Re: access rights in /sbin and /bin [Update]

2008-12-09 Thread Thomas Preud'homme
The Wednesday 10 December 2008 07:04:50 [EMAIL PROTECTED], you wrote :
 On Mon, Dec 08, 2008 at 10:33:02AM -0500, Lennart Sorensen wrote:
  On Sun, Dec 07, 2008 at 04:11:04PM +0100, Hans-J. Ullrich wrote:
   thanks for the list. I checked and found out, that a lot of binaries in
   /sbin got permissions to rwxr-xr-- (root:root), but they should have
   rwxrwxr-x. I wondered, as I never changed the rights manually in the
   past and I am sure, I have not been hacked. So there is only one
   explanation: an applicatiopn must have changed it. Does someone know,
   which application is changing rights of binaries below /sbin ? I
   suppose, it is either bastille (which I installed and deinstalled a
   long time ago) or selinux (which i still installed).
  
   Please, which manual did i miss to read ???
 
  So far the only thing I have ever seen that causes that is silly people
  who mess with the umask of the root user (which causes dpkg to make lots
  of mistakes).

 Perhaps dpkg shouldn't rely on the umask of the root user?  Perhaps is
 should set it itself?  Could this be considered a dpkg bug?

It sounds reasonable indeed that dpkg don't rely on root umask. I don't want 
root to have a umask of 022 because usually I don't want users to read root 
file by default. Even if most of the time it's not a security issue, I don't 
want these file to be readable by users by default in case I forget to 
restrict rights of sensitive files.

Furthermore, AFAIK files in /bin, /sbin and other bin directories aren't 
created, they are untared so that rights of these files are rights they have 
when tared by the debian maintener of the package.


 -- hendrik

  So if you ever set a umask for your root user, well don't and reinstall
  every affected package to fix the permissions.
 
  --
  Len Sorensen
 
 
  --
  To UNSUBSCRIBE, email to [EMAIL PROTECTED]
  with a subject of unsubscribe. Trouble? Contact
  [EMAIL PROTECTED]


Greetings,

Thomas Preud'homme

-- 
Why debian : http://www.debian.org/intro/why_debian


signature.asc
Description: This is a digitally signed message part.


Re: access rights in /sbin and /bin [Update]

2008-12-09 Thread Thomas Preud'homme
The Tuesday 09 December 2008 14:38:40 Hans-J. Ullrich, you wrote :
 Am Dienstag, 9. Dezember 2008 schrieb Thomas Preud'homme:
  The Wednesday 10 December 2008 07:04:50 [EMAIL PROTECTED], you 
wrote :
   On Mon, Dec 08, 2008 at 10:33:02AM -0500, Lennart Sorensen wrote:
On Sun, Dec 07, 2008 at 04:11:04PM +0100, Hans-J. Ullrich wrote:
 thanks for the list. I checked and found out, that a lot of
 binaries in /sbin got permissions to rwxr-xr-- (root:root), but
 they should have rwxrwxr-x. I wondered, as I never changed the
 rights manually in the past and I am sure, I have not been hacked.
 So there is only one explanation: an applicatiopn must have changed
 it. Does someone know, which application is changing rights of
 binaries below /sbin ? I suppose, it is either bastille (which I
 installed and deinstalled a long time ago) or selinux (which i
 still installed).

 Please, which manual did i miss to read ???
   
So far the only thing I have ever seen that causes that is silly
people who mess with the umask of the root user (which causes dpkg to
make lots of mistakes).
  
   Perhaps dpkg shouldn't rely on the umask of the root user?  Perhaps is
   should set it itself?  Could this be considered a dpkg bug?
 
  It sounds reasonable indeed that dpkg don't rely on root umask. I don't
  want root to have a umask of 022 because usually I don't want users to
  read root file by default. Even if most of the time it's not a security
  issue, I don't want these file to be readable by users by default in case
  I forget to restrict rights of sensitive files.
 
  Furthermore, AFAIK files in /bin, /sbin and other bin directories aren't
  created, they are untared so that rights of these files are rights they
  have when tared by the debian maintener of the package.

 Hmm, so if this is really the way, packages are installed, then I should
 always report to the maintainer of the package. On the other hand, I had
 some problems with the package eject. The installed binary got rwxr--r--
 ( with owners root:root), but in the package itself it got rwxr-xr-x
 (root:root).

