Bug#384204: 32-bit biarch support
On Wed, Aug 23, 2006 at 12:10:54AM +0300, Daniel Stone wrote: > On Tue, Aug 22, 2006 at 08:28:19PM +0200, Robert Millan wrote: > > Ok, if I understood correctly, changes in comparison with what we have now > > are: > > > > - A more standarised pathset, instead of just /emul/ia32-linux. According > > to > >http://wiki.debian.org/toolchain-multiarch, there are plans to make the > >toolchain aware of the new paths. Sounds a lot like post-etch, but AFAIK > >we can do the same without it (by using --libdir at build time, and > > adequate > >/etc/ld.so.conf at runtime) > > > > - dpkg / dak / katie major rework (I haven't read in-depth, but I assume > > the > >point is using the same ia32 code from the i386 port without compiling it > >separately. Definitely post-etch..) > > > > Did I miss something? Is someone able to guess which of them is Daniel > > concerned with? > > And the crucial points: > - generalised: works on any architecture, not just amd64/i386, > - non-intrusive: dak/dpkg changes are easy. changing every package to >have such horrendous hacks as this, with hardcoded >triplets, is never going to fly, > - cleaner: see previous two. Ok I give up. Would be nice if someone had told me when I filed #381342 19 days ago, though. -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Wed, Aug 23, 2006 at 07:39:17AM +0200, Robert Millan wrote: > On Wed, Aug 23, 2006 at 12:10:54AM +0300, Daniel Stone wrote: > > On Tue, Aug 22, 2006 at 08:28:19PM +0200, Robert Millan wrote: > > > Ok, if I understood correctly, changes in comparison with what we have now > > > are: > > > > > > - A more standarised pathset, instead of just /emul/ia32-linux. > > > According to > > >http://wiki.debian.org/toolchain-multiarch, there are plans to make the > > >toolchain aware of the new paths. Sounds a lot like post-etch, but > > > AFAIK > > >we can do the same without it (by using --libdir at build time, and > > > adequate > > >/etc/ld.so.conf at runtime) > > > > > > - dpkg / dak / katie major rework (I haven't read in-depth, but I assume > > > the > > >point is using the same ia32 code from the i386 port without compiling > > > it > > >separately. Definitely post-etch..) > > > > > > Did I miss something? Is someone able to guess which of them is Daniel > > > concerned with? > > > > And the crucial points: > > - generalised: works on any architecture, not just amd64/i386, > > - non-intrusive: dak/dpkg changes are easy. changing every package to > >have such horrendous hacks as this, with hardcoded > >triplets, is never going to fly, > > - cleaner: see previous two. > > Ok I give up. Would be nice if someone had told me when I filed #381342 19 > days > ago, though. Oh, as for the bugs, the patches don't really harm, so I won't close them. Feel free to apply, ignore or keep open untill the good solution is ready. -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 07:19:29PM +0300, Daniel Stone wrote: > On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > > > (needs lib32ice first) > > Um, why don't you focus your energies on proper multiarch support > instead of trying to propagate hacks like this further than they need to > be? Despite Daniel's rather gruff treatment of the topic, from what little I've heard on the subject I'm inclined to agree with him. Multiarch is widely accepted to be the proper fix, so I'm just going to close these reports in the next few days unless there's a good reason to leave them around. - David Nusinow -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 08:28:19PM +0200, Robert Millan wrote: > Ok, if I understood correctly, changes in comparison with what we have now > are: > > - A more standarised pathset, instead of just /emul/ia32-linux. According to >http://wiki.debian.org/toolchain-multiarch, there are plans to make the >toolchain aware of the new paths. Sounds a lot like post-etch, but AFAIK >we can do the same without it (by using --libdir at build time, and > adequate >/etc/ld.so.conf at runtime) > > - dpkg / dak / katie major rework (I haven't read in-depth, but I assume the >point is using the same ia32 code from the i386 port without compiling it >separately. Definitely post-etch..) > > Did I miss something? Is someone able to guess which of them is Daniel > concerned with? And the crucial points: - generalised: works on any architecture, not just amd64/i386, - non-intrusive: dak/dpkg changes are easy. changing every package to have such horrendous hacks as this, with hardcoded triplets, is never going to fly, - cleaner: see previous two. signature.asc Description: Digital signature
