.0 +0400
+++ busybox-1.20.0/debian/changelog 2012-07-22 12:30:25.0 +0400
@@ -1,3 +1,15 @@
+busybox (1:1.20.0-6) unstable; urgency=low
+
+  * reorder patches in debian/patches/series: all upstream first,
+debian-specific next.  cmp(1) shows no changes in the resulting
+sources (after applying patches both ways)
+  * dont-force-no-alignment-for-s390.patch: do not use ALIGN* macros
+on s390 and s390x because gcc generates wrong code (for wrong
+declarations).  No effect for anything but s390(x), where the
+resulting package does not work anyway. (Closes: 681760)
+
+ -- Michael Tokarev m...@tls.msk.ru  Sun, 22 Jul 2012 12:30:02 +0400
+
 busybox (1:1.20.0-5) unstable; urgency=low
 
   * enable various xargs features for all flavours, to make it work
diff -Nru busybox-1.20.0/debian/patches/dont-force-no-alignment-for-s390.patch 
busybox-1.20.0/debian/patches/dont-force-no-alignment-for-s390.patch
--- busybox-1.20.0/debian/patches/dont-force-no-alignment-for-s390.patch
1970-01-01 03:00:00.0 +0300
+++ busybox-1.20.0/debian/patches/dont-force-no-alignment-for-s390.patch
2012-07-22 11:58:39.0 +0400
@@ -0,0 +1,29 @@
+Subject: dont force no alignment for s390
+From: Michael Tokarev m...@tls.msk.ru
+Bugs-Debian: http://bugs.debian.org/681760
+Forwarded: no
+
+This fix hasn't been forwarded upstream, because might be more
+correct fix is to use the ALIGN* marks properly in the declarations
+too, or just remove whol ALIGN* mess:
+
+ http://lists.busybox.net/pipermail/busybox/2012-July/078155.html
+ http://lists.busybox.net/pipermail/busybox/2012-July/078163.html
+
+--- a/include/platform.h
 b/include/platform.h
+@@ -284,7 +284,13 @@
+ #define fdprintf dprintf
+ 
+ /* Useful for defeating gcc's alignment of char message[]-like data */
+-#if 1 /* if needed: !defined(arch1)  !defined(arch2) */
++/* #if 1 /o if needed: !defined(arch1)  !defined(arch2) */
++/* on s390 and s390x, GCC may generate incorrect code when accessing
++ * variables defined with ALIGN* but declared without -- assuming alignment
++ * on access which isn't actually present.  The right fix is to use ALIGN*
++ * properly in all declarations too, but there are quite some of such places.
++ */
++#if !defined(__s390__)
+ # define ALIGN1 __attribute__((aligned(1)))
+ # define ALIGN2 __attribute__((aligned(2)))
+ # define ALIGN4 __attribute__((aligned(4)))
diff -Nru busybox-1.20.0/debian/patches/series 
busybox-1.20.0/debian/patches/series
--- busybox-1.20.0/debian/patches/series2012-07-06 19:21:19.0 
+0400
+++ busybox-1.20.0/debian/patches/series2012-07-22 11:59:35.0 
+0400
@@ -1,5 +1,14 @@
 1.20.1.patch
 
+# upstream stable patches
+busybox-1.20.1-ash.patch
+busybox-1.20.1-ifupdown.patch
+busybox-1.20.1-man.patch
+busybox-1.20.1-tar.patch
+busybox-1.20.1-ps.patch
+busybox-1.20.1-mke2fs.patch
+busybox-1.20.1-1.20.2.patch
+
 shell-ash-export-HOME.patch
 # we need to get rid of this one:
 #applets-fallback.patch
@@ -10,11 +19,5 @@
 u-mount-FreeBSD-support.patch
 swaponoff-FreeBSD-support.patch
 
-# upstream stable patches
-busybox-1.20.1-ash.patch
-busybox-1.20.1-ifupdown.patch
-busybox-1.20.1-man.patch
-busybox-1.20.1-tar.patch
-busybox-1.20.1-ps.patch
-busybox-1.20.1-mke2fs.patch
-busybox-1.20.1-1.20.2.patch
+# http://bugs.debian.org/681760
+dont-force-no-alignment-for-s390.patch


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#684005: unblock: busybox/1:1.20.0-6

2012-08-06 Thread Philipp Kern
On Mon, Aug 06, 2012 at 12:33:36PM +0400, Michael Tokarev wrote:
 Please unblock package busybox.
 
 The version in unstable fixes a single bug - busybox basically was
 unusable on s390(x) due to a programming error.  The single fix
 merely removes __attribute__(aligned(1)) and similar qualifiers
 for a few strings on s390(x) where it was problematic.
 
 The other change in the package - reordering of patches in debian/series -
 makes no effect whatsoever on the resulting source (after applying all
 patches - I diffed the result to be sure), but makes it easy to group
 upstream and non-upstream patches.
 
 So the only change in this release is to remove forced-alignment of
 static strings on s390(x).

This is fine from a RT point of view, thanks. But given that this package needs
the d-i RM ACK and I couldn't find a definite answer in the archives, this
still needs confirmation. KiBi?

Kind regards
Philipp Kern


signature.asc
Description: Digital signature


Bug#684005: unblock: busybox/1:1.20.0-6

2012-08-06 Thread Cyril Brulebois
Michael Tokarev m...@tls.msk.ru (06/08/2012):
 Package: release.debian.org
 Severity: normal
 User: release.debian@packages.debian.org
 Usertags: unblock
 
 Please unblock package busybox.

ACK for d-i. Thanks!

Mraw,
KiBi.


signature.asc
Description: Digital signature


Bug#684005: unblock: busybox/1:1.20.0-6

2012-08-06 Thread Cyril Brulebois
Hello,

just replying for the sake of people reading -boot@ and wondering:

Philipp Kern pk...@debian.org (06/08/2012):
 This is fine from a RT point of view, thanks. But given that this package
 needs the d-i RM ACK and I couldn't find a definite answer in the archives,
 this still needs confirmation. KiBi?

I've just sent an ACK to that very bug some minutes ago. ;-)

Mraw,
KiBi.


signature.asc
Description: Digital signature