I'm almost sure of what I said. For example I installed nicotine recently and 
my root umask is 027 but /usr/bin/nicotine is 755 so the umask is not used. I 
think umask is used in package installation only for files created by package 
scripts (postrm, prerm, etc.)


 When I reinstalled it, the permissions did not change. Then I deinstalled
 it completely and reinstalled it again. Now it got the correct permissions:
 rwxr-xr-x ! So far, so well. But after an upgrade some weeks later, the
 permissions were wrong again (set as before). I could not explain that to
 myself, but I could verify this behaviour. Lots of other binaries were
 showing the same effects (for example the kde screensaver suddenly could
 not have been unlocked any more, as the sticky bit was missing).

 So, if I would understand, how is and who is setting the permissions at
 upgrades or installing, I could find out, what is going wrong.

 Who knows it ?

As you said before you must have a program that change these permissions 
later. It must have a program to monitor change on a file, like a permanent 
lsof which would say you which program did the modifications.



 Cheers

 Hans

Regards,

Thomas Preud'homme

-- 
Why debian : http://www.debian.org/intro/why_debian


signature.asc
Description: This is a digitally signed message part.


Re: What is the best java for debian-amd64 ?

2008-09-16 Thread Thomas Preud'homme
Tuesday 16 September 2008, Hans-J. Ullrich wrote :
 Hi all,

 as there are many java versions installable, what is the best one to
 use in amd64-systems ? What are the differences ? I saw gcj, gij,
 gdk, java. Can someone tell me, which should be used ? Ia ma very
 confused. Update-alternatives is giving me these choices:

 There are 8 alternatives which provide `java'.

   SelectionAlternative
 ---
   1/usr/bin/gij-4.1
   2/usr/lib/jvm/ia32-java-6-sun/jre/bin/java
   3/usr/lib/jvm/java-gcj/jre/bin/java
   4/usr/bin/gij-4.3
 * 5/usr/lib/jvm/java-6-sun/jre/bin/java
   6/usr/bin/gij-4.2
   7/usr/lib/jvm/ia32-java-1.5.0-sun/jre/bin/java
  +8/usr/lib/jvm/java-6-openjdk/jre/bin/java

 Press enter to keep the default[*], or type selection number:

 I choose 5. What is this +-sign meaning ?

It was your current alternative when you ran update-alternatives


 Questions, questions, questions

 Thanks for any help !

I think the two most important ones are the sun alternative because it 
is widely used and the openjdk which is the free sun JVM (v6 I think) + 
the free java libraries. It should replace the comming java 7 (where 
jvm + libraries is supposed to be free) in all distros I think.

Not that I'm not completely sure of what I say.



 Kind regards

 Hans-J. Ullrich



Regards,

Thomas Preud'homme

-- 
Why Debian : http://www.debian.org/intro/why_debian


signature.asc
Description: This is a digitally signed message part.


Re: What are these packages doing ?

2008-08-08 Thread Thomas Preud'homme
Saturday 09 August 2008, Sandro Tosi wrote :
  What are these packages good for ?
 
  linux-image-amd64
  linux-image-2.6-amd64
 
  Are these meta-packages ? Or are these packages intend to let
  aptitude or apt-get automatically install the newest
  kernel-versions ? (If yes, then these are the packages I am looking
  for)

 for sure linux-image-2.6-amd64 lets you have always the latest 2.6
 kernel for amd64 installed. apt-cache show it and check Depends line:
 it should have (if your machine is enough up-to-date) the version
 2.6.25-2.

And linux-image-amd64 as a dependance on the latest linux branch (here 
linux-image-2.6.amd64). If one day Linux jump to 2.8 version, then 
linux-image-amd64 dependance will be modified accordingly.

 Kindly,
 --
 Sandro Tosi (aka morph, Morpheus, matrixhasu)
 My website: http://matrixhasu.altervista.org/
 Me at Debian: http://wiki.debian.org/SandroTosi



Regards,

Thomas Preud'homme

-- 
Why Debian : http://www.debian.org/intro/why_debian


signature.asc
Description: This is a digitally signed message part.


Re: Heavy bug in kde !