Bug#384204: 32-bit biarch support
Please take a look at http://wiki.debian.org/multiarch . -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) Finger [EMAIL PROTECTED] (NOT a valid e-mail address) for more info. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 02:06:36PM -0400, Aaron M. Ucko wrote: > Please take a look at http://wiki.debian.org/multiarch . Thanks. -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 08:12:14PM +0200, Robert Millan wrote: > On Tue, Aug 22, 2006 at 02:06:36PM -0400, Aaron M. Ucko wrote: > > Please take a look at http://wiki.debian.org/multiarch . > > Thanks. Ok, if I understood correctly, changes in comparison with what we have now are: - A more standarised pathset, instead of just /emul/ia32-linux. According to http://wiki.debian.org/toolchain-multiarch, there are plans to make the toolchain aware of the new paths. Sounds a lot like post-etch, but AFAIK we can do the same without it (by using --libdir at build time, and adequate /etc/ld.so.conf at runtime) - dpkg / dak / katie major rework (I haven't read in-depth, but I assume the point is using the same ia32 code from the i386 port without compiling it separately. Definitely post-etch..) Did I miss something? Is someone able to guess which of them is Daniel concerned with? -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 08:23:23PM +0300, Daniel Stone wrote: > On Tue, Aug 22, 2006 at 07:12:33PM +0200, Robert Millan wrote: > > On Tue, Aug 22, 2006 at 07:49:34PM +0300, Daniel Stone wrote: > > > On Tue, Aug 22, 2006 at 06:35:08PM +0200, Robert Millan wrote: > > > > On Tue, Aug 22, 2006 at 07:19:29PM +0300, Daniel Stone wrote: > > > > > On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > > > > > > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > > > > > > > > > > > (needs lib32ice first) > > > > > > > > > > Um, why don't you focus your energies on proper multiarch support > > > > > instead of trying to propagate hacks like this further than they need > > > > > to > > > > > be? > > > > > > > > Wait, this is not proper multiarch? > > > > > > No. > > > > Can you point me to any of: > > > > a) documentation/description of the "right thing". > > http://www.google.com/search?q=multiarch&btnI=I'm+Feeling+Lucky > > > b) current status / timeline. > > > > c) reason why everyone else (including glibc and gcc maintainers) seems > > to be > > following this scheme instead. > > These two questions are left as an exercise to the reader. Thank you for nothing. -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 07:12:33PM +0200, Robert Millan wrote: > On Tue, Aug 22, 2006 at 07:49:34PM +0300, Daniel Stone wrote: > > On Tue, Aug 22, 2006 at 06:35:08PM +0200, Robert Millan wrote: > > > On Tue, Aug 22, 2006 at 07:19:29PM +0300, Daniel Stone wrote: > > > > On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > > > > > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > > > > > > > > > (needs lib32ice first) > > > > > > > > Um, why don't you focus your energies on proper multiarch support > > > > instead of trying to propagate hacks like this further than they need to > > > > be? > > > > > > Wait, this is not proper multiarch? > > > > No. > > Can you point me to any of: > > a) documentation/description of the "right thing". http://www.google.com/search?q=multiarch&btnI=I'm+Feeling+Lucky > b) current status / timeline. > > c) reason why everyone else (including glibc and gcc maintainers) seems to > be > following this scheme instead. These two questions are left as an exercise to the reader. signature.asc Description: Digital signature
