No middle-click on GNOME app title bar in KDE
So I have configured KDE to send a window to the background when I middle click in its title bar [1]. But when I do on the 'title bar' of a GNOME application (e.g. Gedit) nothing happens. [2] My understanding is that this is because GNOME applications don't have a window manager-controlled title bar. Instead they have an application-controlled "header bar" with a hardcoded behavior that does not match the KDE configuration. [3] BUT... A right-click in the "header bars" brings up a different menu depending on whether the GNOME application is running in GNOME or in KDE. So they are somehow aware of the environment they are in and appear to be able to adapt to it. The content of the right-click menu still does not match the one I get when right-clicking on a real title bar. So on the one hand GNOME header bars seem to be able to adapt to the desktop environment they are in, but on the other hand they still don't match it. Is it because I misconfigured something? Is there a special package that needs to be installed to let GNOME applications better integrate with KDE? SO... Do I have discrepencies because KDE never provided the required support? Or is it that GNOME applications lack the support for integrating with any environment besides GNOME? (and revert to a hardcoded default behavior in such cases) [1] System Settings -> Window Management -> Window Behavior -> Titlebar actions -> Titlebar and Frame Actions -> Middle click -> Lower. [2] In GNOME a middle-click on a GNOME header bar is treated as a left-click, that is it raises the window. But the same action on a GNOME header bar in KDE gives focus the first time, and raises it the second time. So if the window is already raised and has the focus nothing happens. [3] In fact one can configure Mutter so a middle-click lowers the window: gsettings set org.gnome.desktop.wm.preferences \ action-middle-click-titlebar 'lower' (set to 'none' for the default behavior) But this only works on real window manager-controlled title bars. So it works on KDE's Konsole for instance, but not on GNOME's own GNOME Terminal or Gedit! -- Francois Gouget http://fgouget.free.fr/ You don't liberate a people. A people liberates itself. Youssoupha
Re: No DNS in Fedora Podman image on Debian 11
On Sun, 7 Nov 2021, Ulf Volmer wrote: [...] > podman run --rm --security-opt=seccomp=unconfined -it fedora:latest > > solves the issue for me. That does work. Thanks! Today I also found that this is actually a known issue: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=995777 It indicates that this is going to hit all Linux containers using glibc >= 2.33. Fedora 35 is just the first casualty. The upstream bug that's referenced there also provides a fix and that has been integrated in the Podman that's in Debian Testing. (I checked that there is no issue on Debian Testing) However I don't know how to convert the 'correct' fix into something usable with the Debian 11 Podman; and the Debian Testing Podman (3.4) is not easily installable on Debian 11 (needs a newer libc). So until a fix makes its way into Debian 11 your workaround will be quite useful. So thanks again. -- Francois Gouget http://fgouget.free.fr/ Hell is empty and all the devils are here. -- Wm. Shakespeare, "The Tempest"
No DNS in Fedora Podman image on Debian 11
So I'm trying to use a fedora Podman image on my Debian 11 machine but for some reason DNS lookups do not seem to be working in the container environment. Specifically: $ podman run --rm -it fedora:latest # dnf install gzip [...] Fedora 35 - x86_640.0 B/s | 0 B 00:00 Errors during downloading metadata for repository 'fedora': - Curl error (6): Couldn't resolve host name for https://mirrors.fedoraproject.org/metalink?repo=fedora-35=x86_64 [getaddrinfo() thread failed to start] * I have the same issue on two Debian 11 systems (one of which is not administered by me). * The container can retrieve web pages with curl if I type in the IP address. So that confirms it's just the DNS that does not work. * debian:testing containers have no network or DNS issue. So it's just fedora:latest that's broken. * But I also have no issue with fedora:latest if I run it inside a Fedora 35 VM (Libvirt+QEmu specifically). * So it's the combination of a Debian 11 host + a Fedora container that's broken. * For good measure I tested with an "iptables -I (IN|OUT)PUT -j ACCEPT" on the host and it makes no difference. * In the guest /etc/resolv.conf has the domain line and "nameserver 10.0.2.3". * I see mentions of systemd-resolved on the Internet but I see no trace of systemd in the Fedora container. I don't know how to specifically test whever DNS lookups go through systemd-resolved though. Does anyone know what's up? Can anyone reproduce this issue? -- Francois Gouget http://fgouget.free.fr/ Un western sans indien c'est comme une police sans serif. -- John Wayne
Re: DNS server won't talk to me
On Fri, 20 Apr 2018, Greg Wooledge wrote: [...] > You misunderstand. That's not the resolver that Francois is using. > It's the authoritative name server for the domain he's trying to resolve > (maibokun.com). [...] > As a *workaround*, sure, he could use a public resolver like Google's > 8.8.8.8 as a sort of "proxy" that the Japanese name server is willing > to talk to. But short of that, he is completely cut off by the > router on the Japanese end. Yep. So much for the Internet being a peer-to-peer network :-( So in the end I configured things to go through a public resolver (but not Google). And while I was at it I installed dnscrypt-proxy (which was in the news recently and which Bob Weber also mentioned). -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ 145 = 1! + 4! + 5!
Re: DNS server won't talk to me
On Fri, 20 Apr 2018, Glenn English wrote: > On Thu, Apr 19, 2018 at 11:44 PM, Francois Gouget <fgou...@free.fr> wrote: > > > Are DNS servers banning queries from some residential addresses or > > something like this? > > I'm banning some, off and on, (I see massive hits from all over the > globe on my DNS server -- ~100K hits a day above my rate limit). Have > you tried to ping that unresponsive one to see if it's alive? Or a TCP > Telnet connection to its port 53? Indeed I cannot ping their DNS server (210.143.111.171) but I just thought they blocked ICMP. However I noticed I can in fact ping it from another host so I did a traceroute and the packets get blocked at the penultimate hop: $ traceroute -n 210.143.111.171 traceroute to 210.143.111.171 (210.143.111.171), 30 hops max, 60 byte packets [...] 21 60.37.54.202 296.022 ms 60.37.54.198 278.166 ms 122.1.245.126 274.472 ms 22 122.1.246.106 270.430 ms 275.228 ms 122.1.246.110 277.430 ms 23 211.0.221.30 273.257 ms 279.265 ms 277.767 ms 24 * * * On the other host the traceroute finishes with: 19 60.37.54.202 158.630 ms 122.1.245.130 161.021 ms 122.1.245.126 154.684 ms 20 122.1.246.110 147.979 ms 122.1.246.106 149.896 ms 122.1.246.110 155.476 ms 21 211.0.221.30 156.153 ms 144.694 ms 148.812 ms 22 210.143.111.171 156.433 ms 156.363 ms 159.304 ms > Is it possible that you've exceeded their rate limit? I have a script that would try to resolve the maibokun.com hostname once a day and the TTL on that appears to be 83334. So I would end up accessing their name server once a day. Of course now that it's not working and I have tried to figure out what's going on it's been quite a bit more. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ May your Tongue stick to the Roof of your Mouth with the Force of a Thousand Caramels.
DNS server won't talk to me
So I'm running a bind server and while it works I ran into a domain name that it refuses to resolve: maibokun.com. Digging into it, it looks like one DNS server is refusing to talk to me: On my box: $ host maibokun.com ;; connection timed out; no servers could be reached $ host maibokun.com 210.143.111.171 ;; connection timed out; no servers could be reached Same thing on my laptop. But if I connect the laptop to another Wifi network (thus changing it public IP address) or run the command on a computer on the other side of the atlantic I get: $ host maibokun.com maibokun.com has address 210.188.220.102 maibokun.com mail is handled by 10 mail.maibokun.com. $ host maibokun.com 210.143.111.171 Using domain server: Name: 210.143.111.171 Address: 210.143.111.171#53 Aliases: maibokun.com has address 210.188.220.102 maibokun.com mail is handled by 10 mail.maibokun.com. Are DNS servers banning queries from some residential addresses or something like this? Anyone else seeing the same issue? -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Hell is empty and all the devils are here. -- Wm. Shakespeare, "The Tempest"
Re: GNOME 3.20 dropped support for tap-to-click?
On Mon, 1 Aug 2016, Francois Gouget wrote:
[...]
> I tried some more mouse/touchpad drivers to no avail. To summarize:
> * xserver-xorg-input-synaptics: tap-to-click but no right click
> * xserver-xorg-input-libinput: right click but no tap-to-click
That should have been the opposite:
> * xserver-xorg-input-synaptics: right click but no tap-to-click
> * xserver-xorg-input-libinput: tap-to-click but no right click
[...]
> So no matter what, touchpads are broken in GNOME and so far as I can
> tell nobody cares :-(
But I finally found a workaround using the synaptics driver.
I don't have an xorg.conf file so I tried adding Michael's suggested
settings in /etc/X11/xorg.conf.d/50-synaptics.conf but that had no
effect. Maybe the identifier string was wrong or more likely the
configuration section just cannot stand on its own.
So then I copied /usr/share/X11/xorg.conf.d/70-synaptics.conf to
/etc/X11/xorg.conf.d and added the TapButton lines to the section titled
"Default clickpad buttons". So for me that section now reads:
# This option enables the bottom right corner to be a right button on clickpads
# and the right and middle top areas to be right / middle buttons on clickpads
# with a top button area.
# This option is only interpreted by clickpads.
Section "InputClass"
Identifier "Default clickpad buttons"
MatchDriver "synaptics"
Option "SoftButtonAreas" "50% 0 82% 0 0 0 0 0"
Option "SecondarySoftButtonAreas" "58% 0 0 15% 42% 58% 0 15%"
Option "TapButton1" "1"
Option "TapButton2" "2"
Option "TapButton3" "3"
EndSection
And that finally got me tap-to-click with the synaptics driver, without
breaking right-clicks. It's sad that getting such basic functionality
working feels like a breakthrough :-(
--
Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/
$live{free} || die "";
Re: GNOME 3.20 dropped support for tap-to-click?
On Fri, 1 Jul 2016, Francois Gouget wrote: > On Thu, 2 Jun 2016, Francois Gouget wrote: > > > On Thu, 2 Jun 2016, Michael Biebl wrote: > > [...] > > > Unfortunately the synaptics package in Debian was changed to have a > > > higher priority then libinput [2]. So if you have > > > xserver-xorg-input-synaptics installed you can't configure your touchpad > > > in gnome-control-center atm. > > > > Removing xserver-xorg-input-synaptics worked. Thanks! > > I've been meaning to send this for a while. > > The problem with removing xserver-xorg-input-synaptics is that then I > can no longer right-click using the touchpad, neither through a hard > click, nor by tapping. It looks like xserver-xorg-input-libinput either > does not support it, or is not configured to support it but then I have > not found any configuration option for it. > > So things are still not working as they should :-( I tried some more mouse/touchpad drivers to no avail. To summarize: * xserver-xorg-input-synaptics: tap-to-click but no right click * xserver-xorg-input-libinput: right click but no tap-to-click * xserver-xorg-input-mtrack:dead touchpad, mouse works * xserver-xorg-input-multitouch:dead touchpad, mouse works * xserver-xorg-input-mutouch: dead touchpad, mouse works So no matter what, touchpads are broken in GNOME and so far as I can tell nobody cares :-( -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Linux: It is now safe to turn on your computer.
