Bug#899318: virtualbox: FTBFS: vboxssdt-standard.hex:16:23: error: expected initializer before '-' token

[Mattia Dongili]

On Fri, May 25, 2018 at 07:07:53AM +0900, Mattia Dongili wrote:
[...]
> Gianfranco,
> if you need a work-around, changing the file prefix to something that is
> a legal C identifier might help.

Gianfranco,
so are you going to patch virtualbox?
In terms of severity, I think this bug against acpica-unix should be
"Important" rather than "Serious".

Thanks
-- 
mattia
:wq!

Bug#899318: virtualbox: FTBFS: vboxssdt-standard.hex:16:23: error: expected initializer before '-' token

[Mattia Dongili]

On Thu, May 24, 2018 at 06:33:45AM +0900, Mattia Dongili wrote:
> On Wed, May 23, 2018 at 02:56:55PM +0200, Gianfranco Costamagna wrote:
> > control: reassign -1 src:acpica-unix
> > control: found -1 20180508-1
> > control: affects -1 virtualbox
> > 
> > On Tue, 22 May 2018 19:12:32 +0200 Emilio Pozuelo Monfort 
> > <po...@debian.org> wrote:
> > > Package: virtualbox
> > > Version: 5.2.12-dfsg-1
> > > Severity: serious
> > > 
> > > virtualbox fails to build on a clean sid chroot here:
> > > 
> > > In file included from 
> > > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:45:0:
> > > /build/virtualbox-5.2.12-dfsg/out/obj/VBoxDD/vboxssdt-standard.hex:16:23: 
> > > error: expected initializer before '-' token
> > >  unsigned char vboxssdt-standard_aml_code[] =
> > >^
> > > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp: In 
> > > function 'int acpiPrepareDsdt(PPDMDEVINS, void**, size_t*)':
> > > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:371:32:
> > >  error: 'AmlCode' was not declared in this scope
> > >  cbAmlCodeDsdt = sizeof(AmlCode);
> > > ^~~
> > > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp: In 
> > > function 'int acpiPrepareSsdt(PPDMDEVINS, void**, size_t*)':
> > > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:423:30:
> > >  error: 'AmlCodeSsdtCpuHotPlug' was not declared in this scope
> > >  pabAmlCode = AmlCodeSsdtCpuHotPlug;
> > >   ^
> > > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:428:30:
> > >  error: 'AmlCodeSsdtStandard' was not declared in this scope
> > >  pabAmlCode = AmlCodeSsdtStandard;
> > >   ^~~
> > > kmk: *** [/usr/share/kBuild/footer-pass2-compiling-targets.kmk:226: 
> > > /build/virtualbox-5.2.12-dfsg/out/obj/VBoxDD/PC/ACPI/VBoxAcpi.o] Error 1
> 
> For the record, the DSL files are here:
> https://salsa.debian.org/pkg-virtualbox-team/virtualbox/tree/master/src/VBox/Devices/PC
> 
> and are compiled and post-processed as:
> https://salsa.debian.org/pkg-virtualbox-team/virtualbox/blob/master/src/VBox/Devices/Makefile.kmk#L829-857

So, apparently upstream is aware of this. See comments on
https://github.com/acpica/acpica/commit/f9a88a4c1cd020b6a5475d63b29626852a0b5f37#diff-95e84a7c99c268ce5433dea47cf17ec0

Gianfranco,
if you need a work-around, changing the file prefix to something that is
a legal C identifier might help.
Either way, unless the offending change is not backed out (as Bob suggested
in the discussion I linked), the "AmlCode" identifier is gone for good,
so things like
$(QUIET)$(SED) -e "s/AmlCode/AmlCodeSsdtStandard/g" \

$(QUIET)$(SED) -e "s/AmlCode/AmlCodeSsdtCpuHotPlug/g" \
won't work.
-- 
mattia
:wq!

Bug#899318: virtualbox: FTBFS: vboxssdt-standard.hex:16:23: error: expected initializer before '-' token

[Mattia Dongili]

On Wed, May 23, 2018 at 02:56:55PM +0200, Gianfranco Costamagna wrote:
> control: reassign -1 src:acpica-unix
> control: found -1 20180508-1
> control: affects -1 virtualbox
> 
> On Tue, 22 May 2018 19:12:32 +0200 Emilio Pozuelo Monfort  
> wrote:
> > Package: virtualbox
> > Version: 5.2.12-dfsg-1
> > Severity: serious
> > 
> > virtualbox fails to build on a clean sid chroot here:
> > 
> > In file included from 
> > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:45:0:
> > /build/virtualbox-5.2.12-dfsg/out/obj/VBoxDD/vboxssdt-standard.hex:16:23: 
> > error: expected initializer before '-' token
> >  unsigned char vboxssdt-standard_aml_code[] =
> >^
> > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp: In 
> > function 'int acpiPrepareDsdt(PPDMDEVINS, void**, size_t*)':
> > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:371:32: 
> > error: 'AmlCode' was not declared in this scope
> >  cbAmlCodeDsdt = sizeof(AmlCode);
> > ^~~
> > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp: In 
> > function 'int acpiPrepareSsdt(PPDMDEVINS, void**, size_t*)':
> > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:423:30: 
> > error: 'AmlCodeSsdtCpuHotPlug' was not declared in this scope
> >  pabAmlCode = AmlCodeSsdtCpuHotPlug;
> >   ^
> > /build/virtualbox-5.2.12-dfsg/src/VBox/Devices/PC/ACPI/VBoxAcpi.cpp:428:30: 
> > error: 'AmlCodeSsdtStandard' was not declared in this scope
> >  pabAmlCode = AmlCodeSsdtStandard;
> >   ^~~
> > kmk: *** [/usr/share/kBuild/footer-pass2-compiling-targets.kmk:226: 
> > /build/virtualbox-5.2.12-dfsg/out/obj/VBoxDD/PC/ACPI/VBoxAcpi.o] Error 1

For the record, the DSL files are here:
https://salsa.debian.org/pkg-virtualbox-team/virtualbox/tree/master/src/VBox/Devices/PC

and are compiled and post-processed as:
https://salsa.debian.org/pkg-virtualbox-team/virtualbox/blob/master/src/VBox/Devices/Makefile.kmk#L829-857

-- 
mattia
:wq!

Bug#857126: cpufrequtils: Fails to install/upgrade (syntax error in init script)

[Mattia Dongili]

On Sat, Mar 11, 2017 at 05:06:39PM +0100, Sebastian Humenda wrote:
> Hi

Hi,
it looks like we almost overlapped in our replies.

> Andreas Henriksson schrieb am 08.03.2017, 21:55 +0100:
> >On Wed, Mar 08, 2017 at 10:57:38AM +0100, Sebastian Humenda wrote:
...
> >> Mar 08 10:54:47 kraftkrust cpufrequtils[13461]: /etc/init.d/cpufrequtils: 
> >> 3: /etc/default/cpufrequtils: Syntax error: Unterminated quoted string
...
> >Could you please attach your /etc/init.d/cpufrequtils ? Do you know
> >if this file was ever modified by you or not?
> It is attached. It has not been modified by me, especially I have just purged 
> it
> and the issue persists.

Just to reiterate, it's /etc/default/cpufrequtils that has an
unterminated quoted string, not /etc/init.d/cpufrequtils.

Could you check that?

Thanks
-- 
mattia
:wq!


signature.asc
Description: PGP signature

Bug#857126: cpufrequtils: Fails to install/upgrade (syntax error in init script)

[Mattia Dongili]

On Wed, Mar 08, 2017 at 09:55:02PM +0100, Andreas Henriksson wrote:
> Control: tags -1 + moreinfo
> 
> Hello Sebastian Humenda,
> 
> Thanks for your bug report
> 
> On Wed, Mar 08, 2017 at 10:57:38AM +0100, Sebastian Humenda wrote:
> > Package: cpufrequtils
> > Version: 008-1+b1
> > Severity: grave
> > Justification: renders package unusable
> [...]
> > Mar 08 10:54:47 kraftkrust cpufrequtils[13461]: /etc/init.d/cpufrequtils: 
> > 3: /etc/default/cpufrequtils: Syntax error: Unterminated quoted string
> [...]
> 
> The /etc/init.d/cpufrequtils file is not shipped in the package since
> cpufrequtils 002-2.1 (a very long time ago, prior to old-old-stable).

as the error suggests, the unterminated string is in
/etc/default/cpufrequtils rather than in the init script.

/etc/init.d/cpufrequtils sources the defaults file, so that too has to be
correctly parseable by /bin/sh.

if [ -f /etc/default/cpufrequtils ] ; then
. /etc/default/cpufrequtils
fi

/etc/default/cpufrequtils is not a file managed by the package.

Sebastian,
Let me know if you need help writing the defaults file.

I'm closing this bug report.
-- 
mattia
:wq!

Bug#840852: please make libcpupower-dev depend on libcpupower1 (= ${binary:Version})

[Mattia Dongili]

tags 840852 + patch
thanks

Ben,

On Sat, Oct 15, 2016 at 06:13:55PM +0300, Vlad Orlov wrote:
...
> Please make libcpupower-dev depend on libcpupower1 (= ${binary:Version}),
> like it's usually done with all the other -dev packages.

I realize the patch is trivial, but anyway here's one that is build
tested and working.

Regards
-- 
mattia
:wq!
commit 8cee65c3229f5f692ad97df9bca117ef194c9ac7
Author: Mattia Dongili <malat...@linux.it>
Date:   Sun Oct 16 18:08:41 2016 -0700

libcpufreq-dev depends on its corresponding binary package

Closes: #840852
Signed-off-by: Mattia Dongili <malat...@linux.it>

diff --git a/debian/changelog b/debian/changelog
index 995c950..eef97ea 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -32,6 +32,10 @@ linux (4.8.1-1~exp1) UNRELEASED; urgency=medium
   * [arm64] Enable SERIAL_8250_EXTENDED, SERIAL_8250_SHARE_IRQ and
 SERIAL_8250_BCM2835AUX, needed for Raspberry Pi 3.
 
+  [ Mattia Dongili ]
+  * make libcpufreq-dev depend on its corresponding binary package
+(Closes: #840852)
+
  -- Ben Hutchings <b...@decadent.org.uk>  Sat, 01 Oct 2016 21:51:33 +0100
 
 linux (4.8~rc8-1~exp1) experimental; urgency=medium
diff --git a/debian/templates/control.tools.in b/debian/templates/control.tools.in
index 7c95172..e9318e3 100644
--- a/debian/templates/control.tools.in
+++ b/debian/templates/control.tools.in
@@ -31,7 +31,7 @@ Package: libcpupower-dev
 Build-Profiles: 
 Section: libdevel
 Architecture: linux-any
-Depends: ${shlibs:Depends}, ${misc:Depends}
+Depends: libcpupower1 (= ${binary:Version}), ${shlibs:Depends}, ${misc:Depends}
 Provides: libcpufreq-dev
 Conflicts: libcpufreq-dev
 Replaces: libcpufreq-dev

Bug#817708: uml-utilities marked for removal

[Mattia Dongili]

On Tue, Jul 12, 2016 at 06:29:14PM +0200, Ritesh Raj Sarraf wrote:
...