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 07:49:34PM +0300, Daniel Stone wrote: > On Tue, Aug 22, 2006 at 06:35:08PM +0200, Robert Millan wrote: > > On Tue, Aug 22, 2006 at 07:19:29PM +0300, Daniel Stone wrote: > > > On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > > > > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > > > > > > > (needs lib32ice first) > > > > > > Um, why don't you focus your energies on proper multiarch support > > > instead of trying to propagate hacks like this further than they need to > > > be? > > > > Wait, this is not proper multiarch? > > No. Can you point me to any of: a) documentation/description of the "right thing". b) current status / timeline. c) reason why everyone else (including glibc and gcc maintainers) seems to be following this scheme instead. -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 06:35:08PM +0200, Robert Millan wrote: > On Tue, Aug 22, 2006 at 07:19:29PM +0300, Daniel Stone wrote: > > On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > > > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > > > > > (needs lib32ice first) > > > > Um, why don't you focus your energies on proper multiarch support > > instead of trying to propagate hacks like this further than they need to > > be? > > Wait, this is not proper multiarch? No. signature.asc Description: Digital signature
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 07:19:29PM +0300, Daniel Stone wrote: > On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > > > (needs lib32ice first) > > Um, why don't you focus your energies on proper multiarch support > instead of trying to propagate hacks like this further than they need to > be? Wait, this is not proper multiarch? -- Robert Millan My spam trap is [EMAIL PROTECTED] Note: this address is only intended for spam harvesters. Writing to it will get you added to my black list. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384204: 32-bit biarch support
On Tue, Aug 22, 2006 at 05:22:48PM +0200, Robert Millan wrote: > Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev). > > (needs lib32ice first) Um, why don't you focus your energies on proper multiarch support instead of trying to propagate hacks like this further than they need to be? signature.asc Description: Digital signature
Bug#384204: 32-bit biarch support
Package: libsm6
Version: 1:1.0.1-1
Severity: wishlist
Tags: patch
Patch attached for 32-bit biarch support (lib32sm6 and lib32sm-dev).
(needs lib32ice first)
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy: (500, 'testing'), (1, 'experimental')
Architecture: amd64 (x86_64)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.17-1-amd64-k8
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) (ignored: LC_ALL
set to en_US.UTF-8)
Versions of packages libsm6 depends on:
ii libc6 2.3.6-15 GNU C Library: Shared libraries
ii libice6 1:1.0.1-1 X11 Inter-Client Exchange library
ii x11-common1:7.0.22 X Window System (X.Org) infrastruc
libsm6 recommends no packages.
-- no debconf information
diff -Nur libsm-1.0.1.old/debian/control libsm-1.0.1/debian/control
--- libsm-1.0.1.old/debian/control 2006-08-21 20:06:06.0 +0200
+++ libsm-1.0.1/debian/control 2006-08-22 17:08:57.0 +0200
@@ -3,7 +3,7 @@
Priority: optional
Maintainer: Debian X Strike Force
Uploaders: David Nusinow <[EMAIL PROTECTED]>, Branden Robinson <[EMAIL
PROTECTED]>, Fabio M. Di Nitto <[EMAIL PROTECTED]>
-Build-Depends: debhelper (>= 5.0.0), pkg-config, x11proto-core-dev (>=
6.8.99.15+cvs.20050722-1), xtrans-dev, libice-dev (>= 1:1.0.0-1)
+Build-Depends: debhelper (>= 5.0.0), libc6-dev-i386 [amd64], pkg-config,
x11proto-core-dev (>= 6.8.99.15+cvs.20050722-1), xtrans-dev, libice-dev (>=
1:1.0.0-1), lib32ice-dev (>= 1:1.0.0-1) [amd64]
Standards-Version: 3.7.2.0
Package: libsm6
@@ -57,3 +57,41 @@
.
This module can be found as the module 'lib/SM' at
:pserver:[EMAIL PROTECTED]:/cvs/xorg
+
+Package: lib32sm6
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}, x11-common
+Description: X11 Session Management library
+ This package provides the main interface to the X11 Session Management
+ library, which allows for applications to both manage sessions, and make use
+ of session managers to save and restore their state for later use.
+ .