Re: No PTP camera support in KDE
On Fri, 1 Jul 2016, Francois Gouget wrote: [...] > So every time I plug in either the Canon EOS 600 or Canon Ixus 970 IS > cameras via USB on my desktop computer I get a notification with a green > button to open the camera in the file browser. But clicking on that > button has no effect. Plugging in a regular USB key also produces a > notification and green button and in that case clicking on it works. > > When I plug in the same cameras on my laptop I can access the pictures > from the file manager normally. The difference is that the laptop is > running GNOME 3.2.0 while the desktop is running KDE 5.22 from Debian > Testing. I found a workaround: running Nautilus on KDE works just fine and can access the pictures on my PTP cameras! 1 point for GNOME, 0 for KDE -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Be careful of reading health books, you might die of a misprint. -- Mark Twain
Re: Truncating MPGs or MP4s
On Sat, 2 Jul 2016, brian wrote: [...] > I just need to remove that dead space - doesn't matter to me whether I > do that on the MPG or after conversion to the MP4. I don't need all > the bells and whistles that most video programs are likely to have, I > just need to be able to see the video and then mark a point after > which everything will be deleted - in other words, a very SIMPLE video > editor. I usually use avidemux for that, telling it to copy the audio and video rather then reencode them. Then I do the transcoding with handbrake. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Sufficently advanced incompetence is indistinguishable from malice.
No PTP camera support in KDE
So every time I plug in either the Canon EOS 600 or Canon Ixus 970 IS cameras via USB on my desktop computer I get a notification with a green button to open the camera in the file browser. But clicking on that button has no effect. Plugging in a regular USB key also produces a notification and green button and in that case clicking on it works. When I plug in the same cameras on my laptop I can access the pictures from the file manager normally. The difference is that the laptop is running GNOME 3.2.0 while the desktop is running KDE 5.22 from Debian Testing. As far as I know both cameras are seen as PTP devices and that seems to be the problem for KDE. From what I read KDE uses libgphoto which is installed (libgphoto2-6 and libgphoto2-port12) but I'm not sure which part KDE package would be specifically responsible for the KDE side. I have kamera 4:16.04.1-1 installed and when I start it it does see the Ixus (for instance) and it tests ok. Kamera's help says to type camera:/ into the Konqueror location bar. However: * The KDE file manager is Dolphin, not Konqueror. * Typing that address in Konqueror does not work (Improperly Formatted URL). * Typing that address in Dolphin gets me an icon for the camera but double-clicking on it ends up with an error saying it cannot communicate with the USB peripheral. Has anyone managed to access PTP devices in KDE on Debian? -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ RFC 2549: ftp://ftp.isi.edu/in-notes/rfc2549.txt IP over Avian Carriers with Quality of Service
Re: GNOME 3.20 dropped support for tap-to-click?
On Thu, 2 Jun 2016, Francois Gouget wrote: > On Thu, 2 Jun 2016, Michael Biebl wrote: > [...] > > Unfortunately the synaptics package in Debian was changed to have a > > higher priority then libinput [2]. So if you have > > xserver-xorg-input-synaptics installed you can't configure your touchpad > > in gnome-control-center atm. > > Removing xserver-xorg-input-synaptics worked. Thanks! I've been meaning to send this for a while. The problem with removing xserver-xorg-input-synaptics is that then I can no longer right-click using the touchpad, neither through a hard click, nor by tapping. It looks like xserver-xorg-input-libinput either does not support it, or is not configured to support it but then I have not found any configuration option for it. So things are still not working as they should :-( -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Cahn's Axiom: When all else fails, read the instructions.
Re: GNOME 3.20 dropped support for tap-to-click?
On Thu, 2 Jun 2016, Michael Biebl wrote: [...] > Unfortunately the synaptics package in Debian was changed to have a > higher priority then libinput [2]. So if you have > xserver-xorg-input-synaptics installed you can't configure your touchpad > in gnome-control-center atm. Removing xserver-xorg-input-synaptics worked. Thanks! > This situation is unfortunate, but we are aware of it and hope to figure > something out for stretch. xserver-xorg-input-all depends on xserver-xorg-input-synaptics. So maybe the first step would be to remove that dependency since it does not seem to be needed anymore. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ I haven't lost my mind, it's backed up on tape around here somewhere...
GNOME 3.20 dropped support for tap-to-click?
Since the upgrade to GNOME 3.20.1 I lost tap-to-click in GNOME.
I cannot reestablish it in gnome-control-center as that only offers to
switch the left and right buttons.
Does anyone know of a workaround to that?
Note that:
* xserver-xorg-input-libinput is installed.
* Tap-to-click still works in gdm3. See bug 731439:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=731429
* However the same command I used for bug 731439 has no effect in my
GNOME session. That is:
gdm gsettings set org.gnome.desktop.peripherals.touchpad tap-to-click true
has no effect
* I also tried the following command with values 1 through 12 with no
result:
xinput set-prop "ETPS/2 Elantech Touchpad" "Synaptics Tap Action" 1
('ETPS/2 Elantech Touchpad' is the name of my touchpad btw)
* I did not find anything related to tap-to-click in gnome-tweak-tools.
--
Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/
1 + e ^ ( i * pi ) = 0
Re: Reporting unmaintained packages
On Thu, 21 Jan 2016, Chris Bannister wrote: [...] > Also, I'm under the impression that the bug submitter doesn't > automatically get sent any responses to the bug report. I don't know if I receive them all, but I do receive responses to bug reports I submitted. I'm not so sure about bug reports where I just sent additional information. I think for those one has to explicitly subscribe to the bug which seems like a good idea. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ La terre est une bêta...
Re: Setting up tap-to-click in gdm3
On Mon, 14 Dec 2015, Sven Arvidsson wrote: [...] > Try this: > http://forums.debian.net/viewtopic.php?p=347944 I finally got time to test this but none of the recommendations on that post worked. I finally found GNOME bug 747811 which indicates that this should really be configured through gsettings. https://bugzilla.gnome.org/show_bug.cgi?id=747811 The trick is to issue a xhost command first: $ xhost +SI:localuser:Debian-gdm $ su # DISPLAY=:0 sudo -u gdm gsettings set org.gnome.desktop.peripherals.touchpad tap-to-click true I'm not sure why the access permissions of the logged in user are relevant to the gdm3 configuration but that worked. I'd also expect to be able to set this up by adding a file in /usr/share/gdm/dconf/ but I tried the lines below and they had no effect :-( cat >/usr/share/gdm/dconf/50-tap-to-click < http://fgouget.free.fr/ A black hole is just God dividing by zero.
Re: Reporting unmaintained packages
On Mon, 18 Jan 2016, Francesco Ariis wrote: > On Mon, Jan 18, 2016 at 12:36:34PM +0100, Francois Gouget wrote: > >> The clamav-unofficial-sigs package has quite important bugs that cause > >> it to fail to retrieve the SecuriteInfo virus signatures and send cron > >> spam every 4 hours. > >> > >> [..] > >> > >> So what's the proper way to report this issue? > > Hello Francois, > I assume the bug you are talking about is #783228 [1]. > clamav-unofficial-sigs is not maintained by a single person, but by > ClamAV Team. Actually I think the following three bugs are duplicates of each other. At least now if not initially (various SecuriteInfo databases went offline progressively so symptoms changed over time). * 783228: clamav-unofficial-sigs: securiteinfo databases not available any more https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=783228 * 784832: clamav-unofficial-sigs: Multiple error message at each execution https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=784832 * 774763: clamav-unofficial-sigs: Updating the databases timeouts on a regular basis https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=774763 (the timeouts are now 404s) Here is the activity for these bugs: Bug| Reported | User-provided workaround | ClamAV Team reply 774763 | 2015/01/07 | 2015/04/24 | none 783228 | 2015/04/24 | 2015/04/24 | none 784832 | 2015/05/09 | 2016/01/18 | none So the the issues were reported over a year ago, workarounds provided over 8 months ago, but the ClamAV team is nowhere to be found, hasn't asked for more details, hasn't closed duplicate bugs, hasn't made any new release of this package. So I did send more data for bug 774763 and 784832 but I'm mostly just repeating information that's already available on bug 783228. So given that information was available 9 months ago I'm not too hopeful. I could also send a patch but is it really necessary when the 'fix' is as simple as setting si_dbs="" in 00-clamav-unofficial-sigs.conf as was described in bug 783228 (again, 9 months ago)? The right fix might be to upgrade to the newer upstream version available from GitHub as reported in bug 785130, 9 months ago (that bug got no reply at all). https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=785130 But then is it really the place of a user to provide a brand new package for the maintainer to just push out? And I'm not willing to take over maintainership because a) I'm not a Debian developer and b) I know I won't have time to keep doing it. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ La terre est une bêta...
Reporting unmaintained packages
The clamav-unofficial-sigs package has quite important bugs that cause it to fail to retrieve the SecuriteInfo virus signatures and send cron spam every 4 hours. These issues were reported in april 2015 and changing the package so it no longer tries to retrieve these virus signatures if they are no longer available would be quite simple. Despite this the package has not been updated since 2014. I'm not sure the developer is MIA but he does have quite a few other packages to maintain so may he's swamped or just lost interest in this particular package. https://qa.debian.org/developer.php?login=Paul+Wise Having that having out of date or missing virus signatures has security implications (more for some users that others, I'll grant you), getting a handle on these bugs seems quite important. So what's the proper way to report this issue? -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ E-Voting: It's not the people who vote that count. It's the people who count the votes.
Re: Setting up tap-to-click in gdm3
On Mon, 7 Dec 2015, Sven Arvidsson wrote: > On Mon, 2015-12-07 at 10:36 +0100, Francois Gouget wrote: > > Does anyone know how to set up tap-to-click in gdm3? > > > > The goal is to be able to tap on the trackpad to perform the swipe > > gesture that's now required to get the list of users, and tap again > > to > > to pick the user to log in as. > > https://wiki.archlinux.org/index.php/GDM#Enabling_tap-to-click > > Haven't tried it myself, so let us know if it works! It did not work :-( I my regular account a grep for 'tap-to-click' in ~/.conf does find a match in ~/.conf/dconf/user. So setting this option through dconf or gsettings is probably the right basic idea. On Debian the gdm3 user is called 'Debian-gdm' and its shell is '/bin/false'. So I su-ed to root and ran the following command: sudo -u Debian-gdm gsettings set org.gnome.desktop.peripherals.touchpad tap-to-click true But I got the following error: (process:26300): dconf-WARNING **: failed to commit changes to dconf: Error spawning command line 'dbus-launch --autolaunch=f856067f8571c63f807f75310009 --binary-syntax --close-stderr': Child process exited with code 1 I'm not sure where to go from there. ~/.conf/dconf/user is a binary file so editing it directly is out. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Cahn's Axiom: When all else fails, read the instructions.