> PS: And I just noticed you did some work with uml-utilities. Thanks.

Yes, I'm trying to upload a new revision.

-- 
mattia
:wq!

Bug#817708: uml-utilities marked for removal

[Mattia Dongili]

On Tue, Jul 05, 2016 at 05:58:58PM +0200, Ritesh Raj Sarraf wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA512
> 
> Hi,
> 
> As per the tracker, uml-utilities is marked for removal from the archive on 
> the
> 19th of July.
> https://tracker.debian.org/pkg/uml-utilities
> 
> 
> Are there any plans for fixing it ?

No. At the moment it doesn't make much sense to keep uml-utilities in
the archive. Even if/when UML is re-added, it should be considered
whether uml-utilities are all that useful.

> PS: I'm trying to bring back user-mode-linux back into the archive. So any 
> help
> on retaining uml-utilities in the archive will be nice.

Thanks for doing that. I'd like to say that I'm trying too, but
effectively I have a lot of ideas but very little time to go after them.
Specifically, the biggest problem I had with the previous packaging was
that it required too much maintenance burden. UML should be one of the
artifacts created by the regular Debian kernel builds.

-- 
mattia
:wq!

Bug#742016: user-mode-linux: build-depends on linux-source-3.12 which is no longer in the archive

[Mattia Dongili]

On Tue, Apr 15, 2014 at 07:44:08PM +0200, Jakub Wilk wrote:
 * Jakub Wilk jw...@debian.org, 2014-04-15, 15:39:
 The package builds successfully and seems to work on i386. I'll test on
 amd64 later.
 
 It works on amd64 works, too. :)
 
 Would it be okay if I NMUed the package?

sure, without any delay would be great. 
Please, make sure this bug report has the latest patch so that I can
import it in the git repository.

Thanks a lot for your help!
-- 
mattia
:wq!


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

Bug#742016: user-mode-linux: build-depends on linux-source-3.12 which is no longer in the archive

[Mattia Dongili]

On Mon, Apr 14, 2014 at 09:48:01PM +0200, Jakub Wilk wrote:
 * Ansgar Burchardt ans...@debian.org, 2014-03-18, 12:04:
 user-mode-linux build-depends on linux-source-3.12 which is no longer in
 the archive. The current version is linux-source-3.13.
 
 Any news on this? Do you need help?

Yes, at the moment I'm busy and cannot work on this (I might in the next
two weeks though). And unfortunately I'm the only active maintainer in
the team.

Updating the user-mode-linux package is just a matter of setting the
correct version in debian/{rules,control}, updating the kernel
configuration (there are debian/rules targets to help with that -
oldconfig{32,64} - policy is M for everything) and building the package.
If there are no major issues, this is it.

Let me know if you can help. Else I can try towards the weekend.

Thanks
-- 
mattia
:wq!


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

Bug#741075: user-mode-linux: Occasional memory corruption on startup under high load

[Mattia Dongili]

On Sat, Mar 08, 2014 at 07:04:56AM +, Anton Ivanov wrote:
 Package: user-mode-linux
 Version: 3.2-2um-1+deb7u2+b1
 Severity: grave
 Tags: patch
 Justification: causes non-serious data loss
 
 Dear Maintainer,
 
 This bug is perennial. If we go through old bugs with
 cannot reproduce tag 50% of them are this one, the other
 50% are the you should not use pipe for interprocess IPC 
 which we will submit shortly.
 
 Manifestation of the problem - UML dies on startup for no
 reason with a memory corruption message. Occurs only on 
 heavily loaded systems and usually when running a lot of 
 UMLs.

Thanks for the patch.
I have noticed that you submitted these patch-set (together with the
other two you sent here and more) upstream and they will be in the
stable branch.
The easiest path here is also to go through the stable release of
linux-source where uml is built from. I'll keep an eye on the stable
tree but it'd be very helpful if you could add the stable tree commit
ids once the patches get included. Same story for the other two bugs.

Thanks!
-- 
mattia
:wq!


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

Bug#712234: user-mode-linux: build-depends on linux-source-3.2

[Mattia Dongili]

On Sun, Oct 20, 2013 at 12:15:59PM +0200, Ansgar Burchardt wrote:
 Ansgar Burchardt ans...@debian.org writes:
  Ansgar Burchardt ans...@debian.org writes:
  user-mode-linux build-depends on linux-source-3.2 which is no longer
  available in unstable.
 
  That's still the case. And user-mode-linux in testing/unstable now seems
  to get security updates via stable as well...
 
user-mode-linux | 3.2-2um-1+deb7u1 | unstable | source, amd64, i386
 
 And nothing has changed so far, except more updates to testing/unstable
 that arrived via stable:
 
   user-mode-linux |3.2-2um-1+deb7u2 |   unstable | source
   user-mode-linux | 3.2-2um-1+deb7u2+b1 |   unstable | amd64, i386
 
 I'm considering to request removal as the package doesn't seem to be
 maintained any longer. There are three reverse dependencies that would

Apologies for the delay, I just uploaded a package based on 3.11.

Unfortunatly the way UML is built right now is far from ideal as it's a
separate package from linux-image and friends and so it's a lot of
chasing and rebuilding.
Trying to integrate the package with the debian kernel build system
requires time that I currently don't have.
-- 
mattia
:wq!


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

Bug#700169: non-free license: requires to obey US export regulation even, when not in the US

[Mattia Dongili]

On Thu, Mar 28, 2013 at 02:17:18PM +0100, Michael Stapelberg wrote:
...
 Can we just ignore this bug for wheezy? To me, the licensing intention
 seems very clear.

Can it be closed even? I don't think this bug applies at all.

-- 
mattia
:wq!


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

Bug#700169: non-free license: requires to obey US export regulation even, when not in the US

[Mattia Dongili]

On Wed, Mar 27, 2013 at 11:47:33AM +0100, Michael Stapelberg wrote:
 Hi Ansgar, Mattia,
 
 Ansgar Burchardt ans...@debian.org writes:
  I also checked the initial Debian package on snapshot.debian.org
  (version 20050930-1). It also has only the non-free license in the
  individual files, but states Dual GPLv2/ACPICA Licence in d/copyright.
  It also has the BSD-3-clause-or-GPL-2 bit in d/copyright.
 
  It's likely that it was already dual-licensed, but that this wasn't
  documented in the tarball itself. I'm not sure why they now have two
  tarballs instead of one with both licenses... The GNU General Public
  License or via a separate license that may be more favorable to
  commercial OSVs (from the FAQ) seems also wrong given there are *three*
  licenses: the non-free one, a 3-clause BSD and the GPL-2
 Well, according to https://github.com/acpica/acpica/commit/84b8d0fd, the
 dual-license tarballs are only available starting from version
 20110211. That version can indeed be downloaded as unix2 tarball.
 
 Mattia: is it reasonable to update this package to a newer version,
 based on one of the unix2 tarballs?

yes it is, that's what Al did already:
http://ftp-master.debian.org/new/acpica-unix_20130214-0.3.html

In any case, most of the code of the old packages have been included in
the linux kernel for years and the original download page states:

The Linux package includes the same functionality as the
previous two, but has been modified to integrate smoothly with
the Linux kernel source. This includes conversion of the ACPI CA
source code to the Linux kernel coding standard, and licensing
under the GNU General Public License.
...

Linux
The latest IASL compiler for Linux can be built from the Unix
source package:
download acpica-unix-VERSION.tar.gz
$ tar xzf acpica-unix-VERSION.tar.gz
$ cd acpica-unix-VERSION/compiler
$ make

Starting with Linux kernel 2.4, ACPI CA is in the Linux kernel.

(taken from a randomly old snapshot:
http://web.archive.org/web/20050911035003/http://developer.intel.com/technology/iapc/acpi/downloads.htm)

from the commit log you referenced the explicit licence on source files
was requested by FreeBSD but ACPICA has always been dual licensed.

-- 
mattia
:wq!


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

Bug#700169: non-free license: requires to obey US export regulation even, when not in the US

[Mattia Dongili]

On Wed, Mar 27, 2013 at 05:06:35PM +, Adam D. Barratt wrote:
 On 27.03.2013 13:44, Michael Stapelberg wrote:
 Mattia Dongili malat...@debian.org writes:
 yes it is, that's what Al did already:
 http://ftp-master.debian.org/new/acpica-unix_20130214-0.3.html
 I see.
 
 release-team: What’s your take on this? Can we get the new version
 into
 Debian in time for wheezy or how should we handle this?
 
 It's somewhat difficult to tell without seeing what's involved. (We
 can't exactly debdiff against NEW...)

Michael,
I don't see a valid reason to get a newer version in wheezy at this
stage of the freeze.

Regards,
-- 
mattia
:wq!


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

Bug#697300: src:user-mode-linux: missing Built-Using on linux (for linux-source-*)

[Mattia Dongili]

On Fri, Jan 04, 2013 at 10:12:07AM +0100, Ansgar Burchardt wrote:
 Hi,
 
 I looked at your changes and found one error:
 
 +BUILT_USING := $(shell dpkg-query -f '$${binary:Package} (= 
 $${source:Version}), ' -W linux-source-$(kernel_version))
 
 It needs to be $${source:Package} here as the Built-Using fields
 references other source packages.  You should get something like
 
   Built-Using: linux (= 3.2.35-2)
 
 in the resulting binary package.

Yes, I actually changed it on purpose compared to the gcc-avr package
you pointed at.
Re-reading the policy though I get your point. I'll change that before
upload.

Thanks for checking
-- 
mattia
:wq!


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

Bug#697300: src:user-mode-linux: missing Built-Using on linux (for linux-source-*)

[Mattia Dongili]

tags 697300 + pending
thanks

On Thu, Jan 03, 2013 at 07:21:19PM +0100, Ansgar Burchardt wrote:
 Package: src:user-mode-linux
 Version: 3.2-1um-1
 Severity: serious

Thanks, I made the necessary changes to the package, I will upload a
fixed version later today.
-- 
mattia
:wq!


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

Bug#599950: acpidump: FTBFS on ia64: error: unknown register name 'ebx' in 'asm'

[Mattia Dongili]

On Sat, Oct 30, 2010 at 12:15:20PM +0200, Mike Hommey wrote:
 On Tue, Oct 12, 2010 at 01:55:26PM +0200, Cyril Brulebois wrote:
  Source: acpidump
  Version: 20100513-1
  Severity: serious
  Justification: FTBFS
  User: debian-i...@lists.debian.org
  Usertags: ia64
  
  Hi,
  
  your package no longer builds on ia64:
  | make[2]: Entering directory 
  `/build/buildd-acpidump_20100513-1-ia64-CL0nDr/acpidump-20100513/turbostat'
  | cc -Wall -g -O2 turbostat.c -o turbostat
...
 Note it doesn't make sense to build turbostat on ia64 as it is a tool to
 gather data on intel CPUs with the turbo boost technology. AFAIK, no
 ia64 CPU has that.

yes, you are right. let's see if I can find a minute to fix the package
to skip turbostat on ia64.

thanks
-- 
mattia
:wq!



