Bug#469035: libc6-i386: include mutiarch ld.conf from i386
tags 469035 patch thanks This generates the requested file on the amd64 build. --- glibc-2.7.orig/debian/sysdeps/amd64.mk 2008-04-05 02:05:33.0 +0200 +++ glibc-2.7/debian/sysdeps/amd64.mk 2008-04-05 02:35:38.0 +0200 @@ -40,5 +40,9 @@ ln -sf /emul/ia32-linux/lib/ld-linux.so.2 debian/libc6-i386/lib ln -sf /emul/ia32-linux/lib debian/libc6-i386/lib32 ln -sf /emul/ia32-linux/usr/lib debian/libc6-i386/usr/lib32 +mkdir -p debian/libc6-i386/etc/ld.so.conf.d +echo "# Multiarch support" > debian/libc6-i386/etc/ld.so.conf.d/i486-linux-gnu.conf +echo /lib/i486-linux-gnu >> debian/libc6-i386/etc/ld.so.conf.d/i486-linux-gnu.conf +echo /usr/lib/i486-linux-gnu >> debian/libc6-i386/etc/ld.so.conf.d/i486-linux-gnu.conf endef
Processed: Re: libc6-i386: include mutiarch ld.conf from i386
Processing commands for [EMAIL PROTECTED]: > tags 469035 patch Bug#469035: libc6-i386: include mutiarch ld.conf from i386 There were no tags set. Tags added: patch > thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
tzdata 2008b-1 MIGRATED to testing
FYI: The status of the tzdata source package in Debian's testing distribution has changed. Previous version: 2008a-1 Current version: 2008b-1 -- This email is automatically generated; [EMAIL PROTECTED] is responsible. See http://people.debian.org/~henning/trille/ for more information. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#474293: glibc: Add udeb: lines in shlibs files to improve udeb dependencies
Package: glibc
Version: 2.7-10
Severity: wishlist
Tags: d-i patch
For Etch we fixed debhelper and most library packages that provide udebs to
improve the dependencies generated for udebs, but we skipped glibc then as
it's less important when the installer is running (as libc is always
included in the D-I initrds anyway and thus pulled in at build time).
However, it would be great to have this fixed before Lenny as it will help
with the implementation of britney support for udebs.
The attached patch will add udeb: lines in the various libc packages, for
example for libc6:
udeb: ld-linux 2 libc6-udeb (>= 2.7-1)
udeb: libm 6 libc6-udeb (>= 2.7-1)
udeb: libdl 2 libc6-udeb (>= 2.7-1)
udeb: libresolv 2 libc6-udeb (>= 2.7-1)
udeb: libc 6 libc6-udeb (>= 2.7-1)
udeb: libutil 1 libc6-udeb (>= 2.7-1)
udeb: libcrypt 1 libc6-udeb (>= 2.7-1)
udeb: librt 1 libc6-udeb (>= 2.7-1)
udeb: libpthread 0 libc6-udeb (>= 2.7-1)
udeb: libnss_dns 2 libnss-dns-udeb (>= 2.7-1)
udeb: libnss_files 2 libnss-files-udeb (>= 2.7-1)
Applying the patch should be safe and there are no transition issues.
Possibly the change should be checked with Release Masters, but IMO it's
not a problem to implement this at this stage of the release of Lenny.
After glibc has been uploaded with this patch, I plan to request binNMUs for
all D-I packages that depend on libc to get their dependencies fixed.
Example of the effect of the patch
--
fdisk-udeb_2.13.1-3_i386.udeb currently has:
Depends: libc6 (>= 2.7-1)
when built against glibc with this patch this becomes:
Depends: libc6-udeb (>= 2.7-1)
Comments
We are aware the patch is a bit of a hack and the list of libs in the helper
script will require some maintenance. We have discussed whether this could
be implemented in debhelper instead, but this solution was preferred:
http://lists.debian.org/debian-boot/2008/02/msg00336.html
However, if you see alternative solutions, I'd be more than willing to
discuss them and help develop/test them.
I removed the commented out dh_makeshlibs line in udebs section of
debhelper.mk as I felt that to be better than adding a commented out call
to the shlibs-add-udebs script.