+ More information about X.Org can be found at:
+ http://xorg.freedesktop.org>
+ http://lists.freedesktop.org/mailman/listinfo/xorg>
+ .
+ This module can be found as the module 'lib/SM' at
+ :pserver:[EMAIL PROTECTED]:/cvs/xorg
+ .
+ This package supports the 32-bit ABI variant of your system's architecture.
+
+Package: lib32sm-dev
+Architecture: amd64
+Depends: ${shlibs:Depends}, ${misc:Depends}, lib32sm6 (= ${Source-Version}),
x11proto-core-dev, lib32ice-dev, lib32c-dev, libsm-dev (= ${Source-Version})
+Pre-Depends: x11-common (>= 1:7.0.0)
+Description: X11 Inter-Client Exchange library (development headers)
+ This package provides the main interface to the X11 Session Management
+ library, which allows for applications to both manage sessions, and make use
+ of session managers to save and restore their state for later use.
+ .
+ This package contains the development headers for the library found in libsm6.
+ Non-developers likely have little use for this package.
+ .
+ More information about X.Org can be found at:
+ http://xorg.freedesktop.org>
+ http://lists.freedesktop.org/mailman/listinfo/xorg>
+ .
+ This module can be found as the module 'lib/SM' at
+ :pserver:[EMAIL PROTECTED]:/cvs/xorg
+ .
+ This package supports the 32-bit ABI variant of your system's architecture.
diff -Nur libsm-1.0.1.old/debian/lib32sm6.install
libsm-1.0.1/debian/lib32sm6.install
--- libsm-1.0.1.old/debian/lib32sm6.install 1970-01-01 01:00:00.0
+0100
+++ libsm-1.0.1/debian/lib32sm6.install 2006-08-22 17:09:16.0 +0200
@@ -0,0 +1 @@
+emul/ia32-linux/usr/lib/libSM.so.6*
diff -Nur libsm-1.0.1.old/debian/lib32sm-dev.install
libsm-1.0.1/debian/lib32sm-dev.install
--- libsm-1.0.1.old/debian/lib32sm-dev.install 1970-01-01 01:00:00.0
+0100
+++ libsm-1.0.1/debian/lib32sm-dev.install 2006-08-22 17:09:54.0
+0200
@@ -0,0 +1,3 @@
+emul/ia32-linux/usr/lib/libSM.a
+emul/ia32-linux/usr/lib/libSM.so
+emul/ia32-linux/usr/lib/pkgconfig/sm.pc
Binary files
libsm-1.0.1.old/debian/libxau6-dbg/usr/lib/debug/emul/ia32-linux/usr/lib/libSM.so.6.0.0
and
libsm-1.0.1/debian/libxau6-dbg/usr/lib/debug/emul/ia32-linux/usr/lib/libSM.so.6.0.0
differ
Binary files
libsm-1.0.1.old/debian/libxau6-dbg/usr/lib/debug/usr/lib/libSM.so.6.0.0 and
libsm-1.0.1/debian/libxau6-dbg/usr/lib/debug/usr/lib/libSM.so.6.0.0 differ
diff -Nur libsm-1.0.1.old/debian/rules libsm-1.0.1/debian/rules
--- libsm-1.0.1.old/debian/rules2006-08-21 20:06:06.0 +0200
+++ libsm-1.0.1/debian/rules2006-08-22 17:12:33.0 +0200
@@ -25,12 +25,12 @@
DEB_HOST_ARCH ?= $(shell dpkg-architecture -qDEB_HOST_ARCH)
DEB_HOST_GNU_TYPE ?= $(shell dpkg-architecture -qDEB_HOST_GNU_TYPE)
DEB_BUILD_GNU_TYPE ?= $(shell dpkg-architecture -qDEB_BUILD_GNU_TYPE)
-ifeq ($(DEB_BUILD_GNU_TYPE), $(DEB_HOST_GNU_TYPE))
- confflags += --build=$(DEB_HOST_GNU_TYPE)
-else
-