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

Bug#587609: new version breaks virtualbox-ose compilation

[Mattia Dongili]

retitle 587609 FTBFS: DSDT errors
reassign 587609 virtualbox-ose
thanks

On Wed, Jun 30, 2010 at 12:12:22PM +0200, Michael Meskes wrote:
 Package: iasl
 Version: 20100528-2
 Severity: serious
 Tags: sid
 
 Since upgrading iasl virtualbox-ose doesn't compile anymore:
 
 ...
 kBuild: iasl DevicesR3 - 
 /home/michael/technik/sources/archive/virtualbox-ose/virtualbox-ose/src/VBox/Devices/PC/vbox.dsl
 /home/michael/technik/sources/archive/virtualbox-ose/virtualbox-ose/src/VBox/Devices/PC/vbox.dsl
998:  0xdfdf,  // Range Length 
 (calculated
 Error4118 -   
 Length is not equal to fixed Min/Max window ^ 
 
 ASL Input:  
 /home/michael/technik/sources/archive/virtualbox-ose/virtualbox-ose/src/VBox/Devices/PC/vbox.dsl
  - 1305 lines, 46189 bytes, 288 keywords
 Compilation complete. 1 Errors, 0 Warnings, 0 Remarks, 404 Optimizations
 ...

where can I find this DSDT?

 Please note that I do not know if this is a iasl problem or a virtualbox-ose
 one, although all older versions up to 20090521-1 worked without a problem. I
 set this report to serious just in case it's an iasl problem to prevent
 migration, so we have a working version left in squeeze.

A couple of points to note:
- 20090521 didn't have these sanity checks on resource templates
- the aml file should be generated despite the error(s) that iasl reports

Ok, never mind. While searching for the DSDT file to take a look myself
I found this:
http://www.mail-archive.com/freebsd-emulat...@freebsd.org/msg00196.html
read the whole thread it appears VBox's DSDT has a number of problems.

Bug reassigned accordingly
-- 
mattia
:wq!



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

Bug#539427: Segmentation fault of iasl on non i386/amd64

[Mattia Dongili]

On Tue, Oct 27, 2009 at 10:11:10PM +0900, Mattia Dongili wrote:
 On Sun, Oct 25, 2009 at 01:15:06AM +0200, Guillem Jover wrote:
  Hi!
  
  [ Removing #516057 as it's a closed bug about a bison error. ]
  
  On Sat, 2009-10-24 at 18:19:51 +0200, Luk Claes wrote:
   Can you please have a look at fixing the segfaults of iasl on non
   i386/amd64 or do you think it would be better to remove the support of
   iasl on non i386/amd64. If so, bochs is the only affected reverse dep of
   iasl if you would restrict iasl to i386/amd64.
  
  I don't see why iasl would not be able to run on any architecture,
  it's just a compiler for byte-code. We ported it some time ago, but
  it seems it has regressed in the latest upload (probably problems with
  unaligned accesses or little endian assumptions, as before).

I uploaded a new package with a pile of fixes for BE support. Let's see
how it goes.
-- 
mattia
:wq!



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

Bug#577822: cpufreqd: causes hardware damage by ignoring sensor lines after removing libsenors3

[Mattia Dongili]

severity 577822 normal
tags 577822 - patch
thanks

On Wed, Apr 14, 2010 at 11:04:50PM +0200, Helmut Grohne wrote:
 Package: cpufreqd
 Version: 2.4.1-1
 Severity: critical
 Tags: patch
 Justification: causes hardware damage
 
 Since no package on my system depends on libsensors3 I removed it. With
 it the file /etc/sensors.conf went away. This is a file cpufreqd
 requires to use for the sensors plugin. Without that file it will not
 trigger on temperature changes and possibly cause your hardware to
 overheat.

Would you mind describing what you're actually experiencing rather than
an ipotetical (and quite unlinkely) situation?
I can assume you're having problems with the sensors plugin from your
submission...

Also please, provide logs (running cpufreqd with -V7 generates a lot of
useful output to trak down issues).

 A quick solution is to depend on libsensors3 as it provides the required
 file.

cpufreqd depends on libsensors4, have you considered using
/etc/sensors3.conf?
I guess a possible issue here is that cpufreqd _requires_ a sensors
config file while it's not really necessary to use one when initializing
the library..

-- 
mattia
:wq!



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

Bug#557420: cpufrequtils: Tries to remove non-existing symlinks

[Mattia Dongili]

forcemerge 557420 557410
thanks

On Sun, Nov 22, 2009 at 12:25:08AM +0100, Michael Biebl wrote:
 Package: cpufrequtils
 Version: 006-1
 Severity: serious
 Justification: fails to upgrade
 
 pluto:~# ls /etc/rc?.d/???cpufrequtils
 /etc/rc2.d/S20cpufrequtils  /etc/rc3.d/S20cpufrequtils
 /etc/rc4.d/S20cpufrequtils  /etc/rc5.d/S20cpufrequtils

yeah, shame on me. 006-2 is queued since last night already.

-- 
mattia
:wq!



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

Bug#557410: cannot remove `/etc/rc0.d/K20cpufrequtils': No such file or directory

[Mattia Dongili]

On Sat, Nov 21, 2009 at 11:41:05PM +0100, Frank Gevaerts wrote:
 Package: cpufrequtils
 Version: 006-1
 Severity: normal
 
 Seems to be caused by
 
 [ -f $link ] || rm $link
 
 which obviously needs to be  instead

Spotted :)
the new package is queued and built on most architectures already.

Regards,
-- 
mattia
:wq!



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

Bug#557252: negate the match for keyboard capabilities

[Mattia Dongili]

severity 557291 grave
forcemerge 557252 557291
thanks

On Fri, Nov 20, 2009 at 03:20:54PM -0500, Aleksey Kliger wrote:
 --- 11-x11-synaptics.fdi.orig 2009-11-20 15:18:09.0 -0500
 +++ 11-x11-synaptics.fdi  2009-11-20 15:08:49.0 -0500
 @@ -4,7 +4,7 @@
  match key=info.capabilities contains=input.touchpad
!-- do not use the synaptics driver for devices advertising
  themselves as keyboards --
 -  match key=info.capabilities contains=input.keyboard
 +  match key=info.capabilities contains_not=input.keyboard
merge key=input.x11_driver type=stringsynaptics/merge

apologies, turns out that the laptop where I tested the change had a
custom fdi file that was adding the synaptics driver anyway.

I'm about to upload a fixed package.

-- 
mattia
:wq!



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

Bug#539427: Segmentation fault of iasl on non i386/amd64

[Mattia Dongili]

On Sun, Oct 25, 2009 at 01:15:06AM +0200, Guillem Jover wrote:
 Hi!
 
 [ Removing #516057 as it's a closed bug about a bison error. ]
 
 On Sat, 2009-10-24 at 18:19:51 +0200, Luk Claes wrote:
  Can you please have a look at fixing the segfaults of iasl on non
  i386/amd64 or do you think it would be better to remove the support of
  iasl on non i386/amd64. If so, bochs is the only affected reverse dep of
  iasl if you would restrict iasl to i386/amd64.
 
 I don't see why iasl would not be able to run on any architecture,
 it's just a compiler for byte-code. We ported it some time ago, but
 it seems it has regressed in the latest upload (probably problems with
 unaligned accesses or little endian assumptions, as before).

yeah, dispite the fact that the patch applied, there were other changes
that are causing a consistent segfault.

 I might try to take a look at fixing it, but not now. Anyway regarding
 bochs, it's not really a problem as iasl is only used when building
 architecture independent packages, thus the Build-Depends-Indep.

I tried debugging it when the bug first appeared after I uploaded the
package with not much success but now I'm definitely lacking the time to
look at it.
As an emergency fix, if it's not a big deal for bochs, disabling the
failing architectures is an option.

thanks
-- 
mattia
:wq!



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

Bug#537208: Fails to upgrade (cannot access `/usr/lib/dbus-1.0/dbus-daemon-launch-helper': No such file or directory)

[Mattia Dongili]

severity 537208 serious
retitle  537208 Fails to upgrade (cannot access 
`/usr/lib/dbus-1.0/dbus-daemon-launch-helper': No such file or directory)
thanks

confirmed, today's upgrade here failed due to dbus.

Cheers
-- 
mattia
:wq!



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

Bug#527711: xf86-input-evtouch: FTBFS: evtouch.c:33:25: error: xf86Version.h: No such file or directory

[Mattia Dongili]

On Thu, May 14, 2009 at 08:36:26PM +0900, Mattia Dongili wrote:
 On Wed, May 13, 2009 at 05:52:01PM +0200, Julien Cristau wrote:
...
  Also debian/copyright confuses me, it says the code is MIT, but somehow
  upstream has GPL-2 in COPYING, and the above patch says it's GPL as
  well.  Could you clarify this?
 
 what a mess! the source files are definitly MIT, I'm pretty sure I asked
 upstream about the COPYING file long time ago, I'll see if I can find
 that discussion again, or even ask him again.
 Although I'm quite sure our conversation ended up with him saying the
 software is indeed MIT licensed.

I found the old discussion. At that time I pointed out the misleading
COPYING file, but I guess he never got around to amend it.

I'll point it out again.
-- 
mattia
:wq!



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

Bug#527711: xf86-input-evtouch: FTBFS: evtouch.c:33:25: error: xf86Version.h: No such file or directory

[Mattia Dongili]

On Sun, May 17, 2009 at 07:42:17PM +0200, Julien Cristau wrote:
 On Thu, May 14, 2009 at 13:51:31 +0200, Julien Cristau wrote:
 
  On Thu, May 14, 2009 at 20:36:26 +0900, Mattia Dongili wrote:
  
   On Wed, May 13, 2009 at 05:52:01PM +0200, Julien Cristau wrote:
Hi Mattia,

I started to look at this as evtouch is one of the few drivers not
building against server 1.6, but I'm a bit stuck.
Patch 06_xf86-input-evtouch-0.8.7-misc.patch doesn't apply on 0.8.8,
http://www.postnuklear.de/xorg-patches/ hasn't been updated in quite a
long time.
   
   yeah, I think there is no need for that patch in 0.8.8.
   I removed it from git a while ago after I imported the 0.8.8 sources.
   
  Bah it looks like I somehow forgot to git fetch yesterday before looking
  at it, so I didn't notice you had already imported 0.8.8.  Sorry about
  that.
  
 After a bit more patching, I got it to build.  Can you check/upload, or
 should I do it (I don't have any way to test the package though)?

I can test it later tonight. If everything is good I'll upload the
package straight away.

Thanks
-- 
mattia
:wq!



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

Bug#527711: xf86-input-evtouch: FTBFS: evtouch.c:33:25: error: xf86Version.h: No such file or directory

[Mattia Dongili]

On Wed, May 13, 2009 at 05:52:01PM +0200, Julien Cristau wrote:
 Hi Mattia,
 
 I started to look at this as evtouch is one of the few drivers not
 building against server 1.6, but I'm a bit stuck.
 Patch 06_xf86-input-evtouch-0.8.7-misc.patch doesn't apply on 0.8.8,
 http://www.postnuklear.de/xorg-patches/ hasn't been updated in quite a
 long time.

yeah, I think there is no need for that patch in 0.8.8.
I removed it from git a while ago after I imported the 0.8.8 sources.