Setting up tap-to-click in gdm3
Does anyone know how to set up tap-to-click in gdm3? The goal is to be able to tap on the trackpad to perform the swipe gesture that's now required to get the list of users, and tap again to to pick the user to log in as. -- Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/ Advice is what we ask for when we already know the answer but wish we didn't -- Eric Jong
Re: Status of repository debian/testing?
On Sun, 27 Sep 2015, David Wright wrote:
[...]
> So waiting could still be a sensible option at this time. Some of us
> have yet to finish sorting out jessie (in my case, as a production
> system, not as an upgrade target).
As far as I can tell waiting is unlikely to solve the upgrade issues.
That's what I did for a while, manually holding back packages until
there was no conflict anymore (since aptitude was totally overwhelmed),
but things did not get better. Finally I got enough insight into the
conflicts to solve them. The issues I found fall in just a few
categories:
* A lot of packages have been renamed from 'foo' to 'foov5'. If your
'foo' package was not marked as automatically installed ('i ' rather
than 'i A' in aptitude), then aptitude and apt-get will think it's
because you explicitly want the 'foo' package rather than the 'foov5'
one. So it will get stuck because the new packages want 'foov5' but it
cannot remove 'foo', resulting in a conflict. This is the issue that
waiting will never solve.
-> The fix: In aptitude use 'b' to go from one broken package to the
next, open the package page and go to the 'Conflicts' list.
If this shows the package conflicting with a 'foov5' package,
go back to the top of the page and press 'A' to mark the package as
automatically installed. It should be scheduled for deletion and
the 'foov5' should no longer be marked as broken.
(You can also go to the 'foov5' package to mark that you really
want it: '+')
* I had the same issue with the 'mysql-*-5.5' versus 'mysql-*-5.6' set
of packages and possibly with a couple others I forgot.
-> Fix: Again, mark the 5.5 packages as automatically installed and
aptitude will figure it out.
* I have both the 32 and 64 bit versions of libxml2-dev installed so I
can do proper Wine development work. The problem is libxml2-dev
depends on libicu-dev which depends on g++. But neither g++ nor g++-5
are not a multiarch package, not even 'MultiArch: foreign' ones. So
although libicu-dev is marked as multiarch, only one version can be
installed at any given time, unlike in previous versions (see bug
799100). This then caused a cascade of breakage.
-> Fix: The only way for now is to give up on multiarch for
libxml2-dev and revert to manually creating libFoo.so links in the
relevant /usr/lib directory :-( And to think multiarch,
"including cross-compiling environments for embedded systems" was
a release goal of the two year old Debian 7 and is marked as
"Completed in Wheezy". What a bad joke! [1]
* I had BuildBot installed but it had trouble with python-sqlalchemy: it
depends on 'python-sqlalchemy (< 0.10)' which is only available in
Debian Stable. It also depends on python-migrate which, if installing
the version available in Debian Testing, depends on python-sqlalchemy
>=1.0~). Instant conflict (see bug 794300). Combined this confused
aptitude and me enough that I removed buildbot for the upgrade.
-> Fix: Pick the python-sqlalchemy and python-migrate packages from
Debian Stable.
* Once I completed the upgrade I got a broken sddm (and if the results
when I replaced it with lightdm are anything to go by, a broken KDE
/ Plasma too). This is because of a missing version requirement of
some package on libqt5sql5 (see bug 802811).
-> Fix: Install libqt5sql5 and the packages it depends on from Debian
Unstable. There's already a fix for that which should get to Debian
Testing one day.
With all that I now have an _almost_ usable desktop.
- sddm does not start on boot, nor from the console. But I can log in
remotely and start it manually from there (see bug 803324).
- Switching the desktop effects off often kills kwin such that it does
not restart. But I really only need it to work once and I can start it
manually.
- plasmashell crashes multiple times per day, just as before the
upgrade (see bug 794110 and bug 801501). Fortunately it restarts
automatically so it just means I have to dismiss the crash dialog
regularly.
- KWallet keeps asking me to set it up whenever I *quit* Google
Chrome or *stop* openvpn as root (note: quit or stop, not start in
both cases! see bug 797877) even though I have no intention of using
it.
- And I have lost sound. Alsa does see the Intel HDA soundcard but
PulseAudio/Phonon don't and stick to the dummy sound output. Somehow
my user account got removed from the audio group but adding it back
and a logout+login did not solve the issue. I still have not figured
that one out yet.
But at least I should now be able to apply updates normally.
[1] https://wiki.debian.org/ReleaseGoals/
https://wiki.debian.org/ReleaseGoals/MultiArch
--
Francois Gouget <fgou...@free.fr> http://fgouget.free.fr/
Any sufficie
Re: Trouble getting net-snmp to return lm-sensors data (solved)
On Fri, 31 May 2013, Francois Gouget wrote: So I'm trying to get net-snmp to return lm_sensors data so I can monitor the temperature and fan speed of my PC. But currently it won't return any data: $ snmpwalk -v2c -c public localhost lmSensors LM-SENSORS-MIB::lmSensors = No more variables left in this MIB View (It is past the end of the MIB tree) It turns out this was an access control issue. Uncommenting the following line fixed it: rocommunity public localhost -- Francois Gouget fgou...@free.fr http://fgouget.free.fr/ It really galls me that most of the computer power in the world is wasted on screen savers. Chris Caldwell from the GIMPS project http://www.mersenne.org/prime.htm -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/alpine.DEB.2.10.1306120019490.31015@amboise.dolphin
Trouble getting net-snmp to return lm-sensors data
So I'm trying to get net-snmp to return lm_sensors data so I can monitor the temperature and fan speed of my PC. But currently it won't return any data: $ snmpwalk -v2c -c public localhost lmSensors LM-SENSORS-MIB::lmSensors = No more variables left in this MIB View (It is past the end of the MIB tree) Note that it's not saying that it does not know about lmSensors so it seems like it found the MIB alright. Also sensors does return what seems like suitable data to me: $ sensors coretemp-isa- Adapter: ISA adapter Core 0: +50.0°C (high = +74.0°C, crit = +100.0°C) Core 1: +50.0°C (high = +74.0°C, crit = +100.0°C) Core 2: +46.0°C (high = +74.0°C, crit = +100.0°C) Core 3: +50.0°C (high = +74.0°C, crit = +100.0°C) it8718-isa-0290 Adapter: ISA adapter Vcore:+1.12 V (min = +0.00 V, max = +4.08 V) Vram: +1.87 V (min = +0.00 V, max = +4.08 V) +3.3V:+3.15 V (min = +0.00 V, max = +4.08 V) +5V: +4.70 V (min = +0.00 V, max = +6.85 V) +12V:+12.16 V (min = +0.00 V, max = +16.32 V) Vbat: +3.26 V CPU Fan: 1207 RPM (min = 10 RPM) F/Low Fan: 0 RPM (min = 10 RPM) F/High Fan: 1300 RPM (min =0 RPM) NBr Temp: +45.0°C (low = +127.0°C, high = +127.0°C) sensor = thermistor CPU Temp: +34.0°C (low = +127.0°C, high = +80.0°C) sensor = thermal diode intrusion0: OK In particular notice the 'V' and 'Temp' labels which I believe should allow net-snmp to recognize these as voltages and temperaturs. Here's what I did so far: * I'm running Testing so I installed: snmp 5.4.3~dfsg-3 snmpd5.4.3~dfsg-3 snmp-mibs-downloader 1.1 libsnmp155.4.3~dfsg-3 lm-sensors 1:3.3.3-1 * Commented out the 'export MIBS=' line in /etc/default/snmpd. I also tried pointing it to specific MIB files to no avail. * Commented out the 'mibs :' line in /etc/snmp/snmp.conf. * Here's a summary of /etc/snmp/snmpd.conf: agentAddress udp:127.0.0.1:161 view systemonly included .1.3.6.1.2.1.1 view systemonly included .1.3.6.1.2.1.25.1 rocommunity public default-V systemonly rouser authOnlyUser sysLocationSitting on the Dock of the Bay sysContact Francois Gouget fgou...@free.fr sysServices72 proc mountd proc ntalkd4 proc sendmail 10 1 disk / 1 disk /var 5% includeAllDisks 10% load 12 10 5 trapsink localhost public iquerySecName internalUser rouser internalUser defaultMonitors yes linkUpDownNotifications yes master agentx * I found a comment that claims that net-snmp 5.5 or greater is required to get integration with lm-sensors. Yet I believe that functionality was present long ago and could not find confirmation. Also net-snmp 5.5 is not available in Testing. http://www.net-snmp.org/wiki/index.php/Net-SNMP_and_lm-sensors_on_Ubuntu_10.04 So has anyone else been able to get net-snmp and lm-sensors to play nice together. Any special tricks? -- Francois Gouget fgou...@free.fr http://fgouget.free.fr/ The nice thing about meditation is that it makes doing nothing quite respectable -- Paul Dean
Re: ssh and X
On 25 Dec 2001, Jens Müller wrote: Since upgrading to Woody, X forwarding over ssh does not work any longer, which is a bad thing, since I cannot run X programs as super-user now. Finding why ssh -X does not work anymore is a good thing. But your comment that says you cannot run programs as root anymore suggests that your are using ssh as a kind of 'su' alternative. In that case I suggest that you look at sux, an su-wrapper script which will do the xauth magic necessary for root to get access to your screen. Instead of typing 'su -', just type 'sux -'. You can get sux there: http://fgouget.free.fr/sux/sux and the Readme: http://fgouget.free.fr/sux/sux-readme.shtml One advantage is that with sux your X connection does not through the encryption/decryption cycle. But one thing it does not do is forward you ssh-agent. I which there was an ssu so that I could do 'ssu -A -X - foo' and have ssu login into my foo account (using a public key for instance), set up the xauth stuff and forward the ssh-aget connection. But not go through the encryption/decryption of each X packet! Well, I hope you will find sux useful. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ RFC 2549: ftp://ftp.isi.edu/in-notes/rfc2549.txt IP over Avian Carriers with Quality of Service
Re: vncserver ignoring ~/.vnc/xstartup :(
On Sun, 18 Nov 2001, Martin Rowe wrote: Hi all I'm trying to get VNC running on my home box and vncserver sems to be bypassing the xstartup script. Instead of running Blackbox, it pulls up KDE. The .vnc/ session log doesn't report any problems, and I can run xstartup within the VNC window and it runs (just lanches an aterm) fine - just on top of KDE. This has worked fine for me in the past, though on Mandrake, and runs okay at work on Debian (though launched from inetd instead of manually starting each session, and using a login manager, so no xstartup involved). The only similar problem I've seen hunting via Google has also been on a Debian system, so I was wondering if there was a difference in the packaging of VNC for Debian - or more likely I've missed something obvious somewhere. Replying late but... I believe this can be caused by x-session-manager. When KDE is installed it seems to systematically start its own window manager, which means that when twm gets to start there is already a window manager running so it fails. Well, in any case I had this problem when I installed KDE on my box, x-window-manager was not used anymore so I deleted the x-session-manager symlink. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Any sufficiently advanced Operating System is indistinguishable from Linux
Re: X-forwarding lazy
On 3 Nov 2001, Jens Müller wrote: Since I have upgraded to Woody, X-Forwarding via slogin is VERY lazy. I use it as slogin localhost su in order to be able to start X programs as root ... You should use sux instead: http://fgouget.free.fr/sux/sux-readme.shtml http://fgouget.free.fr/sux/sux -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Any sufficiently advanced bug is indistinguishable from a feature. -- from some indian guy
Re: QuickTime
On Wed, 10 Oct 2001, dooshiant wrote: http://openquicktime.sourceforge.net/ I don't know much Spanish but I guess that's what you're looking for. :-) Or you could also have a look at CrossOver (I don't know Spanish either). It's not open source but it will let you view Sorenson movies too. http://www.codeweavers.com/products/crossover/ You will also be able to use other Windows plugins like Shockwave and the Word/Excel and Powerpoint viewers. (disclaimer: I work for CodeWeavers) -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ We are Pentium of Borg. You will be approximated. Division is futile.