2008-08-06 Thread Thomas Preud'homme
Wednesday 06 August 2008, Hans wrote :
 Dear maintainers,

 I discovered a heavy bug in kde.

 As I do not know, which part is responsible for it, I hope, someone
 other know and will send a bugreport for the package (or tell it to
 me, so that I can report it).

 ---
 Bug description:

 When you want to save your settings of kde (I mean your personal
 settings, the one, you save with the button save settings in the
 K-Menu), and you restart kde again, your settings are lost.

 Worse, you loose all your virtual desktops, there is only one left.
 All windows lost their frames (looks similar to the windows, when you
 start compiz) and it is no more possible to move windows !

 Deleting ~/.kde or ./kde4 does NOT help ! So be warned: Get your old
 ~/.kde and ~/.kde4 as backup, otherwise you crash your whole kde !!!

 I could not find out, which part and files of KDE are responsible for
 this mess.

 It is the same, when you try this as the user root. You will crash
 kde of user root, too !

 Again, if you want to find out what happens, be very, very careful
 and get your old kde-settings as backup.

 At the moment I can confirm this behaviour only on an amd64-system,
 as for normal users, for root and as well as for newly added users.

 -


 If you know, please send this bugreport to the right maintainer or
 give me an advice, whom I have to send best.

 Thank you very much !

 regards

 Hans

Look at these messages on debian-user : 

http://lists.debian.org/debian-user/2008/07/msg02501.html

If I remember well I saw the solution in the last messages.

Regards,

Thomas Preud'homme

-- 
Why Debian : http://www.debian.org/intro/why_debian


signature.asc
Description: This is a digitally signed message part.


Re: hal buggy ? [Update]

2008-08-01 Thread Thomas Preud'homme
Le jeudi 31 juillet 2008, Hans-J. Ullrich a écrit :
 Hiho !

 I managed to get hal running again!

 What did I do ?

 Just downgraded hal and hal-info to the version of debian/stable.
 Doing so, all start and stop scripts were changed (although I never
 changed them myself) = no success !

 Then, upgraded to the version of testing = no success !

 Last, upgraded to the version of sid = success ! Yeah !

 So, it seems, there were some changes in the scripts (initscripts,
 start- and stop-scripts , whatever) between stable and sid, which
 might conflict with the binary of hal and hal-info.

Probably the init scripts which became dependancy based, like in gentoo 
I think. I didn't knew the change was already done.


 Anyway, this message shall just be for your info. Problem is solved !

 Cheers

 Hans

Regards

-- 
Why Debian : http://www.debian.org/intro/why_debian


signature.asc
Description: This is a digitally signed message part.


Re: EM64T compiling options?

2008-07-15 Thread Thomas Preud'homme
The Tuesday 15 July 2008 14:39:41 Lennart Sorensen, You wrote :
 You almost certainly wouldn't.  Very few people have any reason to write
 in assembly anymore.  You might write a critical section of code in
 assembly, but all the glue ought to be in C, inluding anything to manage
 pthreads to do multiple threads, which is how you take advantage of
 multiple cores.

You also can take a look at openmp which allow you to parallelize for loops 
and make easy synchronization. The support of openmp is included in latests 
gcc versions.


 --
 Len Sorensen



-- 
Thomas Preud'homme

Why debian : http://www.debian.org/intro/why_debian


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



Re: EM64T compiling options?

2008-07-15 Thread Thomas Preud'homme
The Tuesday 15 July 2008 15:46:06 E. Rens, you wrote :
 The loop sharing looks exciting but openmp seems difficult to use too.
 Does openmp replace pthreads or work in combination ?

AFAIK it use pthread to work but it create all necessary pthread at startup to 
avoid creating them at each loop parallelization. It should work properly 
with pthreads, it wouldn't be packaged with gcc if it wasn't the case. 
However I didn't test this combination.

openmp is really simple to use, easier than programming the same thing with 
pthreads. You can see examples on wikipedia :

http://en.wikipedia.org/wiki/OpenMP

The purpose is to do simple parallelization. If you want to write an algorithm 
which needs modification to support multi-cores, you have to do it yourself 
with pthreads.

Regards


 Emmanuel


-- 
Thomas Preud'homme

Why debian : http://www.debian.org/intro/why_debian


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