 Also debian/copyright confuses me, it says the code is MIT, but somehow
 upstream has GPL-2 in COPYING, and the above patch says it's GPL as
 well.  Could you clarify this?

what a mess! the source files are definitly MIT, I'm pretty sure I asked
upstream about the COPYING file long time ago, I'll see if I can find
that discussion again, or even ask him again.
Although I'm quite sure our conversation ended up with him saying the
software is indeed MIT licensed.

 (debian/copyright also has an outdated upstream url,
 http://www.conan.de/touchscreen/evtouch.html has the new release)
 
 ubuntu has 0.8.8 building on new xserver, so that would probably be
 worth looking at as well.

Sure, I'll take a look.
Thanks
-- 
mattia
:wq!



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

Bug#516057: NMU?

[Mattia Dongili]

On Wed, Mar 18, 2009 at 09:15:04PM -0400, Mike O'Connor wrote:
 Maintainer,
 
 The attached patch does indeed seem to get this package building again.
 Do you intend to make an uplaod soon?  Would you like to have this bug
 NMUed?

Feel free to go on but make sure that the test case runs successfully.

Thanks
-- 
mattia
:wq!


signature.asc
Description: Digital signature

Bug#510656: [Pkg-uml-pkgs] unmerging 478717, cloning 478717, reassign -1 to linux-2.6, found -1 in 2.6.18.dfsg.1-23etch1 ...

[Mattia Dongili]

On Sun, Jan 04, 2009 at 02:21:36AM +, Ben Hutchings wrote:
 unmerge 478717
 clone 478717 -1 -2 -3
...
 reassign -3 user-mode-linux 
 found -3 2.6.18-1um-2etch.23etch1
 found -3 2.6.26-1um-2
 fixed -3 2.6.18-1um-2etch.24

not sure UML is really affected by this bug, after all UML is x86 and
x86_64 only ;)
-- 
mattia
:wq!



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

Bug#509355: cpufreqd: segfault while reading acpi battery info

[Mattia Dongili]

On Tue, Dec 23, 2008 at 07:57:00PM +0100, Stefan Bühler wrote:
 The problem is that check_timeout is reset after the first battery got
 updated, so the values for the second battery are never read.
 
 This results in status-value == NULL, and strncmp segfaults.
 
 The solution is to update check_timeout after all battery values are updated.

excellent, thanks
-- 
mattia
:wq!


signature.asc
Description: Digital signature

Bug#509355: cpufreqd: segfault while reading acpi battery info

[Mattia Dongili]

On Sun, Dec 21, 2008 at 04:40:28PM +0100, Jonas Zeiger wrote:
 Package: cpufreqd
 Version: 2.3.3-3
 Severity: grave
 Justification: renders package unusable
 
 
 I've been trying to run cpufreqd on a Dell Latitude D530
 Laptop with a Intel Core2Duo 2.00Ghz CPU, 3GB of DDR2 RAM
 and two attached battery units using the stock Debian kernel
 version 2.6.26, but the problem is exactly the same with my
 custom kernel 2.6.26zlinux (tuned for install on SSD).
 
 Cpufreqd keeps segfaulting when starting up the system:
...
 acpi_battery_init: found 2 Batteries
 acpi_battery_update  : BAT0 - present
 read_battery : BAT0 - reading battery levels
 read_battery : BAT0 - remaining capacity: 520
 acpi_battery_update  : battery life for BAT0 is 109%
 acpi_battery_update  : BAT1 - present
 acpi_battery_update  : BAT1 - estimating battery life (timeout: 30.00 - 
 status: (null))
 Speicherzugriffsfehler   /* german locale, it' a segfault! :(  */

can you run cpufreqd in a gdb session (make sure to start it with -D so
that is doesn't deamonize) and get a stack trace of the segfault?
Also, don't forget to build it with debug symbols.

 Ok, i've already tried to patch cpufreqd_2.3.3-3 with some patch found
 on the net:
 
 (Downloaded cpufreqd_2.3.3-3.diff and cpufreqd_2.3.3.orig.tar.gz)
 $ tar xvzf cpufreqd_2.3.3.orig.tar.gz
 $ patch -p0  cpufreqd_2.3.3-3.diff
 $ cd cpufreqd-2.3.3
 $ sudo apt-get install libsensors-dev libcpufreqd-dev
 $ patch -p0
 Index: src/cpufreqd_acpi.c
 ===
 RCS file: /cvsroot/cpufreqd/sources2/src/cpufreqd_acpi.c,v
 retrieving revision 1.8
 diff -u -r1.8 cpufreqd_acpi.c
 --- src/cpufreqd_acpi.c 31 Aug 2008 01:17:25 -  1.8
 +++ src/cpufreqd_acpi.c 1 Oct 2008 12:22:23 -
 @@ -246,8 +246,8 @@
  
 /* read `clsname` devices */
 devs = sysfs_get_class_devices(cls);
 -   if (!cls) {
 -   clog(LOG_INFO, class '%s' not found (%s)\n, clsname,
 +   if (!devs) {
 +   clog(LOG_INFO, class device '%s' not found (%s)\n, clsname,
 strerror(errno));
 sysfs_close_class(cls);
 return -1;

this shouldn't even apply as it's already applied to the debian package
when building.

thanks
-- 
mattia
:wq!



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

Bug#495655: cpufreqd: uses energy_full instead of charge_full under /sys/...

[Mattia Dongili]

On Wed, Aug 20, 2008 at 09:19:06AM +0200, alberto maurizi wrote:
 Package: cpufreqd
 Followup-For: Bug #495655
 
 
 the error message is:
 cpufreqd: get_class_device_attribute: couldn't open 
 /sys/class/power_supply/BAT0/energy_full (No such file or directory)
 [...]
 kernel: [19344.165939] cpufreqd[24816]: segfault at 40 ip b7e04283 sp 
 bfb30770 error 4 in libc-2.7.so[b7d8e000+155000]
 
 under /sys/class/power_supply/BAT0/ path I can see a charge_full
 rather than energy_full. Maybe some inconsistency in naming
 convention?

well, apparently both of them are acceptable names.
There is already a patch available attached to this bug report
bus.debian.org/495655

I'm going to upload a new version as soon as possible
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#484467: cpufreqd, 484467 (grave) and CONFIG_ACPI_PROCFS_POWER in Lenny

[Mattia Dongili]

On Tue, Aug 12, 2008 at 09:59:23PM +0200, Luk Claes wrote:
 Mattia Dongili wrote:
 Hi,

 Hi

 from the debian linux-2.6 svn repository it looks like
 CONFIG_ACPI_PROCFS_POWER is not going to be set for 2.6.26.

 cpufreqd has a grave bug filed (#484467) for the broken ACPI battery
 level support.
...

 Sorry for the late answer.

 I unblocked cpufreqd and set age-days to 15.

Ohey! no worries and actually thanks. I was going to ping the list
again once I verified that cpufreqd was running with no major issues in
unstable.

thanks a lot.
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#494721: [Pkg-uml-pkgs] Bug#494721: linux-patch-skas: Doesn't apply against Lenny release kernel

[Mattia Dongili]

On Mon, Aug 11, 2008 at 07:52:58PM +0200, Moritz Muehlenhoff wrote:
 Package: linux-patch-skas
 Severity: grave
 Justification: renders package unusable
 
 This package only provide a kernel patch up to 2.6.18 and doesn't
 apply against later kernels (especially due to the unification
 of the x86 architecture).

I'm filing a request for removal now. I haven't seen updates to the
skas3 patchset in quite a long time and development has moved to skas4
which, at some point, will be pushed upstream.

cheers
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#484467: cpufreqd, 484467 (grave) and CONFIG_ACPI_PROCFS_POWER in Lenny

[Mattia Dongili]

Hi,

from the debian linux-2.6 svn repository it looks like
CONFIG_ACPI_PROCFS_POWER is not going to be set for 2.6.26.

cpufreqd has a grave bug filed (#484467) for the broken ACPI battery
level support.
I (with the cpufreqd upstream hat on) released a new cpufreqd (2.3.0)
that reads from sysfs rather than /proc but the change is quite invasive
and I (with the cpufreqd maintainer hat on) am now wondering if I should
upload it (which I will do anyway in the next days) and ask for a freeze
exception or not.
The other option would be to ask for CONFIG_ACPI_PROCFS_POWER to be
enabled in Lenny which I'm unsure if it's a good idea but it'd also
allow battery support in gkrellm (#463779).

The diff for the new cpufreqd version is attached (gzipped), in
short it touches most of the acpi info reading code to make it go to
sysfs (using libsysfs2) plus some trivial fixes here and there.

Any suggestion?
-- 
mattia
:wq!


cpufreqd-2.3.diff.gz
Description: Binary data

Bug#491921: user-mode-linux: FTBFS on amd64 -- no more __memcpy with gcc 4.3

[Mattia Dongili]

On Tue, Jul 22, 2008 at 03:20:41PM -0400, Aaron M. Ucko wrote:
 Package: user-mode-linux
 Version: 2.6.25-1um-1
 Severity: serious
 Tags: patch
 Justification: no longer builds from source
 
 user-mode-linux fails to build on amd64 with the following error:
 
 |   CC  arch/um/sys-x86_64/ksyms.o
 | arch/um/sys-x86_64/ksyms.c:16: error: '__memcpy' undeclared here (not in a 
 function)
 | arch/um/sys-x86_64/ksyms.c:16: warning: type defaults to 'int' in 
 declaration of '__memcpy'
 | make[2]: *** [arch/um/sys-x86_64/ksyms.o] Error 1
 | make[1]: *** [arch/um/sys-x86_64] Error 2
 
 The problem is that arch/um/sys-x86_64/ksyms.c unconditionally tries
 to export __memcpy even though include/asm-x86/string_64.h declares it
 only for GCC 4.2 and older, newer versions presumably being smart
 enough not to need it.  I'm attaching a patch that conditionalizes the
 EXPORT_SYMBOL call accordingly, and have confirmed that it allows the
 build to complete without errors.  (FTR, arch/um/os-Linux/user_syms.c
 already exports memcpy itself, so ksyms.c need not do so in any case.)
 
 FWIW, you can find a full log illustrating the error at
 http://buildd.debian.org/fetch.cgi?pkg=user-mode-linux;ver=2.6.25-1um-1;arch=amd64;stamp=1216481207
 
 -- System Information:
 Debian Release: lenny/sid
   APT prefers unstable
   APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable')
 Architecture: amd64 (x86_64)
 
 Kernel: Linux 2.6.25.11 (SMP w/2 CPU cores)
 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1)
 Shell: /bin/sh linked to /bin/dash

 Index: linux-source-2.6.25/arch/um/sys-x86_64/ksyms.c
 ===
 --- linux-source-2.6.25.orig/arch/um/sys-x86_64/ksyms.c   2008-04-16 
 22:49:44.0 -0400
 +++ linux-source-2.6.25/arch/um/sys-x86_64/ksyms.c2008-07-22 
 14:10:47.0 -0400
 @@ -13,4 +13,6 @@
  EXPORT_SYMBOL(__up_wakeup);
  
  /*XXX: we need them because they would be exported by x86_64 */
 +#if (__GNUC__ == 4  __GNUC_MINOR__  3) || __GNUC__  4
  EXPORT_SYMBOL(__memcpy);
 +#endif

current linux git uses a slightly different solution
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=8bfd04b974689f700bbd053ad6e66b0a95fb80c9
which is after all conditionally exports memcpy or __memcpy with the
same logic as their definition.