Re: konqueror config error
On Thu, 30 Aug 2001, Greg Ward wrote: [...] First, I did an apt-get --purge remove list_of_all_kde_packages, then a reinstall of the same list. The config problem still persisted. I then repeated the purge, removed ~/.kde, /var/lib/kde2, /etc/kde2, and every other kde-related file I could find on the system. This time, when I reinstalled kde, things were working. I did this, and now it works! I can run Konqueror, kscd, and kmix (the only KDE apps I use regularly) without problems. This is KDE 2.1.1. I'm not entirely sure where I downloaded the .debs from; for a while I had You must not have reinstalled the exact same set of package then. Or maybe kdebase was not properly installed on your system the first time. I uninstalled all the kde packages, cleaned up and reinstalled the exact same set of packages and it still did not work. So i investigated a bit more and found the kcmkonq.desktop file. It contained a reference to kcontrol/file-manager.html which was not installed on my system. So I searched for it on the Debian site and found that it was provided by kdebase. So I installed kdebase and this solved the problem. Except that kdebase recommends kdewallpapers, kscreensaver and kdebase-audiolibs (plus a few others) which I definitely do not want to install on my system (why would I?). So I am now in recommends~depends hell in dselect. Btw, I entered a bug report about this problem (I mention it here to avoid duplicate bug reports). I don't know yet the bug number but the subject is 'konqueror depends on kdebase...'. I'm sure this will be resolved in due time. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Good judgment comes from experience, and experience comes from bad judgment -- Barry LePatner
Re: Looking for a wireless ethernet solution...
On Thu, 30 Aug 2001, Joey Hess wrote: [...] This would require you also buy a pcmcia adapter for your linux box, since all the networking gear comes pretty much as pcmcia cards. I believe that the access points don't come with a pcmcia card either, so you'll need three in all. Gear list then comes down to (you may find slightly better prices): The access point does not need a pcmcia card. All the electronics for doing wireless stuff is already included. So the list would become: $150+ 1 access point $75 1 pcmcia to ISA or pcmcia to PCI bridge $210 2 pcmcia cards $435 total But the prices I have seen are a bit higher: usually around $250 for the access point and $150 for each pcmcia card. This raises the total cost to $625. Here is an alternative: windows laptop | . | . hub - linux box dsl | . | . friend's box . guy in a van on the street Here you keep the wired link between the linux box and the hub to the windows box, and do not use an access point, instead using Ah-Hoc networking mode between your linux box and the laptop. This requires only 2 pcmcia cards, so the gear is: $75 1 pcmcia to ISA or pcmcia to PCI bridge $140 2 pcmcia cards $215 total Or $375. But there is a problem with this solution: you need 3 available PCI slots. One for the wireless card, one for the ethernet card to the hub, and one for the ethernet card to the dsl. I'm also thinking about getting a wireless setup and this has kind of blocked me for now. But there may be an alternative. The traffic between the dsl and the linux box is likely to be PPP: PPPoE. And some access points can even speak PPP on their wired side. So the idea is to basically connect everything to the hub. dsl | windows ---+ (ppp) || friend's box ---+--(ethernet)-- hub || linux box---+ (ppp) | laptop . . . . . access point . . . . . guy in a van The idea is that the PPP traffic would be relayed from the access point and dsl to the linux box (and all others) and that the linux box would pick it up there. This should be safe in that the 'guy in the van' cannot scan your network since all the access point spits out on your lan is PPP packets. The only security risk is that anyone on the lan can snoop on this traffic, but this may be deemed acceptable. Same for the DSL. But there's a couple of big question marks there: * can an ethernet hub relay PPP traffic? * what about a switch (I have a switch at home) * will the PPP traffic and the IP traffic interfer * in the case of the access point, the access point is expecting the linux box to play the role of a dsl modem. Is there software that can handle that on Linux? * PPP is a point to point protocol. Will it work to have two such devices share just one ethernet segment? (if not we can still put the dsl on it which still saves a PCI slot) Oh, and yes, this is probably much more complex. But you get to play with Linux firewalling, auto-proxying, ... And to complete the picture I would use the wireless network in unencrypted mode so that anyone can access it (and since the 802.11 encryption is relatively useless anyway), but only allow access to the internet (firewalled to prevent too nasty things from happening). Then I would use a VPN between the laptop and the linux box, and would only allow the traffic from this VPN to access the LAN. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ War doesn't determine who's right. War determines who's left.
konqueror config error
Hi, Each time I go to Settings/Configure Konqueror I get the following error messages: * Error - KDE Control Module There was an error loading module 'Settings/FileBrowsing/kcmkonq.desktop' The diagnostic is: And that's all, it never says what the diagnostic really is. * Then I get the same error for 'Settings/FileBrowsing/filetypes.desktop' and 'Settings/WebBrowsing/konqhtml.desktop'. So what's wrong? Did I forget to install some module or something? I cannot find any of these files on my system. Also I see that Greg Ward got the same problem but he never got any answer, at least not in public it seems (see http://lists.debian.org/debian-kde/2001/debian-kde-200105/msg00293.html). I would very much like to access the 'filetypes' settings. I believe this is the one I need. Here's what I have on my box (a debian-testing box): ii kdebase-crypto 2.1.1.0-3 KDE core applications (Crypto modules) ii kdebase-doc2.1.1.0-10 Documentation for Applications in kdebase ii kdebase-libs 2.1.1.0-10 KDE libraries amd modules for kdebase ii kdelibs3 2.1.2-3KDE core libraries (runtime files) ii kdelibs3-crypt 2.1.2-1KDE core libraries (Crypto Modules) ii konqueror 2.1.1.0-10 KDE's advanced File Manager, Web Browser and ii lesstif1 0.92.26-2 OSF/Motif implementation released under LGPL ii libc6 2.2.3-9GNU C Library: Shared libraries and Timezone ii libjpeg62 6b-1.3 The Independent JPEG Group's JPEG runtime li ii libkonq3 2.1.1.0-10 Core libraries for KDE's file manager ii libpng21.0.12-2 PNG library - runtime ii libqt2 2.3.0-final-5 Qt GUI Library (runtime version). ii libstdc++2.10- 2.95.4-0.01042 The GNU stdc++ library ii xlibs 4.0.3-4X Window System client libraries ii zlib1g 1.1.3-15 compression library - runtime -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ It really galls me that most of the computer power in the world is wasted on screen savers. Chris Caldwell from the GIMPS project http://www.mersenne.org/prime.htm
Re: NFS alternative
On 17 Jul 2001, Brian May wrote: [...] I have heard though, that NFS uses RPC, and RPC doesn't do anything this sophisticated. Rather, if the server doesn't respond in time, the client sends the request again. This is OK for requests like read file, which don't affect the state of the server, as the worst thing that could happen is that the server will reread the file again. For other requests though, like delete file, confusion could occur. eg. consider the case where the client wants to delete file X, [...] Well, of course you can bet that all these issues have been taken into account in the design of the NFS protocol. But actually this touches an important aspect of NFS servers: by design/philosophy they are stateless. This means that the NFS server can crash, reboot, and to the client it makes strictly no difference (the same goes for the client). And this fits well with UDP since it's a connectionless protocol. Compare this to the SMB protocol used by Windows. This protocol is resolutely stateful. And guess what: SMB uses TCP connections. So if the server crashes then the clients have to reconnect to the server and try to resynchronize their state with the state of the server somehow. Well, of course, the designers of the SMB protocol thought of the issue too and I'm sure it's all taken care of. Now, you are worried that if an NFS UDP packet is lost you'll get into trouble. But what if an SMB client sends an 'append' command to the server, the server performs the operation but crashes before sending the reply, reboots, causing the client to reconnect and resend the command... You see, using TCP did not save you completely. You get exactly the same problem as if the NFS server had crashed and rebooted, or the command packet got lost, or the reply packet got lost. I believe that the work to recover from such problems must be relatively equivalent in both cases. Btw, I doubt that there is an 'append' command as it is so obviously not 'idempotent'. Of course there must be some sort of 'delete_file' operation and I'm not sure how them make that one 'idempotent'. But in multi-user systems how do you know if the file was not deleted by some other user in the first place... Conclusion: (if my sources are correct) never use NFS (unless mounted read-only) over a congested network... I think you need to do some reading. I don't know of a reference book that would be specifically about NFS but you can have a look at The Design and Implementation of the 4.4BSD Operating System by McKusik, Bostic, Karels and Quarterman. Chapter 9 (about 25 pages) is about NFS and should give you a feel for it. And the book as a whole is just awsome. Definitely a must have if you're interested in operating system design. I'm definitely going to re-read this chapter... -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ 1 + e ^ ( i * pi ) = 0
Re: NFS alternative
On Mon, 16 Jul 2001, D-Man wrote: On Mon, Jul 16, 2001 at 12:45:22PM -0500, Dimitri Maziuk wrote: | * D-Man ([EMAIL PROTECTED]) spake thusly: | ... | Ok, that makes sense. How about if probability leaves us behind | and a packet is lost? Does NFS provide any way to correct for | that or will your filesystem be hosed? | | Thankfully, I forget the details[0]. From experience, no, it won't | be exactly hosed: you'll end up with a .nfs004950384672385721380937 | file that will grow and eventually fill up the partition... nothing | an rm -rf / won't fix. And then there's negative cookies and stale | mounts that require a reboot on most unices I've seen... Ok. It sounds like it would still result in data loss :-(. Hmm, I'm not an NFS expert but I'll play one on the mailing-list for you ;-) Please, if there are experts out there, correct me if I'm wrong. AFAIU, NFS has its own mechanism to recover from lost packets. So it won't be a problem if a packet is lost. Similarly I believe NFS RPCs cannot span UDP packets, so there is no chance that a lost packet would change the meaning of an RPC. The RPC will be lost, pure and simple, and NFS will have to reissue it or something similar. So I don't think packet loss is an issue. What NFS is 'lacking' is congestion control, as in the TCP slow star and exponential back-off. This means NFS will blast UDP packets as fast as it cans with no regard for other trafic. This is not really an issue on a lan and actually had a performance at a time (I think). But if you go over multiple links, then you may saturate a slower link, causing the router that is just before it to start dropping packets. Especially if multiple streams converge there. And once you start dropping packets performance degrades very significantly. I believe that's why NFS is bad if there are multiple hops (I get it from a very reliable source that this is also why it's very bad if the traffic will go over ATM, you need buffering/traffic shaping). -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Any sufficiently advanced bug is indistinguishable from a feature. -- from some indian guy
Re: how to run X apps as root?