I have only tested the patch on amd64 and i386, but AFAICT it should work
for other arches too.
The udeb: lines are only added for libc6 and equivalent packages, and *not*
for the "crossarch" packages (libc6-amd64, libc-i386, libc6-xen, etc.).
AFAICT udebs should not be compiled against those variants, so adding them
there seemed redundant.
However, I'm not completely sure that is correct and I'd appreciate your
opinion on this.
Please consider including this patch with your next upload.
Cheers,
FJP
commit e5619fdd8444d5e48d9cdea223f94df43dfa350f
Author: Frans Pop <[EMAIL PROTECTED]>
Date: Fri Apr 4 10:23:29 2008 +0200
Add udeb lines in shlibs files
Currently onther udebs depend on regular libc packages. This will allow
then to correctly depend on the libc udeb instead.
diff --git a/debian/rules.d/debhelper.mk b/debian/rules.d/debhelper.mk
index fe127dd..96cfb98 100644
--- a/debian/rules.d/debhelper.mk
+++ b/debian/rules.d/debhelper.mk
@@ -109,6 +109,9 @@ endif
-o -regex '.*/libc-.*so' \) \
-exec chmod a+x '{}' ';'
dh_makeshlibs -X/usr/lib/debug -p$(curpass) -V "$(call xx,shlib_dep)"
+ # Add relevant udeb: lines in shlibs files
+ chmod a+x debian/shlibs-add-udebs
+ ./debian/shlibs-add-udebs $(curpass)
if [ -f debian/$(curpass).lintian ] ; then \
install -d -m 755 -o root -g root debian/$(curpass)/usr/share/lintian/overrides/ ; \
@@ -152,7 +155,6 @@ $(patsubst %,$(stamp)binaryinst_%,$(DEB_UDEB_PACKAGES)): $(stamp)debhelper
-o -regex '.*lib[0-9]*/.*libpthread.*so.*' \
-o -regex '.*lib[0-9]*/libc[.-].*so.*' \) \
-exec chmod a+x '{}' ';'
- # dh_makeshlibs -X/usr/lib/debug -p$(curpass) -V "$(call xx,shlib_dep)"
dh_installdeb -p$(curpass)
# dh_shlibdeps -p$(curpass)
dh_gencontrol -p$(curpass)
diff --git a/debian/shlibs-add-udebs b/debian/shlibs-add-udebs
new file mode 100755
index 000..651fb8e
--- /dev/null
+++ b/debian/shlibs-add-udebs
@@ -0,0 +1,51 @@
+#! /bin/sh
+set -e
+
+# This script adds "udeb lines" to shlibs files which allows other udebs
+# to get correct dependencies when built against glibc libraries.
+# The script was written by Frans Pop <[EMAIL PROTECTED]>.
+
+package="$1"
+shlibs_file="debian/$package/DEBIAN/shlibs"
+
+# Skip packages that don't have an shlibs file.
+# The "cross-subarch" library packages have an shlibs file, but should
+# not have udeb lines, so skip those as well.