Will apply, and reupload later today. Thanks.
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#484467: query power status in /sys not in /proc

[Mattia Dongili]

retitle 484467 cpufreqd: Depends on obsolete /proc/acpi interface
thanks

On Sat, Jun 28, 2008 at 10:36:26PM +0100, Ben Hutchings wrote:
 /proc/acpi is scheduled for removal in Linux in July 2008, so there is
 little point in reenabling it in the Debian kernel now.
 
 cpufreqd looks dead upstream, so it may be time to remove it.  There

I am upstream, can't say I'm actively developing cpufreqd these days
(months? years?) but changing it to read from /sys is not a huge effort.
Let me see if I can come up with something.

 would need to be a release note to make sure people remove it before
 upgrading if this shutdown rule is part of the standard configuration.

the shutdown rule is not part of the standard configuration, at all (I'm
not that crazy).

cheers
-- 
mattia
:wq!


signature.asc
Description: Digital signature

Bug#484467: query power status in /sys not in /proc

[Mattia Dongili]

On Tue, Jun 10, 2008 at 12:22:19PM +0200, psycheye wrote:
 In the last kernels (2.6.24.x) proc is setted how obsolete, there is /sys 
 directory.
 Using 2.6.25.4, could I've any problems?

apparently yes...

One could argue that removing /proc/acpi now may break other programs
and that it would be nice to have it enabled in the kernel for some more
time.

bah
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#484467: query power status in /sys not in /proc

[Mattia Dongili]

On Tue, Jun 10, 2008 at 01:51:34PM +0200, psycheye wrote:
  One could argue that removing /proc/acpi now may break other programs
  and that it would be nice to have it enabled in the kernel for some more
  time.
 doing: cpufreq-get I see:
 No cpufreqd socket found
 
 (but in /var/run there's the pid about cpufreqd) 

you need a unix socket, not a pid file :)
you can enable this feature in the config file (see also man
cpufreqd.conf)

 so, if I try to set cpu clock with cpufreq-selector (cpufreq-selector -c 0 -c 
 1 -f 800) the command run, but the cpu clock doesn't change :-/

oh, this is not from the cpufreqd package and actually conflicts with
what cpufreqd thinks the current frequency is.

 So, I don't understand if No cpufreqd socket found is about the bug of 
 package or other errors.

no, unrelated. This bug is about the /proc/acpi directory not being
available in recent debian kernels.

 What I will do?

see above. Enable the unix socket and try again.

cheers
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#484467: query power status in /sys not in /proc

[Mattia Dongili]

On Fri, Jun 06, 2008 at 10:56:53AM +0200, maximilian attems wrote:
 reassign  484467 cpufreqd
 severity 484467 serious
 retitle 484467 query power status in /sys not in /proc
 stop
 
 
 On Fri, Jun 06, 2008 at 08:11:16AM +0200, LÉVAI Dániel wrote:
  It reboots because there is a rule in /etc/cpufreqd.conf which halts the 
  system when it detects a 0-1% battery charge. cpufreqd uses the 
  legacy /proc way to gather battery status information, thus with the 
  new kernel it doesn't work, and instantly halts because it thinks that 
  no juice left.
 
 ok thanks a lot for the info, reassigning properly.
 up to the cpufreqd maintainer to react.

Any chance you could instead re-enable the deprecated /proc/acpi
support?
I don't think cpufreqd is the only one breaking.

Cheers
-- 
mattia
:wq!



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#461760: xserver-xorg-input-synaptics: Synaptics touchpad completely stops working under gdm, xdm, kdm ... and X (gnome, openbox, kde...)

[Mattia Dongili]

On Sat, Apr 26, 2008 at 10:24:38AM +0200, David Paleino wrote:
 On Tue, 22 Apr 2008 20:55:07 +0200, Julien Cristau wrote:
 
  Hi,
 
 Hi,
 
  you reported a problem with the synaptics driver not reporting any
  events for your touchpad.
 
 Well, not detecting the touchpad at all would have been a better
 description ;)
 
  Can you attach the contents of /proc/bus/input/devices to this bug?
 
 Sure, but I can anticipate that the touchpad isn't listed in that file.

then the Xorg driver has very little to do here and this sounds like a
different bug. (Sorry I never replied to your other bug submission)

...
 I can't make the touchpad working now, even with the versions reported above. 
 I
 suspect it's something udev(?)-related, or, however, with hardware detection. 
 :(

I agree here.
If the touchpad is not listed in /proc/bus/input/devices then is't most
probably kernel related.
Did you notice anything suspicious in your kernel log?

cheers
-- 
mattia
:wq!


signature.asc
Description: Digital signature

Bug#443726: Fixed upstream?

[Mattia Dongili]

On Sat, Feb 02, 2008 at 01:33:29PM +0900, Mattia Dongili wrote:
 On Wed, Jan 23, 2008 at 09:18:23PM +0900, Mattia Dongili wrote:
  On Wed, Jan 23, 2008 at 12:23:24PM +0100, Per Olofsson wrote:
   Hi,
   
   Upstream's CHANGELOG[1] says:
   
   0.8.8
 - Make it work for current Xorg-GIT
   
   However, no version 0.8.8 has been released yet. Maybe we can ask 
   upstream for a
   patch in the mean time?
  
  sure!!! will do.
 
 sigh... no answer from upstream yet. And no 0.8.8 has been released yet.

wow, I nic euser pointed to this:
http://www.postnuklear.de/xorg-patches/files/xf86-input-evtouch-0.8.7-misc.patch
(description at: http://www.postnuklear.de/xorg-patches/)

I'll give it a try this weekend.

hurray
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#461760: xserver-xorg-input-synaptics: I confirm this bug

[Mattia Dongili]

On Thu, Jan 31, 2008 at 11:53:11PM +0100, Michael Schmitt wrote:
 Package: xserver-xorg-input-synaptics
 Version: 0.14.7~git20070706-2
 Followup-For: Bug #461760
 
 I can confirm this bug. It seems this does not break all touchpads but at 
 least some.
 I have a ALPS touchpad in a Toshiba Satellit 2100 Pro and after uprade it 
 breaks the
 touchpad completely. Downgrade to the current lenny version 0.14.6-1 fixes 
 this issue.
 Here are the refering fragments of Xorg.0.log with a working touchpad (lennys 
 package):
...
 (EE) No Input driver matching `synaptics'

yep, as you can see the synaptics driver is completely bypassed and a
different one is taking over.

...
 (II) Synaptics touchpad driver version 0.14.6 (1406)
 (--) ALPS Touchpad auto-dev sets device to /dev/input/event3
 (**) Option Device /dev/input/event3
 (**) Option SHMConfig On
 (**) Option LeftEdge 150
 (**) Option RightEdge 900
 (**) Option TopEdge 170
 (**) Option BottomEdge 680
 (**) Option VertScrollDelta 50
 (**) Option HorizScrollDelta 50
 (**) Option EdgeMotionMinSpeed 500
 (**) Option EdgeMotionMaxSpeed 500
 (**) Option FingerLow 25
 (**) Option FingerHigh 30
 (**) Option MaxTapTime 180
 (**) Option MaxTapMove 110
 (**) Option EmulateMidButtonTime 75
 (**) Option UpDownScrolling 1
 (**) Option LeftRightScrolling 1
 (**) Option LockedDrags 0
 (**) Option RTCornerButton 3
 (**) Option LTCornerButton 2
 (**) Option CircularScrolling 0
 (**) Option PalmDetect 1
 (**) Option PalmMinWidth 10
 (**) Option PalmMinZ 20
 (--) ALPS Touchpad touchpad found
 (**) Option AlwaysCore
 (**) ALPS Touchpad: doesn't report core events
 (II) evaluating device (ALPS Touchpad)
 (II) XINPUT: Adding extended input device ALPS Touchpad (type: MOUSE)
 (II) evaluating device (USB Mouse)
 (II) XINPUT: Adding extended input device USB Mouse (type: MOUSE)
 (II) evaluating device (Keyboard0)
 (II) XINPUT: Adding extended input device Keyboard0 (type: KEYBOARD)
 (--) ALPS Touchpad auto-dev sets device to /dev/input/event3
 (**) Option Device /dev/input/event3
 (--) ALPS Touchpad touchpad found

this really puzzles me... completely dead?
does synclient report any activity?

 Your assumption was true... if xorg complains about the synaptics module
 it works (another driver taking its part?) and if the synaptics module
 recognizes the touchpad it does not work. Right now I realised those
 touchpad specific options (scrolling at the edges etc.) do not work anymore
 either (with lennys package) what makes sense to me. But I am almost 100%
 sure it worked completely.

what happens if you just try to run with the default options?
i.e.: comment everything except

 Section InputDevice
Driver  synaptics
Identifier  ALPS Touchpad
Option  Protocol  auto-dev
Option  Device/dev/input/mouse1
Option  Always Core   1
Option  SHMConfig On
 EndSection

thanks for the report
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#443726: Fixed upstream?

[Mattia Dongili]

On Wed, Jan 23, 2008 at 09:18:23PM +0900, Mattia Dongili wrote:
 On Wed, Jan 23, 2008 at 12:23:24PM +0100, Per Olofsson wrote:
  Hi,
  
  Upstream's CHANGELOG[1] says:
  
  0.8.8
  - Make it work for current Xorg-GIT
  
  However, no version 0.8.8 has been released yet. Maybe we can ask upstream 
  for a
  patch in the mean time?
 
 sure!!! will do.

sigh... no answer from upstream yet. And no 0.8.8 has been released yet.

-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#443726: Fixed upstream?

[Mattia Dongili]

On Wed, Jan 23, 2008 at 12:23:24PM +0100, Per Olofsson wrote:
 Hi,
 
 Upstream's CHANGELOG[1] says:
 
 0.8.8
   - Make it work for current Xorg-GIT
 
 However, no version 0.8.8 has been released yet. Maybe we can ask upstream 
 for a
 patch in the mean time?

sure!!! will do.

Thanka a lot for the note
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#461760: xserver-xorg-input-synaptics: Synaptics touchpad completely stops working under gdm, xdm, kdm ... and X (gnome, openbox, kde...)

[Mattia Dongili]

On Sun, Jan 20, 2008 at 05:18:20PM +0100, Laurent CARON wrote:
 Julien Cristau wrote:
 You don't say which version works for you.  You also didn't send your X
 log.  Please fix that (ideally with logs for both the working and
 non-working versions).