On Tue, 3 Jul 2001, Arne Goetje wrote: Hi, sometimes I need to run some X apps as root (i.e. ethereal, lprngtool) but Xlib complains that the server refuses the connection. how can I make the Xserver accept programs as root when I'm working as normal user? That because Debian recently changed to disable xhost, and now restricts access on a per user basis (kind of). That's a good move because xhost was pretty insecure. Now to access X you must show that you know a secret 128bit or something number that is generated when the X session starts. So in order for root to be able to access the X server, you must give him this secret value using xauth. But that a pain to do each time you do an su. So I wrote an su wrapper which transfer $DISPLAY and also transfers the X cookies that you must have to gain access to the X server. I called is sux and you can get it from http://fgouget.free.fr/sux/ You use it just like su: $ sux - or $ sux - foo or $ sux --untrusted --timeout 30 - foo xterm - created an untrusted cookie that becomes invalid after 30 seconds of inactivity, typically 30 seconds after you've closed the xterm. 2nd Problem: I tried today to run a second instance of startx with -- :1 to run it as root. today it didn't work anymore... I'm using unstable. Maybe that because the vt is missing: $ startx -- vt8 :1 or maybe it is $ startx -- :1 vt8 -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Before you criticize someone, walk a mile in his shoes. That way, if he gets angry, he'll be a mile away - and barefoot.
Re: how to run X apps as root?
On Tue, 3 Jul 2001, christophe barbé wrote: SUX seems pretty usefull. What about a debian package for it (or integrating it in an another)? I was thinking about proposing to add it to xbase-clients but the response was not very enthusiastic. So yes, maybe I'll package it. Would you be interested about packaging it? Is there something similar for remote X session ? for remote X sessions you should use ssh. For instance: $ slogin -X -l foo remote.dom The -X activates X forwarding if it is not the default (it's not on Debian). With -X the remote application will see a $DISPLAY value that looks like remote.com:15. This is in fact handled byt the ssh server which will forward all such X requests to your local slogin process which will then send them to your X server. This means the remote application communicates with your X server via an encrypted communication channel, it's pretty sweet. Of course, on the remote end, access to this remote.com:15 is subject to an xauth-style authentication mechanism. So here too sux can come in handy. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ The software said it requires Win95 or better, so I installed Linux.
Re: .Xauthority and kdm
On Mon, 2 Jul 2001, tim wrote: hello all I upgraded to kde 2.2beta. I start now to runlevel 5 and kdm. My problem is that I dont know how to configure Xauthority that other users (even root) can acces the display. I get the following message: --/ [EMAIL PROTECTED]:~$ su Password: tim:/home/tim# xawtv This is xawtv-3.53.1, running on Linux/i686 (2.4.5) Xlib: connection to :0.0 refused by server Xlib: Invalid MIT-MAGIC-COOKIE-1 key Error: Can't open display: :0 tim:/home/tim# ---/ When I start into runlevel 2 and startx I set the permission with xhost myhost. This is secure enough for me because its my privat firewalled network. But xhost doesnt work in runlevel 5... I went thru the manual of xauth but cant figure out how to give anyone (or specific user) access to the display? Let's forget about the various more or less good tips and tricks that were provided before. This xauth issue has been discussed on debian-devel last week and I coded an su wrapper script that does all the xauth legwork for you. So here's what I recommend: * go to http://fgouget.free.fr/sux/ * download the sux script * install it in /usr/local/bin * and then use the script to su as root: 'sux -' is the preferred way * you can even do as me and put 'alias su=sux' in your .bash_profile For other readers, note that 'sux - foo' works too, unlike the dangerous XAUTHORITY tricks. I say dangerous because if you set 'XAUTHORITY=/home/foo/.Xauthority' and then do a an xauth add/remove, the .Xauthority file will belong to root, which means foo won't have access to *his* X display anymore!!! Oh, and if you played tricks with xhost, remove them they won't be necessary anymore, and fix the permissions on that 'video device', 0666, ksss! -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Any sufficiently advanced Operating System is indistinguishable from Linux
Re: AVI Player (mplayer)
On Mon, 11 Jun 2001, Reza wrote: Hi... I think someone suggested to use Mplayer for AVI player, but now, I have a little bit problem when I'm playing DivX movies, the sound is not good, it stop sometimes, and then I could hear the sound again, and it only happened if I play DivX movies, and when I type mplayer -vo x11 filaname.avi one of the line says: Detected AVI file format! == Found video stream: 0 == Found audio stream: 1 Can you please help me with this? Thank you :) Run top. - if you have 100% CPU usage then it's not surprising that the sound falls behind sometimes. Try to upgrade to XFree 4, use xv... whatever that would decrease CPU usage and allow mplayer to keep up. - if your CPU usage is way below 100% then it's probably a problem with the sound driver. I have this problem on my laptop (ALSA snd-ymfpci driver) and I have no idea what to do, except upgrade to a more recent release of ALSA one of these days. On my machine this also causes video choppiness: if I run 'mplayer -nosound' the video is fine, otherwise it's choppy and the sound is interrupted every few seconds (even with 20% CPU usage). -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Stolen from an Internet user: f u cn rd ths, u cn gt a gd jb n cmptr prgrmmng !
Re: [OT] Minimum CPU speed for playing DIVX avis?
On Thu, 7 Jun 2001, Viktor Rosenfeld wrote: Hi folks, I got mplayer (http://mplayer.sourceforge.net) working very nicely, only that CPU utilization is constantly at 100% (~ 92% user, 8% system) and sound is very broken after 10 seconds playing. The video is fine though. I have a 333MHz Celeron with MMX, use X 3.3.6 from woody, and have a very shitty 4MB PCI graphic card. Don't worry about your processor, it's fast enough for an 810kbps DivX. Here's what you should do: * Upgrade to XFree 4 This is the most important item: get Xv support! * Check that your sound card is not causing you trouble. I have a P3-600 (with no Xv support unfortunately), but even when the CPU load is 20% the video is choppy. If I use the '-nosound' option the choppyness disappears. * If the above is not enough upgrade your graphics card. Choose a graphics card that hits it well in Xv mode with XFree, i.e. the graphics card can do YUV transformation and scaling and whatever else, and all this is supported by XFree. I have no idea which ones do. * Finally upgrade your processor if all else failed but you probably won't get to that unless you play videos with much higher bit rates. Here's some numbers to 'prove' my claims. A friend of mine introduced me to mplayer. Our test DivX is a 729kbps 592x288 24bpp 25.01 fps movie. On his box, a K6-2 233MHz with an ATI Rage, it plays in fullscreen (-fs) (1280x1024x24) with slightly less than 50% CPU utilization (35% mplayer, 10% XFree). Scaling had almost no effect on the CPU usage. On my PII 333MHz with a Permedia2 graphics card (1600x1200x16) and XFree 3.3.6 (no Xv) it would not play without skipping regularly. With XFree 4 in x11/shm mode it's better but CPU gives: 50% mplayer and 30% XFree. Now, with xv it's much better: 45-50% mplayer and 20% XFree. And with -xy 2 scaling it's just XFree that takes more CPU, about 35%. So eventhough I have a faster CPU I just cannot play movies like on his box. Probably it's possible to make up for the graphics card by going to a much faster processor but it does not seem like a good solution. (If you have an ATI card, look for project Gatos, I'm told they have XFree drivers for all ATI graphics cards) -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ In theory, theory and practice are the same, but in practice they're different.
Re: logcheck gaps in time
On Mon, 4 Jun 2001, Alvin Oga wrote: hi ya jiji - rebooted which should take care of cron / syslogd / logcheck.sh not running - apt-get --reinstall install logcheck just in case - hacker? that is my fear. How can I find evidence that the @#$@ is in there? You may try the chkrootkit package. Something troubles me. IIRC you said that you received one mail per hour whether there was something to report or not. This is not what I get here: I get emails only when there's something to report. Did you change something to the default configuration? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ La terre est une bêta...
Re: 486 router is very slow
On Mon, 4 Jun 2001, Raffaele Sandrini wrote: Hi all I set up a router with my old 486 computer. I have there potato witch kernel 2.4 installed. Normally i have have rates of 7.8 kb/s (ISDN) throughput. But if i try it over my new router the rate is not more than 2 kb/s. Is it possible that the Masquerading process takes too much recources on that machine? It's a 33 Mhz Intel DX Processor. Any hints to speed that up? Just a few random checks that come to mind: * what is the chip controlling the serial port to the ISDN modem? Is it an 8250 or a 16650? I seem to remember that the 8250 were not very good but I don't remember when they stopped using them. * check the serial port configuration. What is its baud rate? * what is the CPU usage? To see if it's really CPU bound? Do the system/user values look odd? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Utilisateur (nom commun) : Mot utilisé par les informaticiens en lieu et place d'idiot.
Re: Snort config
On Fri, 18 May 2001, Iwan Mouwen wrote: * John Galt [EMAIL PROTECTED] [010516 15:24]: DEBIAN_SNORT_STATS_RCPT=root ^ Change this just on principle: using root to check system email is just another thing you can do as a user and not have to be logged in as root so much... Why? # vi /etc/aliases root: yournamehere Agreed. IMHO, this is the right way to do things. Send all maintainance and administrative email to root and set /etc/(sendmail/)aliases accordingly. Then if the administrator changes from foo to bar, you only have one file to change. The exception would be if you have multiple administrators for the same machine, each responsible for a subset of the functions. Then I would send email to account-admin, web-admin, sys-admin, etc, and, again, point these to the appropriate physical person in the aliases file. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ RFC 2549: ftp://ftp.isi.edu/in-notes/rfc2549.txt IP over Avian Carriers with Quality of Service
Re: galeon - skipstone?
On Thu, 17 May 2001, Tommi Komulainen wrote: [...] Hmm, now how on earth can I get all my cookies and bookmarks shared among all (mozilla, galeon, skipstone, konqueror, w3m) browsers..? Import/Export is really not an option anymore :-/ I can't help for cookies. But for bookmarks, I long ago decided to write an html file with all by bookmarks and use it as my home page. At the time I did it to share them between IE and Netscape but the principle is universal. It also lets me move my bookmarks from one machine to another just by copying a file. It's just a bit harder to add a new bookmark. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Stolen from an Internet user: f u cn rd ths, u cn gt a gd jb n cmptr prgrmmng !
net-pf-10 modprobes in /var/log/ksymoops/???