+if [ ! -r "$shlibs_file" ] || \
+ echo "$package" | grep -Eq "^libc[0-9.]+-"; then
+ exit 0
+fi
+
+# $1: regexp to select libraries for which lines should be duplicated
+# $2: name of the udeb the new line should point to
+add_udeb_line() {
+ local regexp udeb line lib soname package rest
+ regexp="$1"
+ udeb="$2"
+
+ if line="$(grep "^$regexp[[:space:]]" $shlibs_file)"; then
+ echo
Bug#474226: glibc: Segfaults in tst-rfc3484 during compilation
On Friday 04 April 2008, Aurelien Jarno wrote: > > During compilation of glibc I noticed two segfaults in my logs (dmesg): > > ld-linux.so.2[4016]: segfault at 4 ip 80499bc sp ffb9c6fc error 4 in > > tst-rfc3484[8048000+6000] ld-linux.so.2[4030]: segfault at 4 ip 8049a8c > > sp fff2b1e4 error 4 in tst-rfc3484-2[8048000+6000] > > > > Is this an issue in glibc or maybe in the kernel? > > Note that I'm running an upstream 2.6.25-rc8 kernel. > > I guess it's a bug in the test, but I am unable to reproduce the problem. I should probably also have mentioned that this is in an i386 chroot on an amd64 box. Same segfault happened during a second compile, so it looks to be reproducible for me. How can I run the test separately? What additional info do you need and how can I get that? Cheers, FJP -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#474226: glibc: Segfaults in tst-rfc3484 during compilation
Frans Pop a écrit : > Package: glibc > Version: 2.7-10 > Severity: normal > > During compilation of glibc I noticed two segfaults in my logs (dmesg): > ld-linux.so.2[4016]: segfault at 4 ip 80499bc sp ffb9c6fc error 4 in > tst-rfc3484[8048000+6000] > ld-linux.so.2[4030]: segfault at 4 ip 8049a8c sp fff2b1e4 error 4 in > tst-rfc3484-2[8048000+6000] > > Is this an issue in glibc or maybe in the kernel? > Note that I'm running an upstream 2.6.25-rc8 kernel. I guess it's a bug in the test, but I am unable to reproduce the problem. -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug in ldd?
Nephila a écrit : > Before opening a bug i would like to have your opinion about this behaviour i > discovered in ldd: > > On a amd64 etch (fully updated) the command "ldd /usr/bin/groups" (or any > other bash script) results in error: > /usr/bin/ldd: line 171: /lib/ld-linux.so.2: No such file or directory > ldd: /lib/ld-linux.so.2 exited with unknown exit code (127) > > /lib/ld-linux.so.2 in fact does not exists (is included in libc6-i386) > > Which is the right solution? Installing libc6-i386 ? Fixing ldd ? > Should be any stronger dependency on libc6-i386? > That's most probably a bug in ldd -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug in ldd?
Before opening a bug i would like to have your opinion about this behaviour i discovered in ldd: On a amd64 etch (fully updated) the command "ldd /usr/bin/groups" (or any other bash script) results in error: /usr/bin/ldd: line 171: /lib/ld-linux.so.2: No such file or directory ldd: /lib/ld-linux.so.2 exited with unknown exit code (127) /lib/ld-linux.so.2 in fact does not exists (is included in libc6-i386) Which is the right solution? Installing libc6-i386 ? Fixing ldd ? Should be any stronger dependency on libc6-i386? -- Regards Nephila sas PGP key block: http://www.nephila.it/pgp -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#474226: glibc: Segfaults in tst-rfc3484 during compilation
Package: glibc Version: 2.7-10 Severity: normal During compilation of glibc I noticed two segfaults in my logs (dmesg): ld-linux.so.2[4016]: segfault at 4 ip 80499bc sp ffb9c6fc error 4 in tst-rfc3484[8048000+6000] ld-linux.so.2[4030]: segfault at 4 ip 8049a8c sp fff2b1e4 error 4 in tst-rfc3484-2[8048000+6000] Is this an issue in glibc or maybe in the kernel? Note that I'm running an upstream 2.6.25-rc8 kernel. Cheers, FJP -- System Information: Debian Release: lenny/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.25-rc8 (SMP w/2 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#444866: Bug #444866: tzdata: Please provide tzdata-source package with original Olsen database
Torsten Werner a écrit : > Hi, > > I second the request. A source package with the original timezone data > would be very useful for the jdk packages and libjoda-time-java, too. > Note that a tzdata-java package will be provided soon (when the openjdk package is uploaded to unstable). I am clearly not in favor of providing tzdata-source (and I guess it is the same for the release team), as it means each time a new tzdata is uploaded (and that happens often), other packages depending on tzdata-source have to be rebuilt. -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' [EMAIL PROTECTED] | [EMAIL PROTECTED] `-people.debian.org/~aurel32 | www.aurel32.net -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#444866: Bug #444866: tzdata: Please provide tzdata-source package with original Olsen database
Hi, I second the request. A source package with the original timezone data would be very useful for the jdk packages and libjoda-time-java, too. Thank you, Torsten -- http://twerner.blogspot.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