 Hi,

 Working version: working.log
 By working i mean: X starts, but the synaptics functionality is broken
 The working version is:
 xserver-xorg-input-synaptics_0.14.6-1_i386.deb

 Non-Working version: non-working.log

hmmm. Odd, you non-working version has no errors while your working has.
I guess in the working case the device is taken by a different driver.

This is from the working.log:

 (II) LoadModule: synaptics
 (II) Loading /usr/lib/xorg/modules/input//synaptics_drv.so
 dlopen: /usr/lib/xorg/modules/input//synaptics_drv.so: undefined symbol: 
 miPointerGetMotionEvents
 (EE) Failed to load /usr/lib/xorg/modules/input//synaptics_drv.so
 (II) UnloadModule: synaptics
 (EE) Failed to load module synaptics (loader failed, 7)

while the non-working has everything in place.
What does 
  synclient -l
issue with the current xserver-xorg-input-synaptics?
Also, do you have any gsynaptics/ksynaptics/qsynaptics installed?

Thanks
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#457931: /etc/init.d/cpufrequtils fails because of translations

[Mattia Dongili]

On Thu, Dec 27, 2007 at 12:03:44PM +0100, Torsten Marek wrote:
 Package: cpufrequtils
 Version: 002-6
 Severity: serious
 
 Hi,
 
 because of #449307, cpufrequtils now uses cpufreq-info to enumerate CPUs and 
 cores. Unfortunately,
 the sed pattern only works for the English messages of cpufreq-info and fails 
 if LANG != (C|en) and
 a translation is present (which is the case for de, fr  it). In this case, 
 the init script
 fails without any error message and does not set and governors etc. 

gah... very sorry about that.
I hope I'll be able to upload a new package in the next couple of days.

cheers
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#457753: loadcpufreq does nothing by default, cpufrequtils unusable

[Mattia Dongili]

On Tue, Dec 25, 2007 at 12:53:11PM +0100, Kai Weber wrote:
 Package: cpufrequtils
 Version: 002-6
 Severity: grave
 Justification: renders package unusable
 
 
 The script init.d/loadcpufreq is disabled by default, so no modules for
 cpu frequency scaling are loaded. This makes init.d/cpufrequtils fail
 as well.
 
 If I read REDAME.Debian correctly this is not the desired behaviour.
 loadcpufreq should be enabled by default and if someone wishes to
 override this he should create a file /etc/default/loadcpufreq like
 the provided example.
 
 I guess in the loadcpufreq script a default ENABLED=true is missing.

damn, you're right...
thanks for spotting it so quickly, will upload an new package asap

cheers
-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#446729: [Pkg-uml-pkgs] Bug#446729: rootstrap: MAKEDEV: No such file or directory

[Mattia Dongili]

On Mon, Oct 22, 2007 at 04:03:11PM +0200, Marcus Better wrote:
 package rootstrap
 tag 446729 patch
 thanks
 
 This patch appears to fix it.
 
 --- modules/uml~2006-09-26 10:57:22.0 +0200
 +++ modules/uml 2007-10-22 10:18:43.0 +0200
 @@ -5,7 +5,7 @@
  for i in 1 2 3 4 5 6 7 8; do
 NEW_TTY=$NEW_TTY tty$i
  done
 -chroot $TARGET /bin/sh -c cd /dev  ./MAKEDEV ubd std pty $NEW_TTY
 +chroot $TARGET /bin/sh -c cd /dev  /sbin/MAKEDEV ubd std pty $NEW_TTY
  chroot $TARGET umount /proc

thanks, I hope I'll heve some time to update the package soon

-- 
mattia
:wq!



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#443726: xserver-xorg-input-evtouch: input coordinates not scaled

[Mattia Dongili]

Package: xserver-xorg-input-evtouch
Version: 0.8.7-2
Severity: grave
Justification: renders package unusable


the driver's conversion_proc is not called in xorg-1.4
xf86PortMotionEvent making the driver fail to scale the input
coordinates to the screen size.
The net result is the pointer not responding correctly to taps.

I'm already working on a fix, but this is to avoid that the package
enters lenny if I fail to fix it in a timely fashion.

mattia

-- System Information:
Debian Release: lenny/sid
  APT prefers testing
  APT policy: (990, 'testing'), (300, 'unstable')
Architecture: i386 (i686)

Kernel: Linux 2.6.23-rc1-mm1-1 (SMP w/2 CPU cores; PREEMPT)
Locale: LANG=en_GB, LC_CTYPE=ja_JP (charmap=EUC-JP)
Shell: /bin/sh linked to /bin/bash



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#398091: php[45]-xcache for PHP 5.2

[Mattia Dongili]

Hello Risko,

since I needed to have a working xcache for etch I updated you package
to upstream version 1.2.0.
You can find my work here:
http://people.debian.org/~malattia/packages/xcache/
(It was built in a clean etch chroot)

Something that the changelog doesn't list is the fact that I added an
xcache.ini to be installed in /etc/php[45]/conf.d, it installs with the
cache actually disabled but loading the extension.

Thanks
-- 
mattia
:wq!


signature.asc
Description: Digital signature

Bug#407708: optimization problem

[Mattia Dongili]

severity 407708 important
retitle 407708 iasl optimization generates a broken DSDT
thanks

On Sat, Jan 20, 2007 at 07:38:33PM +0100, emisca wrote:
 I was wrong. I tried older version of iasl, and all have this behaviour.

I discovered the same with a more recent one (20061109).

 I tried disabling optimizations (iasl -oa -tc), and the code generated
 seems good. Perhaps there is something wrong with optimization logic.

I see, I'll try to ping upstream about that. By the way, I assume
booting with the mis-optimized DSDT has't been successful, correct?

Thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#401153: NMU uploaded

[Mattia Dongili]

On Tue, December 19, 2006 1:09 am, Andreas Barth said:
 * Andreas Barth ([EMAIL PROTECTED]) [061214 08:41]:
 I uploaded an NMU of your package to experimental to see if the test
 case really works on buildds.

 It failed as expected on powerpc and also on sparc, and worked on at
 least amd64 and powerpc - so I think this package can be uploaded to
 unstable, unless there is reason not to do it.

 In other words, I intend to upload it to unstable during the next days.

please go on, I won't have time enough in the next days.
Thanks
-- 
mattia
:wq!

Bug#401153: backtrace for iasl bug #401153

[Mattia Dongili]

On Mon, Dec 04, 2006 at 04:38:26PM +0100, Andreas Henriksson wrote:
 On Sun, Dec 03, 2006 at 08:58:31PM +0200, Guillem Jover wrote:
 [...]
  The attached patch fixes the segfault, and corrects the CFLAGS
  handling for upstream and Debian, it also adds alpha to the list of
  64 bit arches. The fix for compiler/aslopcodes.c is needed because
  the macros ACPI_UINT32_MAX: and ACPI_INTEGER_MAX are the same if
  ACPI_INTEGER is defined to be 32 bits.
  
  regards,
  guillem
 
 I can confirm that the patch provided by Guillem Jover seems to fix
 the testcase in the original bug-report for me on Debian Testing
 PowerPC.
 
 $ iasl -tc ../acpi-dsdt.dsl
 
 Intel ACPI Component Architecture
 ASL Optimizing Compiler version 20060912 [Dec  4 2006]
 Copyright (C) 2000 - 2006 Intel Corporation
 Supports ACPI Specification Revision 3.0a
 
 ASL Input:  ../acpi-dsdt.dsl - 561 lines, 18338 bytes, 209 keywords
 AML Output: acpi-dsdt.aml - 2098 bytes 104 named objects 105 executable
 opcodes
 
 Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 68 Optimizations
 $

Ok, I'll see if it doesn't introduce evident regressions (eg by
compiling some of the DSDT's available on acpi.sf.net + my own here) and
eventually include it in the package.

Robert: does the patch looks correct to you?

Thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#401153: backtrace for iasl bug #401153

[Mattia Dongili]

On Mon, Dec 04, 2006 at 07:53:49PM +0100, Andreas Barth wrote:
 * Andreas Henriksson ([EMAIL PROTECTED]) [061204 19:44]:
  I can confirm that the patch provided by Guillem Jover seems to fix
  the testcase in the original bug-report for me on Debian Testing
  PowerPC.
 
 I would propose to upload the fixed package ASAP. In case you want, I
 can do an NMU as well.

No problem, I'm already working on it.
Thanks.
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#401153: backtrace for iasl bug #401153

[Mattia Dongili]

On Sat, Dec 02, 2006 at 02:27:40AM +0100, Andreas Henriksson wrote:
 On fre, 2006-12-01 at 15:43 -0800, Moore, Robert wrote:
  I'm not sure I understand what you are describing, please clarify.
 
 Please see http://bugs.debian.org/401153 for the original bug report.
 Basically, iasl successfully compiles the atteched dsl file on a regular
 i386 pc, but when running the same program on PowerPC (Debian GNU/Linux)
 then you get a segmentation fault. There is probably a bug in the
 program somewhere which seems to only trigger on powerpc at the moment.
 
 It has been tracked down to being related to Asl.Value.String being
 NULL. This state seems invalid and makes the program crash. Do you have
 any idea what might have failed and how this variable could be
 uninitialized? It would be really appreciated if you have any hints
 about possible failures that we can investigate...

Just as a side note, I tried compiling the mentioned DSDT on an ARM
platform and I get a different error:

$ ./acpica-unix-20060912/compiler/iasl acpi-dsdt.dsl

Intel ACPI Component Architecture
ASL Optimizing Compiler version 20060912 [Oct  2 2006]
Copyright (C) 2000 - 2006 Intel Corporation
Supports ACPI Specification Revision 3.0a

ACPI Warning (nssearch-0421): Found bad character(s) in name, repaired: [_GP*]
 [20060912]

And no AML is generated.
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#401153: backtrace for iasl bug #401153

[Mattia Dongili]

On Fri, Dec 01, 2006 at 04:45:37PM +0100, Andreas Henriksson wrote:
 On Fri, Dec 01, 2006 at 03:42:51PM +0100, Andreas Henriksson wrote:
  Program received signal SIGSEGV, Segmentation fault.
  0x10019570 in TrAmlTransformWalk ()
  (gdb) bt
  #0  0x10019570 in TrAmlTransformWalk ()
  #1  0x1001758c in TrWalkParseTree ()
  #2  0x1000b728 in CmDoCompile ()
  #3  0x10011250 in main ()
  (gdb)
  
 
 I forgot to mention, this is on Debian Testing PowerPC with iasl
 recompiled with DEB_BUILD_OPTIONS=noopt, nostrip.
 
 
 Some additional printf debugging shows this:
 TrAmlTransformWalk - TrTransformSubtree - TrDoDefinitionBlock
 
 File compiler/asltransform.c line 432:
   if (!ACPI_COMPARE_NAME (Next-Asl.Value.String, ACPI_SIG_DSDT))
 
 it blows up here Next-Asl.Value.String is NULL... All the
 macro/pointer obfuscation got me lost here... 