I keep getting these reports in '/var/log/ksymoops/*.log'. 20010504 02 start /sbin/modprobe -s -k net-pf-10 safemode=1 20010504 02 probe ended 20010504 000203 start /sbin/modprobe -s -k net-pf-10 safemode=1 20010504 000203 probe ended 20010504 000712 start /sbin/modprobe -s -k net-pf-10 safemode=1 20010504 000712 probe ended 20010504 001002 start /sbin/modprobe -s -k net-pf-10 safemode=1 20010504 001002 probe ended ... What's up with this? It seems I get these messages whether have 'alias net-pf-10 off' or not (my two computers get these messages one has the above line commented out and the other not). -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ The greatest programming project of all took six days; on the seventh day the programmer rested. We've been trying to debug the *^%$#@ thing ever since. Resume: design before you implement.
Dead links in sendmail package
I'm using sendmail 8.11.3+8.12.0.Beta7-3 from testing and cruft is complaining about the following dead links: /usr/lib/sm.bin/deliver - ../../bin/deliver /usr/lib/sm.bin/maildrop - ../../bin/maildrop /usr/lib/sm.bin/vacation - ../../bin/vacation Now, the destination of these links is provided by the 'deliver' package which is not installed on my system (sendmail only suggests it). So I have the following questions: * is it acceptable for a package to create dead links * should sendmail be creating these dead links? If not then I'll submit a bug report against sendmail (why does it need them btw?) * or should the cruft be updated to not complain about these links? In which case I'll submit an explanation file for inclusion in cruft (hmm, I'm not sure you can explain dead links with an explanation file actually). Btw, cruft also complains that the following files are 'unexplained: /'. I guess it needs a '/usr/lib/cruft/filters/sendmail' file. /var/lib/sendmail/dead.letter /var/lib/sendmail/host_status /var/run/sendmail /var/run/sendmail/sendmail.pid /var/run/sendmail/smcontrol -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ We are Pentium of Borg. You will be approximated. Division is futile.
Logcheck enhancement needs a volunteer
#$LOGTAIL $LOGDIR/auth.log $TMPDIR/check.$$
#$LOGTAIL $LOGDIR/daemon.log $TMPDIR/check.$$
#$LOGTAIL $LOGDIR/kern.log $TMPDIR/check.$$
#$LOGTAIL $LOGDIR/lpr.log $TMPDIR/check.$$
#$LOGTAIL $LOGDIR/mail.log $TMPDIR/check.$$
#$LOGTAIL $LOGDIR/syslog.log $TMPDIR/check.$$
#$LOGTAIL $LOGDIR/user.log $TMPDIR/check.$$
#fi
#
# END CONFIGURATION SECTION. YOU SHOULDN'T HAVE TO EDIT ANYTHING
# BELOW THIS LINE.
function make_pattern_file()
{
pat_file=$1
pat_dir=$2
if [ -f $pat_file ]; then
cat $pat_file
fi
if [ -d $pat_dir -a `echo $pat_dir/*` != $pat_dir/* ]; then
cat $pat/*
fi
}
function filter_log()
{
src=$1
dst=$2
match=$3
ignore=$4
if [ ! -s $src ]; then
return 0;
fi
if [ -n $match -a -s $match ]; then
if [ -n $ignore -a -s $ignore ]; then
$GREP -i -f $match $src | $GREP -v -f $ignore $dst
else
$GREP -i -f $match $src $dst
fi
elif [ -n $ignore -a -s $ignore]; then
$GREP -v -f $ignore $src $dst
else
ln -s $src $dst
fi
return test -s $dst
}
# Set the flag variables
FOUND=0
ATTACK=0
# See if the tmp file exists and actually has data to check,
# if it doesn't we should erase it and exit as our job is done.
if [ ! -s $TMPDIR/check.$$ ]; then
rm -f $TMPDIR/check.$$
exit 0
fi
# Perform Searches
# first sort the logs to remove duplicate lines (from different logfiles with
# the same lines) and reduce CPU and memory usage afterwards.
sort -u $TMPDIR/check.$$ $TMPDIR/check-sorted.$$
mv $TMPDIR/check-sorted.$$ $TMPDIR/check.$$
# Add an identification line at the beginning of the sent mail
# Rene Mayrhofer, 2000-11-24, text by Rainer Dorsch
echo This mail is sent by logcheck. If you do not want to receive it any
more, \
$TMPDIR/checkreport.$$
echo please modify the configuration files in /etc/logcheck or deinstall
logcheck. \
$TMPDIR/checkreport.$$
# Check for blatant hacking attempts
make_pattern_file $HACKING_FILE $HACKING_DIR $TMPDIR/pat_hack.$$
if filter_log $TMPDIR/check.$$ $TMPDIR/checkoutput.$$
$TMPDIR/pat_hack.$$; then
echo $TMPDIR/checkreport.$$
echo Active System Attack Alerts $TMPDIR/checkreport.$$
echo =-=-=-=-=-=-=-=-=-=-=-=-=-= $TMPDIR/checkreport.$$
cat $TMPDIR/checkoutput.$$ $TMPDIR/checkreport.$$
FOUND=1
ATTACK=1
fi
# Check for security violations
make_pattern_file $VIOLATIONS_FILE $VIOLATIONS_DIR
$TMPDIR/pat_violations.$$
make_pattern_file $VIOLATIONS_IGNORE_FILE $VIOLATIONS_IGNORE_DIR
$TMPDIR/pat_violations_ignore.$$
if filter_log $TMPDIR/check.$$ $TMPDIR/checkoutput.$$
$TMPDIR/pat_violations.$$ $TMPDIR/pat_violations_ignore.$$; then
echo $TMPDIR/checkreport.$$
echo Security Violations $TMPDIR/checkreport.$$
echo =-=-=-=-=-=-=-=-=-= $TMPDIR/checkreport.$$
cat $TMPDIR/checkoutput.$$ $TMPDIR/checkreport.$$
FOUND=1
fi
# Do reverse grep on patterns we want to ignore
make_pattern_file $IGNORE_FILE $IGNORE_DIR $TMPDIR/pat_ignore.$$
if filter_log $TMPDIR/check.$$ $TMPDIR/checkoutput.$$
$TMPDIR/pat_ignore.$$; then
echo $TMPDIR/checkreport.$$
echo Unusual System Events $TMPDIR/checkreport.$$
echo =-=-=-=-=-=-=-=-=-=-= $TMPDIR/checkreport.$$
cat $TMPDIR/checkoutput.$$ $TMPDIR/checkreport.$$
FOUND=1
fi
# If there are results, mail them to sysadmin
if [ $ATTACK -eq 1 ]; then
cat $TMPDIR/checkreport.$$ | $MAIL -s $HOSTNAME $DATE ACTIVE SYSTEM
ATTACK! $SYSADMIN
elif [ $FOUND -eq 1 ]; then
cat $TMPDIR/checkreport.$$ | $MAIL -s $HOSTNAME $DATE system check
$SYSADMIN
fi
# Clean Up
rm -f $TMPDIR/check.$$ $TMPDIR/check-sorted.$$ $TMPDIR/checkoutput.$$ \
$TMPDIR/checkoutput.tmp.$$ $TMPDIR/checkreport.$$ \
$TMPDIR/pat_*
--- cut here ---
--
Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/
Dieu dit: M-x Lumière. Et la lumière fut.
Re: libwine will not upgrade (unstable)
On 1 Mar 2001, Pollywog wrote: Did anyone else have trouble upgrading WINE today? libwine gives the following error on my machine: Preconfiguring packages .. /var/lib/debconf/config.175203: null: command not found libwine failed to configure, with exit code 127 (Reading database ... E: Sub-process /usr/bin/dpkg exited unexpectedly I forwarded your question to Ove and here's his answer: On Fri, 2 Mar 2001, Ove Kaaven wrote: No. See my changelog; I moved the config out of libwine - there simply *is* no config control file in libwine anymore. It seems maybe apt have problems upgrading if config is removed from a package? Removing and reinstalling libwine might help, but the user should probably file a bug against apt. Did you try that and did it solve the problem? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Lotto: A tax on people who are bad at math. -- unknown Windows: Microsoft's tax on computer illiterates. -- WE7U
ARP problems
From time to time my network dies. Most of the time I have two machines on it and the symptom is that they cannot ping each other anymore. At the same time I see that all that a ping triggers is an arp packet. Here's a typical tcpdump trace: 11:16:52.909161 amboise 192.168.1.255: icmp: echo request 11:16:53.936272 amboise 192.168.1.255: icmp: echo request 11:16:54.936322 amboise 192.168.1.255: icmp: echo request 11:16:59.247968 arp who-has oleron tell amboise 11:17:00.246100 arp who-has oleron tell amboise 11:17:01.246096 arp who-has oleron tell amboise 11:17:02.266239 arp who-has oleron tell amboise 11:17:03.266105 arp who-has oleron tell amboise 11:17:04.266085 arp who-has oleron tell amboise 11:17:05.286220 arp who-has oleron tell amboise 11:17:06.286096 arp who-has oleron tell amboise 11:17:07.286088 arp who-has oleron tell amboise 11:17:08.306238 arp who-has oleron tell amboise 11:17:09.306087 arp who-has oleron tell amboise 11:17:10.306080 arp who-has oleron tell amboise 11:17:23.592631 arp who-has oleron tell amboise 11:17:23.592900 arp reply oleron is-at 0:a0:24:ff:e5:ab 11:17:23.592935 amboise oleron: icmp: echo request 11:17:23.593267 oleron amboise: icmp: echo reply 11:17:24.646275 amboise oleron: icmp: echo request 11:17:24.646652 oleron amboise: icmp: echo reply There must be something wrong on my desktop because: - the only way I have found to fix this is to unplug its RJ45 ethernet cable from the switch and replug it. - other machines can still ping each other (when there's actually more than one 'other machine' on the network) * desktop kernel 2.2.17 3c59x it's also the DHCP server * switch In the middle I have a small Unicom MiniSwitch/5. * laptop kernel 2.4.0test9 / windows 98, it happens with both currently with a 3c589_cs but I had the same problem with a linksys PCMPC100 (now dead, broken dongle) Does anyone know what could be wrong? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ War doesn't determine who's right. War determines who's left.
Re: KDE2 install
On Sun, 25 Feb 2001, Daniel Jones wrote: [...] And, as has already been pointed out, get security.debian.org added as well. Yes, except that the given line was wrong and that, AFAIK, there is no security updates for testing anyway! -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Any sufficiently advanced Operating System is indistinguishable from Linux
console-data or debconf bug?
I got the following message when upgrading to today's testing: Setting up console-data (1999.08.29-14) ... Can't locate Debconf/Client/ConfModule.pm in @INC (@INC contains: /usr/lib/perl5/5.005/i386-linux /usr/lib/perl5/5.005 /usr/local/lib/site_perl/i386-linux /usr/local/lib/site_perl /usr/lib/perl5 .) at /usr/share/console/getkmapchoice.pl line 3. BEGIN failed--compilation aborted at /usr/share/console/getkmapchoice.pl line 3. Is this a bug in console-data or debconf or neither? console-data: 1999.08.29-14 debconf: 0.2.80.17 -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Dieu dit: M-x Lumière. Et la lumière fut.