Thanks, I'm forwading the thing upstream

-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#401153: backtrace for iasl bug #401153

[Mattia Dongili]

Hi Robert,

I've been submitted a bug from the powerpc Debian's folks (see below or
point your browser at [1]).
In short the story is: we enabled building iasl for powerpc in order to
buld DSDT tables for qemu which is available on powerpc and can emulate
the x86.
So we have a DSDT which makes iasl segfault while building on powerpc,
it doesn't on i386. The DSDT is here:
http://bugs.debian.org/cgi-bin/bugreport.cgi/acpi-dsdt.dsl?bug=401153;msg=5;att=1

Can you help putting some light where the problem may hide?

Thanks

[1]: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=401153

- Forwarded message from Andreas Henriksson [EMAIL PROTECTED] -

Subject: Bug#401153: backtrace for iasl bug #401153
From: Andreas Henriksson [EMAIL PROTECTED]
Date: Fri, 1 Dec 2006 16:45:37 +0100
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Message-ID: [EMAIL PROTECTED]
User-Agent: Mutt/1.5.9i
Reply-To: Andreas Henriksson [EMAIL PROTECTED], [EMAIL PROTECTED]

On Fri, Dec 01, 2006 at 03:42:51PM +0100, Andreas Henriksson wrote:
 Program received signal SIGSEGV, Segmentation fault.
 0x10019570 in TrAmlTransformWalk ()
 (gdb) bt
 #0  0x10019570 in TrAmlTransformWalk ()
 #1  0x1001758c in TrWalkParseTree ()
 #2  0x1000b728 in CmDoCompile ()
 #3  0x10011250 in main ()
 (gdb)
 

I forgot to mention, this is on Debian Testing PowerPC with iasl
recompiled with DEB_BUILD_OPTIONS=noopt, nostrip.


Some additional printf debugging shows this:
TrAmlTransformWalk - TrTransformSubtree - TrDoDefinitionBlock

File compiler/asltransform.c line 432:
if (!ACPI_COMPARE_NAME (Next-Asl.Value.String, ACPI_SIG_DSDT))

it blows up here Next-Asl.Value.String is NULL... All the
macro/pointer obfuscation got me lost here... 

HTH

-- 
Regards,
Andreas Henriksson



- End forwarded message -
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#400580: cpufreqd: fails to change governor when goes with battery

[Mattia Dongili]

On Tue, November 28, 2006 7:25 pm, leandro noferini said:
 Mattia Dongili [EMAIL PROTECTED] writes:


 [...]

 the kernel. could you please lower the severity and reassign?

 Ok!

 How could I do that?

first ensure that the bug is reproducible with debian's stock kernels,
then reassign, read here: http://www.debian.org/Bugs/server-refcard.html

anyway if you need help for the second step, just ask :)
-- 
mattia
:wq!

Bug#400580: cpufreqd: fails to change governor when goes with battery

[Mattia Dongili]

On Tue, November 28, 2006 1:44 pm, leandro noferini said:
 Mattia Dongili [EMAIL PROTECTED] writes:

 [...]

 kernel bug? cpufreqd correctly complains about not being able to switch
 to
 ondemand. If this governor doesn't work for your platform, just
 reconfigure cpufreqd to avoid using it.

 Now I read more carefully.

 My platform is a last generation  iBook G4 and cpufreqd worked fine also
 with the  kernel 2.6.17  home compiled. cpufreqd  stops working  from an
 upgrade on.

ok, and your logs say that _the_kernel_ is unable to load the ondemand
governor, hence cpufreqd fails. So your problem is to be searched
elsewhere.

-- 
mattia
:wq!

Bug#400580: cpufreqd: fails to change governor when goes with battery

[Mattia Dongili]

On Tue, November 28, 2006 6:20 pm, leandro noferini said:
 Mattia Dongili [EMAIL PROTECTED] writes:

 [...]

 Now I read more carefully.

 My platform is a last generation  iBook G4 and cpufreqd worked fine
 also
 with the  kernel 2.6.17  home compiled. cpufreqd  stops working  from
 an
 upgrade on.

 ok, and your logs say that _the_kernel_ is unable to load the ondemand
 governor, hence cpufreqd fails. So your problem is to be searched
 elsewhere.

 Any tip from where to begin?

the kernel. could you please lower the severity and reassign?

thanks
-- 
mattia
:wq!

Bug#400580: cpufreqd: fails to change governor when goes with battery

[Mattia Dongili]

On Mon, November 27, 2006 12:33 pm, Debian BTS said:
 This is the log from syslog
...
 Package: cpufreqd
 Version: 2.2.0-2
 Severity: grave
 Justification: renders package unusable

 Nov 27 12:19:35 janni kernel: ondemand governor failed to load due to too
 long transition latency
 Nov 27 12:19:37 janni cpufreqd: cpufreqd_set_profile : Couldn't set
 profile On Demand High set for cpu0
 Nov 27 12:19:37 janni cpufreqd: cpufreqd_loop: Cannot set
 policy, Rule unchanged (AC Rule).
 Nov 27 12:19:37 janni kernel: ondemand governor failed to load due to too
 long transition latency

 The kernel is a 2.6.18 home compiled but this happens also with a 2.6.17

kernel bug? cpufreqd correctly complains about not being able to switch to
ondemand. If this governor doesn't work for your platform, just
reconfigure cpufreqd to avoid using it.

-- 
mattia
:wq!

Bug#397917: gsynaptics - lists s390 as supported

[Mattia Dongili]

On Sat, Nov 11, 2006 at 02:11:24PM +0900, Osamu Aoki wrote:
 Hi,
 
 On Fri, Nov 10, 2006 at 05:59:07PM -0800, Steve Langasek wrote:
 ...
   Option 3)  
* Set Architecture: alpha amd64 arm hppa i386 ia64 m68k mips mipsel 
   powerpc sparc
* Ask ftp-master for removal of package for s390
* Make entry in quinn-diff not to build on s390
  
  This is an acceptable solution,
 ...
   Option 4)
* Set Architecture: any
* Have fancy script to fail in s390.
* Ask ftp-master for removal of package for s390
* Make entry in quinn-diff not to build on s390
  
  This is an acceptable solution.
 
 I like the latter so I do not interfare with builoding kfreebsd-i386 etc.

Well, the synaptics driver is quite Linux specific, isn't it?
And since xserver-xorg-driver-synaptics is also not built on that
architecture the gsynaptics package will be uninstallable anyway.

-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#397917: gsynaptics - lists s390 as supported

[Mattia Dongili]

On Fri, Nov 10, 2006 at 11:17:57PM +0900, Osamu Aoki wrote:
 On Fri, Nov 10, 2006 at 02:00:32PM +0100, Bastian Blank wrote:
  Package: gsynaptics
  Version: 0.9.7-1
  Severity: serious
  
  gsynaptics lists s390 as supported (Architecture: any) but
  xserver-xorg-input-synaptics is not available. This leads to
  uninstallable binary packages.
 
 I see.  But xserver-xorg-input-synaptics has 
 
 Architecture: alpha amd64 arm hppa i386 ia64 m68k mips mipsel powerpc sparc

it was Architecture: any with a quinn-diff[1] entry in the past, and
s390 was not supported because xserver-xfree86 wasn't available there.

 This looks agry.  Because I need to track new architectures.  
 
 Can I do
 
 Architecture: !s390
 
 Anyone have suggestion?

This is ok until you have to add one more unsupported arch to the list.
I think the trade-off here is if you want to add new supported
or unsupported architectures, the latter implying you may need to ask
for removal of the binary (as in the current case for s390), the former
implying a new upload.

Here the discussion I had some time ago:
http://www.mail-archive.com/debian-mentors@lists.debian.org/msg12969.html

Currently (AFAICT) xserver-xorg-input-synaptics seems to be installable
on s390 but maybe it doesn't make much sense...

[1]: http://buildd.debian.org/quinn-diff/Packages-arch-specific

-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#375851: possibly just the wrong encoding?

[Mattia Dongili]

Hi Enrico,

is this the same problem addressed in
/usr/share/doc/festvox-itapc16k/README.Debian?
(Encoding of accented letters)

I did write a stupid program (using glade and gtk2) to play with
festival[1], see launch_festival(void *x) in src/callbacks.c for some
code to do the same as the suggested `recode stuff`. It needs
librecode-dev to build.

[1]: http://people.debian.org/~malattia/packages/parlami-0.1.tar.gz
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#374582: [Pkg-uml-pkgs] Bug#374582: base-config is been invoked in rootstrap, although it doesn't exist anymore

[Mattia Dongili]

severity 374582 minor
thanks

On Tue, Jun 20, 2006 at 08:10:18AM +0300, alex bodnaru wrote:
 Package: rootstrap
 Version: 0.3.22-1
 Severity: grave
 Justification: renders package unusable
 
 base-config has been removed from etch, hence it's invokation fails.
 you may execute it on condition if exists, or just ignore it's failure 
 like with || true.

Uh, but the base-config module is not even included in the default
configuration.
I wouldn't even consider this a bug, at least until sarge (with
base-config) disappears.

I'm very tempted to close this bug, I don't really see any bug here.
Surely this is not grave at all.
Maybe just documenting this issue can help?

thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#372498: News about this bug? (ksynaptics: incompatible with xfree86-driver-synaptics 0.14.5-1)

[Mattia Dongili]

Hello Arnaud,

On Wed, Jun 14, 2006 at 09:22:05PM +0200, Arnaud Quette wrote:
[...]
 A last thanks to Modestas for putting me in touch with Fathi, and to
 Mattia for his work on the synaptics driver.

sorry for not getting back to you earlier, I've been on a 2day business
trip and anyway I didn't hear anything from the guy I asked help for
[kq]synaptics.

BYe
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#372259: user-mode-linux: the uml_switch control file is not been searched as in uml-utilities package

[Mattia Dongili]

On Sun, Jun 11, 2006 at 09:35:51AM +0200, Stefano Melchior wrote:
 On Sun, Jun 11, 2006 at 07:19:46AM +0300, alex bodnaru wrote:
[...]
  fortunately, this has changed in um2 version, and now upgrading is smoother.
  
 thus does this mean that -2um version fixed the bug you submitted? If
 so, if you feel it should be fixed, please let me know it, because I would 
 like to close the bug, please.

it already is.

thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#370155: cpufreqd: init script does not stop the daemon and hangs until user input

[Mattia Dongili]

On Sat, Jun 03, 2006 at 08:48:36PM +0200, Norbert Preining wrote:
 Package: cpufreqd
 Version: 2.0.0-1
 Severity: serious
 
 Hi!
 
 My cpufreqd init script hangs indefinitely when stopping. I have to press 
 the enter key to make it continue, also that doesn't stop the daemon.
 
 I consider this serious as remote reboot is not possible anymore.
 And the shutdown process has to be attended.

I can't reproduce it here, can you run the script with -x and send the
output to see where it hangs?

Thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#370155: cpufreqd: init script does not stop the daemon and hangs until user input

[Mattia Dongili]

reassign 370155 lsb-base 3.1-6
retitle 370155 pidofproc reads from stdin and hangs waiting for user input
severity 370155 critical
thanks

On Sat, Jun 03, 2006 at 10:28:11PM +0200, Norbert Preining wrote:
 Hi Mattia!
 