Re: console-data or debconf bug?
On Sat, 10 Feb 2001, Francois Gouget wrote: I got the following message when upgrading to today's testing: Setting up console-data (1999.08.29-14) ... Can't locate Debconf/Client/ConfModule.pm in @INC (@INC contains: /usr/lib/perl5/5.005/i386-linux /usr/lib/perl5/5.005 /usr/local/lib/site_perl/i386-linux /usr/local/lib/site_perl /usr/lib/perl5 .) at /usr/share/console/getkmapchoice.pl line 3. BEGIN failed--compilation aborted at /usr/share/console/getkmapchoice.pl line 3. Is this a bug in console-data or debconf or neither? console-data: 1999.08.29-14 debconf: 0.2.80.17 Hmm, just noticed, add console-tools to the list: Setting up console-tools (0.2.3-16) ... Can't locate Debconf/Client/ConfModule.pm in @INC (@INC contains: /usr/lib/perl5/5.005/i386-linux /usr/lib/perl5/5.005 /usr/local/lib/site_perl/i386-linux /usr/local/lib/site_perl /usr/lib/perl5 .) at /usr/share/console/getkmapchoice.pl line 3. BEGIN failed--compilation aborted at /usr/share/console/getkmapchoice.pl line 3. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Any sufficiently advanced bug is indistinguishable from a feature. -- from some indian guy
Re: console-data or debconf bug?
On Sat, 10 Feb 2001, Chris Matta wrote: This is a perl package issue, you can clear it up by hand installing perl5.6 and perl-5.6-base. So console-data and console-tools have a packaging bug in that they should depend on perl-5.6. I want for proof that perl-5.6 is not available in testing. If the dependencies of console-data and console-tools had been correct then they would never have made it to testing. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ The nice thing about meditation is that it makes doing nothing quite respectable -- Paul Dean
Re: console-data or debconf bug?
On Sat, 10 Feb 2001, Joey Hess wrote: Francois Gouget wrote: Is this a bug in console-data or debconf or neither? It's a known bug in console data, it depends on a version of debconf that is not in testing. (One of the types of things that testing's automation, sadly, cannot detect.) Ok. I assume that since it's a known bug I don't need to make a bug report. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ RFC 2549: ftp://ftp.isi.edu/in-notes/rfc2549.txt IP over Avian Carriers with Quality of Service
What does it mean: Protocol-Reject for unsupported protocol
Hi, I get a lot of these messages in my ppp.log file (more than two dozens per day, no particular timing pattern): Jan 8 18:27:48 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x2069 Jan 8 18:27:48 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x2020 Jan 8 18:43:08 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x2020 Jan 8 18:50:04 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x5441 Jan 9 00:36:03 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x666f Jan 9 00:36:35 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x666f Jan 9 10:49:37 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x983d Jan 9 15:13:45 amboise pppd[712]: Protocol-Reject for unsupported protocol 0x9f36 I have a DSL and I'm using PPPoE. Can anyone tell me what the above messages mean? And why do I'm getting them? Is my configuration wrong or is it someone trying to probe my computer in some way? Or yet something else? pppoe: 1.7-1 pppd: 2.4.0f-1 -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Nouvelle version : les anciens bogues ont ?t? remplac?s par de nouveaux.
Re: Some refreshing news
On Sat, 6 Jan 2001, Gary Jones wrote: On 5 Jan 2001, you wrote: [EMAIL PROTECTED] writes: upgraded my motherboard and CPU last night (from a Pentium 75 to a Celeron 500). The system is a dual boot system (Debian Linux, and Windows 95). I have been unable to boot into Windows (lots of errors, graphics look horrible, and finally kills me with a fatal error), but I have had no problems with getting back into Linux! I will have to reinstall Windows to see if I can get it working again (if they had kids learning games for Linux I would get rid of Windows entirely, but until then I am stuck with it) Wine? vmware? Or Win4lin, much cheaper than vmware and should probably do the job too (warning I use neither). Or plex86 although it is probably not quite ready yet. Of course my personal preference would be Wine but I'm biased since I'm a Wine developper. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ May your Tongue stick to the Roof of your Mouth with the Force of a Thousand Caramels.
Ad blocking with Squid
Is there a Debian package that would configure Squid to do ad blocking? Background: - I'm running squid at home to provide a unified cache to my local network. - I don't want to run Junkbuster because AFAIK it's not a cache. - I don't want to run both Junkbuster and Squid because I would see this as wasteful. I have searched on the web and I found the following two sites that provide such a configuration for Squid: http://www.taz.net.au/block/ http://www.zip.com.au/~cs/adzap/index.html Does anyone have experience with one of these? Is there a better resource? Now I believe it would be great to have a squid-adzap/squid-adblock package that you could just install to block banner ads. I might (note: *might*) package one of the above myself if I ever find the time but I don't want to reinvent the wheel. If no such package exists, would anyone be interested in such a package? Even better, interested enough to do it? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Dieu dit: M-x Lumi?re. Et la lumi?re fut.
Dependency problems in testing
I'm now following testing and I had two dependency problems already: * the first one is between libc6 and libnss which is required by the distributed-net package: libc6:2.2-6 libnss1-compat: 2.1.3-13 depends libc6 (= 2.1.3-13) distributed-net: 2.7106-7.1 depends libnss1-compat I guess the reasoning is that non-free software should not hold back free software and who cares about non-free software anyway :-/. Well, I uninstalled distribued-net, I run my prime-net package anyway :-) Still, what is libnss1-compat still doing in testing? One cannot install it anymore since there is no libc6 version 2.1.3 in testing (by extension the same goes for distributed-net). * the other problem leaves me completely lost: I want to try Window Maker, and while I'm at it, it would seem to make sense to also install wmakerconf. But it looks like I cannot do that: wmaker: 0.61.1-4 wmakerconf: 2.4-4 depends wmakerconf-data = 0.61.1 wmakerconf-data: 0.62.0-1 conflicts wmaker 1.0, wmaker 0.62.0 So, can anyone tell me how to install wmakerconf from testing? What is the ifupdown package? On my system ifup and ifdown are part of netbase. Or does this indicate a soon to come split of the netbase package? Also when upgrading locales I did what they said: Ctrl-Z, edit /etc/... and then fg. That really did not work well, after the fg the keyboard input was no longer working. Only Ctrl-C worked (badly of course) and then the display had a case of staircase :-(((. A logout, re-dselect, re-install solved it fortunately (shows the installation process is quite resilient, that's good). -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Advice is what we ask for when we already know the answer but wish we didn't -- Eric Jong
Re: Difference Dependency-Recommendation
On 28 Dec 2000, John Hasler wrote: Holger writes: I don't quite understand the difference between Dependency and Recommendation, two terms that commonly occur when talking about Debian's packet management system. 'Depends' means that the package will not work without it. 'Recommends' means that the package maintainer recommends it, but the package will work without it. I think both have the same result: the packet (usually, unless explicitly overridden) gets installed. No. dpkg (which is what actually installs packages regardless of the front-end) will not install packages with unsatisfied dependencies unless you use '--force-depends'. It ignores 'Recommends' (as does apt). dselect gets rather insistent about 'Recommends', but its insistence can be overridden. I have a related question: what is the difference between Recommends and Suggests? IIRC, dselect insists strenuously that you should install 'Suggested' packages. If you de-select the suggested package it will pop-up the conflict resolution window, if you override it it will do so again once you exit select mode, ... From where I stand it looks like 'suggest' is equivalent to 'depends'. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ We are Pentium of Borg. You will be approximated. Division is futile.
kapmd *not* taking all the CPU
I am running the kernel 2.4.0-test9 on my laptop and normally kapmd takes all the CPU when the laptop is idle. But I have noticed that after some use (1-3 days uptime doing compiles + debugging) kapmd no longer takes all the CPU even when the laptop is idle: kapmd: 37-38% idle: 61-62% top: 0.3% Sometimes it even completely stops using the CPU (kapmd usage drops to 0%). I hate it because this causes the fan to start and stop, start and stop, ... Grrr. I have a Vaio F560. Does anyone know what's going on. Is it a known bug? Should I update to a more recent 2.4 kernel? Other info: the laptop is closed and I'm using it remotely from my desktop via X so there is continuously some network use. Everything else is compilations and debugging. I'm running 'testing', and noflushd. The kernel APM config is: CONFIG_PM=y CONFIG_ACPI=y CONFIG_APM_DO_ENABLE=y CONFIG_APM_CPU_IDLE=y CONFIG_APM_DISPLAY_BLANK=y -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Linux: the choice of a GNU generation
Re: es1370 sound stopped working
On Thu, 14 Dec 2000, Ron Hale-Evans wrote: Greetings, all-- I am running Potato with a SoundBlaster PCI64 sound card, which uses the es1370 kernel module. I installed the soundcore and es1370 modules a couple of days ago, did a chmod a+rw to all /dev devices in group 'audio', and everything worked fine. This morning I was using the Windows program Zillions of Games under Wine 20001202 (latest snapshot). I have had similar problems with this module for my Ensoniq Soundcard. It seems to happen when a program that uses the sound does not exit cleanly. The (bad) solution I use is rmmod es1370;insmod es1370. Normally this works. I believe the right solution would be to try and debug this module. Or maybe first check if it has changed in 2.4. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Linux: Because rebooting is for adding new hardware
Re: Keyboard randomly generates a `
On Sun, 10 Dec 2000, Randy Edwards wrote: I've seen the freezing on more than one keyboard that subsequently became completely useless. I'd consider a new keyboard based on that alone. My keyboard has been doing this ever since I got it, about two years ago. So I doubt it will break down any time soon. I've seen this behavior on many keyboards: MS Natural clones, Dell keyboards, random cheapo-type keyboards -- a wide variety. It's nice to see that I'm not the only one with this problem. Have you ever observed this on Windows or only Linux? Do you have instances where a keyboard does it while another one of the same brand and model does not?` ^ here it goes again! [...] Anyone know what is causing the ` to be generated (bash really likes those chars tossed in here or there!:-). Yes. gcc and perl are very picky too. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Linux: Because rebooting is for adding new hardware
Keyboard randomly generates a `
Hi, This is most likely not very Debian specific but here goes. From time to time, I'd say at least once a day, my keyboard just generates a '`'. For instance I would be scrolling in a C file in Emacs and suddenly a '`' appears. Of course I'm absolutely certain that it's not me accidentally hitting this key. I could not find any pattern: it happens in any program, whether the keyboard has been 'idle' for a long time before or not, but it only happens when I hit a key. I don't remember it happening in the console but I don't spend enough time using it to rule it out. Also I know it does not happen in NT 4 and Win2000 (unless the problem is with my keyboard, not the model). The keyboard is the standard keyboard that came with a Gateway E-3110 (it says model 2196003. You can check it out here, http://www.asahi-net.or.jp/~RJ3M-AOK/keyboard/2196003-XX-XXX.html, except mine is a qwerty/US version. My XF86Config 3.3.6 file says: Section Keyboard ProtocolStandard XkbModelpc104 XkbModelmicrosoft EndSection (everything else is commented out) So if anyone knows what could be wrong with it let me know. The second feature of this keyboard is that sometimes it freezes. And this time I know it's the model because it also happens on NT4 and Win2000 machines. Doing a cold keyboard reboot (unpluging and replugging it) plus a kbdrate fixes it. May be related? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Only wimps use tape backup: _real_ men just upload their important stuff on ftp, and let the rest of the world mirror it ;) -- Linus Torvalds
smbfs: inaccessible, un-unmountable directory
I had a Windows fs mounted on ~/smb via smbfs. The Windows machine got put in suspend mode, then woken up, but now I canot access the directory anymore: $ ls smb ls: smb: Input/output error $ smbumount smb Could not open smb: Input/output error When I do 'ls ~' I don't even see smb anymore. So what should I do? Surely I must have some other way of getting out of this situation than rebooting (rmmod smbfs does not work). Kernel: Linux version 2.2.17 smbfs: 2.0.7-3 -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Utilisateur (nom commun) : Mot utilisé par les informaticiens en lieu et place d'idiot.