 On Sam, 03 Jun 2006, Mattia Dongili wrote:
  I can't reproduce it here, can you run the script with -x and send the
  output to see where it hangs?
 
 It hangs in the read pid:

Yes, upgraded my lsb-base (was at -5 here) and it's fully reproducible.

 gandalf:~# bash -x /etc/init.d/cpufreqd stop
 + PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin
[...]
 + pidfile=/var/run/cpufreqd.pid
 + '[' -f /var/run/cpufreqd.pid ']'
 + read pid
 
 ENTER

I suppose it should be
cat /var/run/cpufreqd.pid | read pid
instead.

It seems to be an lsb-base bug, reassigning. Severity is critical as it
breaks unrelated software.

Thanks for reporting.
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#370155: cpufreqd: init script does not stop the daemon and hangs until user input

[Mattia Dongili]

On Sat, Jun 03, 2006 at 11:04:55PM +0200, Norbert Preining wrote:
 Hi all!
 
 On Sam, 03 Jun 2006, Mattia Dongili wrote:
  I suppose it should be
  cat /var/run/cpufreqd.pid | read pid
  instead.
 
 changed it to this, and it didn't work, at least it didn't stop the
 cpufreqd.

oh... running it on the command line[1] seems to work, but you're right, the
script still is not stopping cpufreqd (pidofproc returns 2).

[1]: if [ -f /var/run/cpufreqd.pid ] ; then cat /var/run/cpufreqd.pid |
read pid ; fi ; if [ -n ${pid:-} ]; then echo $pid ; fi

 So let the lsb-base maintainers fix this.

I have to agree :)

-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#365395: user-mode-linux: UML will not boot from ISO

[Mattia Dongili]

severity 365395 important
thank

On Sat, Apr 29, 2006 at 11:48:16AM -0700, Todd A. Jacobs wrote:
 Package: user-mode-linux
 Version: 2.6.16-1um-1
 Severity: grave
 Justification: renders package unusable

it's not really unusable, severity is inappropriate IMO.

 I reported on the UML mailing list that I was unable to boot from an ISO
 image, even after various gyrations. Jeff Dike suggested that the binary
 was probably compiled without proper block support (specifically
 CONFIG_BLK_DEV_UBD, although there may be issues with ISO9660 support as
 well).
 
 Here is the actual error and command-line:
 
 $ linux mem=64M umid=debian ubd0r=debian-testing-i386-netinst.iso

You're mixing things up. You can't boot from the install image, you need
to extract the initrd from the iso [1] and boot from the initrd
[2]. Then you can proceed and install the system (you'll also have to
manually tell the installer where the cdrom is -you'll be asked during
installation, in my example you can use /dev/ubdb).

[1]: mount -o loop debian-testing-i386-netinst.iso /mnt
 cp /mnt/install/2.6/initrd.gz /anywhere
[2]:  linux ramdisk_size=12000 mem=128M umid=debian initrd=initrd.gz 
root=/dev/ram0 ubd0=rootfs ubd1=debian-testing-i386-netinst.iso 
eth0=tuntap,,,172.20.0.1

As said in the uml-devel ML I have the new package ready here, I'll
upload it tomorrow.

[...]
 Console initialized on /dev/tty0
 Initializing software serial port version 1
  ubda: unknown partition table
 Kernel panic - not syncing: VFS: Unable to mount root fs on 
 unknown-block(98,0)

In fact you get this same error running:
linux mem=64M umid=debian ubd0r=/dev/zero
The problem here is that netinst.iso doesn't contain a regular
filesystem to boot with. This works with regular PCs because the
bootloader helps this process (loading and running the initrd).

Thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#362058: [Pkg-uml-pkgs] Bug#362058: uml-utilities: ships port-helper under /usr/lib64 on amd64

[Mattia Dongili]

tags 362058 +confirmed +pending
thanks

On Tue, Apr 11, 2006 at 07:57:40PM -0400, Aaron M. Ucko wrote:
[...]
 uml-utilities goes out of its way to force use of /usr/lib64 on amd64,
 which is a big no-no on Debian and clashes with libc6:

Ops, sorry.

  Unpacking replacement libc6 ...
  dpkg: error processing /var/cache/apt/archives/libc6_2.3.6-6_amd64.deb 
  (--unpack):
   trying to overwrite `/usr/lib64', which is also in package uml-utilities
 
 Could you please fix this?

done, fixed in our svn repository. I'll upload -3 tomorrow.

bye
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#345436: xlibs: Keyboard not working properly after upgrade to 6.9.0.dfsg.1-1

[Mattia Dongili]

On Sat, Dec 31, 2005 at 12:26:39PM -0200, Paulo Marcel Coelho Aragao wrote:
 Package: xlibs
 Version: 6.9.0.dfsg.1-1
 Severity: grave
 Justification: renders package unusable
 
 After the upgrade to 6.9.0.dfsg.1-1, the keyboard stopped working properly 
 under X. I've noticed the following:
 
 1. ctrl-alt-fn doesn't switch to other virtual consoles
 2. The key with characters '/' and '?' is dead
 
 Not switching to virtual consoles can be lived with, but the lack of '/' and 
 '?' makes X unusable.
 
 Needless to say, the keyboard works properly on the consoles, and it worked 
 properly before the upgrade.
[...]
 Versions of packages xlibs depends on:
[...]
 ii  libxtst6 6.9.0.dfsg.1-1  X Window System event recording 
 an
 ii  xlibs-data   6.8.2.dfsg.1-11 X Window System client data
   ^
maybe you only missed the xlibs-data update?

-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#337479: yaird: should use /usr/bin/perl, not /usr/local/bin/perl

[Mattia Dongili]

On Wed, Nov 09, 2005 at 08:16:08PM +0100, Laurent Bonnaud wrote:
 On sam, 2005-11-05 at 16:31 +0100, Mattia Dongili wrote:
[...]
  I's like
  configuring a package with silly ./configure options and pretending it
  works everywhere. :)
 
 I did not intervene in the build process.  I used this command:
 
   apt-get -b source yaird
 
 With this command I expect to get a working package as long as I do not
 mess up the system.  And putting files in /usr/local/ is clearly allowed
 and does not count as messing up the system.

Yes, sorry. I was thinking of a lot of silly corner cases when I wrote
that (I hit the same issue when playing with ./configure and stuff so
that's what I had in mind in the first place).

By the way, Erik, Jonas I'm attaching a patch to the configure.in to add
--with-perl option (defaults to autodetection if not given).
I'm sorry I'm not into cdbs enough to provide a patch for debian/rules
also :)

-- 
mattia
:wq!
Index: configure.in
===
--- configure.in(revision 4760)
+++ configure.in(working copy)
@@ -47,8 +47,29 @@
[Define to nuke initramfs before moving to real root])
 fi
 
+# Force a specific perl binary
+AC_ARG_WITH(perl,
+   AC_HELP_STRING([--with-perl=PATH],
+   [Force a specific perl executable path (autodetected by 
default)]),
+   [
+   case ${withval} in
+   yes|auto)
+   with_perl=auto ;;
+   *)  PERL=${withval}
+   if ! test -x ${PERL} ; then
+   AC_MSG_ERROR(bad value ${PERL} for --with-perl) 
+   fi
+   AC_SUBST(PERL)
+   ;;
+   esac
+   ],[
+   with_perl=auto
+   ])
+if test $with_perl = auto; then
+   AC_PATH_PROG(PERL,perl)
+fi
+
 # Checks for programs.
-AC_PATH_PROG(PERL,perl)
 AC_PROG_CC
 AC_PROG_RANLIB
 

Bug#336598: cannot parse /etc/fstab file with SMB share

[Mattia Dongili]

merge 336585 336598
thanks

On Mon, Oct 31, 2005 at 12:45:25PM +, Martin Michlmayr wrote:
 Package: yaird
 Version: 0.0.11-10
 Severity: grave
 
 yaird fails when /etc/fstab contains a SMB share, such as:
 
 \\mlpc-serv-fs1.eng.cam.ac.uk\homes   /srv/mlpc-serv-fs1 smbfs 
 noauto,user,uid=tbm,gid=tbm,credentials=/etc/samba/private/msm34

this is actually the same as #336585:
The fs_freq and fs_passno fields in /etc/fstab are optional

-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#336598: cannot parse /etc/fstab file with SMB share

[Mattia Dongili]

On Mon, Oct 31, 2005 at 06:19:56PM +0100, Mattia Dongili wrote:
 merge 336585 336598
 thanks

doh! failed because of severity. I'll let kernel mainteiners decide on
that then.

sorry for the noise
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#336598: cannot parse /etc/fstab file with SMB share

[Mattia Dongili]

On Mon, Oct 31, 2005 at 08:09:35PM +0100, Jonas Smedegaard wrote:
 -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1
 
 On Mon, 31 Oct 2005 18:19:56 +0100
 Mattia Dongili [EMAIL PROTECTED] wrote:
 
  this is actually the same as #336585:
  The fs_freq and fs_passno fields in /etc/fstab are optional
 
 Yes. Thanks, Mattia.
 
 The attached patch should fix this issue. Looks like a simple oversight
 from upstream (he noticed they are irrelevant but checks for them
 anyway).

I'm sorry, it seems it take a couple of hours to my messages to be
delivered.
I sent another mail telling that the correct fix is probably checking
if ($#fields  3) as also the fs options are (ehrm...) optional :)

BTW: the mere failed because of mismatching severity
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]

Bug#336598: cannot parse /etc/fstab file with SMB share

[Mattia Dongili]

On Mon, Oct 31, 2005 at 09:06:59PM +0100, Jonas Smedegaard wrote:
[...]
 On Mon, 31 Oct 2005 20:26:35 +0100
 Mattia Dongili [EMAIL PROTECTED] wrote:
 
  I sent another mail telling that the correct fix is probably checking
  if ($#fields  3) as also the fs options are (ehrm...) optional :)
 
 I disagree. Judging from the manpage of fstab the 4th entry contains
 at least the type of mount, so cannot be left out.

anyway mount allows the presence of such lines:
sysfs /sys sysfs
/dev/hda5 /mnt/tmp0 reiser4

the above 2 lines both work here without any complaint from mount (I did
notice I left out opts-dump-pass entried only trying out yaird :))

-- 
mattia
:wq!


signature.asc
Description: Digital signature

Bug#319387: problems starting kde 3.4

[Mattia Dongili]

tags 319387 + moreinfo
severity 319387 important
thanks

On Thu, Jul 21, 2005 at 08:19:18PM +0200, Bastian Venthur wrote:
 Package: xfree86-driver-synaptics
 Version: 0.14.2-2
 Severity: serious
 
 When booting and starting in kdm, the touchpad works fine, but when kde
 3.4 (from alioth) starts after login, the touchpad is disabled again.
 
 The only way to get it back to work is:
 
 sudo modprobe -r psmouse  sudo modprobe psmouse
 
 in kde, and the touchpad is on again.

Could you provide X (/var/log/Xorg.0.log) and kernel (from dmesg) logs? 

Thanks
-- 
mattia
:wq!


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]