Re: smbfs: inaccessible, un-unmountable directory
On Sun, 3 Dec 2000, Nate Amsden wrote: umount and remount .. It works! Why the hell is it that smbumount does not work while umount works just fine? smbfs is VERY unstable i would never trust it to anything ... I don't trust it very much either. I mostly use it to transfer files from Windows machines to my Linux box. I'm (a bit more) disgusted by smbfs but very grateful for your help. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ War doesn't determine who's right. War determines who's left.
Re: hdparm to increase performance
On Fri, 10 Nov 2000, Chris Majewski wrote: [...] What about when I start netscape? I hear the disk crunching, so presumably the binary isn't cached. Shouldn't increasing the disk throughput speed up the time it takes to start netscape? I don't know. They must be doing something wrong ;-) If it doesn't then it must be that the accesses are dominated by seeks. Fragmentation? This should not happen with ext2fs. Maybe random disk accesses (for checking the cache?) Here once it's in the cache it starts in 3.4s (I don't know how it is when it's not in the cache :-). -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Dieu dit: M-x Lumière. Et la lumière fut.
Re: hdparm to increase performance
On 10 Nov 2000, Krzys Majewski wrote: Michael P. Soulier [EMAIL PROTECTED] writes: In case you guys missed this one, check it out. http://www.oreillynet.com/pub/a/272 I just increased my harddrive throughput by 5 times. Mike Do you notice a difference though? I increased mine from about 3Mbps to about 18Mbps, but I haven't felt it yet. -chris Linux does not swap enough. Remove some RAM :-) More seriously, I guess that if Linux has lots of RAM to play with (here swap in use 10% RAM), and a big cache (30-50% RAM), you're not going to saturate your disk bandwidth in the first place. So increasing it by a factor of 6 is not going to help since it was not the bottleneck in the first place. Also under such conditions whether reads or writes are slow is not really important since reads are read ahead, as long as the data arrives before you need it it's fine, and writes go to the cache and don't have to be flushed all that often. Finally the only cases where disk access will have an influence is when the read-ahead occasionally fails. In that case It's most likely going to be a latency issue, i.e. what's your seek time. And I believe hdparm cannot help enhance this aspect. In such a situation, a more interesting aspect of hdparm is to reduce the CPU utilization. That way gcc and others can continue to work while you're accessing the disk. I've seen the usage drop from 75-90% while doing a 'dd if=/dev/null of=foo', which made the machine very unusable, to less than 5%, which made the machine usable again. Of course it depends on your particular situation and usage patterns. Maybe you stream tons of data, copy huge files (50MB+), etc. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Linux: the choice of a GNU generation
Re: To setup display for NeoMagic 256XL on Sony VAIO PCG-XG29
Hi, On Tue, 7 Nov 2000, jimmy sandhar wrote: Hi all, Does anybody know or have done this before, setting up display for Sony VAIO PCG-XG29. Please let me know if so. You should have a look at the Linux Laptops site. There's documentation about the Sony XG28: Linux Laptops: http://www.cs.utexas.edu/users/kharker/linux-laptop/ http://www.bluecoat93.org/tech/linux-xg28.html ^H^H^H scrath the above, the guy did not configure his graphics card correctly. A search on Google gave me: http://www.de-srv.com/linux/sonyF450.html This page has good instructions. Since I have a Sony with the same graphics card so you can ask me if you have trouble. I can also give you my SVGATextMode config: I do 80x60x9. rant Isn't it time that we have a good database on how to configure hardware for Linux! The F450 documentation partly duplicates that of the F490, and the XG28 got things wrong. A laptop/desktop is nothing but a collection of hardware components, so just point to the proper page for that component. /rant -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ In a world without fences who needs Gates?
Re: whence ping -w?
On Thu, 2 Nov 2000, Will Day wrote: A short time ago, at a computer terminal far, far away, Carel Fellinger wrote: Can anyone tell me where to find this ping command that offers a -w option (or any other option to set the timeout)? Or maybe I'm just crazy ping -i some-number No, that's the delay between multiple packets, not the timeout for each packet. Yes (ping, most implementations of it I guess) does not have a 'timeout' (if they're at all similar to the one I studied (bing actually)). Note that the following will not really help you at all, it's just FYI. IIRC, most ping implementations send an 8 byte timestamp obtained with gettimeofday as the payload of their ICMP echo request packets. This payload is copied in the ICMP echo reply packets. So wwhen ping receives an ICMP echo reply message that is addressed to itself (matching icmp id), it just does another call to gettimeofday, substract the timestamp that was returned, and print the RTT. This way ping does not need to keep track of all the packets it sent and thus does not need a timeout. It can even handle packets that come back out of order. I was really confused by ping's code until I realized that. Then I thought it was pretty cool. -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ In theory, theory and practice are the same, but in practice they're different.
dwww rebuilding the menus daily!
I'm using tripwire and every day it would tell me that all the menus of my window managers have been modified. I suspected a cron script and finally found some time to track it down. It turns out that /etc/cron.daily/dwww is the guilty party. More precisely it, invokes /usr/sbin/dwww-build which in turn invokes /usr/bin/update-menus: if [ -x /usr/bin/update-menus ]; then echo -n g) document index (by menu package)... update-menus echo done fi What is this document index (by menu package) and why does it implies rebuilding the menus each day? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Avoid the Gates of Hell - use Linux.
X errors while running make xconfig
Hi, When I'm running 'make xconfig' to configure a new kernel, I regularly get the following error: X Error of failed request: BadLength (poly request too large or internal Xlib length error) Major opcode of failed request: 49 (X_ListFonts) Serial number of failed request: 4170 Current serial number in output stream: 5753 Does anyone know what could be causing it? Note that it is not systematic. It's just very annoying to loose all the changes when you're almost finished. I'm running 'make xconfig' on a laptop and the display goes to my desktop, so it goes through my local ethernet network. This ethernet network has an MTU of 1460 (because of a bad PPP/masquerading interaction). I have: Debian 2.2 kernel 2.2.17 XFree86 3.3.6 -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Computers are like airconditioners They stop working properly if you open WINDOWS
X errors while running make xconfig (fwd)
Ok, I'm retrying to post this message because it seems it did not make it through the first time (and I did not get any email indicating there was an error). -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Nouvelle version : les anciens bogues ont été remplacés par de nouveaux. -- Forwarded message -- Date: Thu, 12 Oct 2000 12:52:35 -0700 (PDT) From: Francois Gouget [EMAIL PROTECTED] To: debian-user@lists.debian.org Subject: X errors while running make xconfig Hi, When I'm running 'make xconfig' to configure a new kernel, I regularly get the following error: X Error of failed request: BadLength (poly request too large or internal Xlib length error) Major opcode of failed request: 49 (X_ListFonts) Serial number of failed request: 4170 Current serial number in output stream: 5753 Does anyone know what could be causing it? Note that it is not systematic. It's just very annoying to loose all the changes when you're almost finished. I'm running 'make xconfig' on a laptop and the display goes to my desktop, so it goes through my local ethernet network. This ethernet network has an MTU of 1460 (because of a bad PPP/masquerading interaction). I have: Debian 2.2 kernel 2.2.17 XFree86 3.3.6 -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ Computers are like airconditioners They stop working properly if you open WINDOWS
Re: X errors while running make xconfig (fwd)
On Thu, 12 Oct 2000, Eric G . Miller wrote: Try menuconfig. Never had a problem with it... This does not solve the problem. It's merely a workaround. I 'make xconfig' is the only case where I have this problem. I commonly run xemacs across the network link and never had any problem. Does it mean there is a bug in Tcl/Tk? -- Francois Gouget [EMAIL PROTECTED]http://fgouget.free.fr/ War doesn't determine who's right. War determines who's left.
[Hamm Install] package dependency problems
Hi, I downloaded Hamm last night from fp.mi.us.debian.org and I got quite a lot of package download problems. Packages that seem to have moved between the ime I selected them and the time I downloaded them. Anyway, I have a lot of package problems, 58 packages are in the state iU, i.e. installed but not configured. I could at least identify the following problems: * the first problem has to do with perl. I have the package perl-base installed but apache which depends on 'perl' claims that perl is not installed. And in fact 'perl-base' does not provide 'perl'. But it 'Replaces: perl', 'Suggests: perl' and 'Conflicts: perl ( 5.004.04-2)'. I tried to find another package that would provide 'perl' but did not find any. To me it looks like there is a screw-up in the package dependencies/information. perl-base5.004.04-6 apache 1.3b6-3 * ddd-smotif depends on 'libg++27'. Fair enough. I installed the package 'libg++27' (Opt oldlibs libg++27). But this package provides 'libg++' not the one I expected. I could not find anything that ould povide 'libg++27'. Packages depending on 'libg++27' staroffice3 3.1-8 rvplayer 1:5.0-2 ddd-dmotif 2.2.3-2 libg++2.8-dev2.90.29-0.3 ddd-smotif 2.2.3-2 circus 0.43-1 libg++27-altdev 2.7.2.1-14.4 netscape44.0-12 Related packages: libg++272-dev2.7.2.8-0.1 provides libg++-dev libg++2.82.90.29-0.3 provides libg++ libg++2.8-dev2.90.29-0.3 provides libg++-dev libg++27 2.7.2.1-14.4 provides libg++ libg++272-dbg2.7.2.8-0.1 provides libg++272 * I'm not sure apache should depend on perl. I know we have apache installed on a Solaris machine were there is no perl. Sometimes when doing a first install it feels like every other package requires some other package that you (I) have not yet installed (and don't really want to install but...). This is typically things like perl, perl extensions, tcl, tk... * These problems seem to have blocked the installation of tcl8.0, tk8.0 which I could 'dpkg --configure' manually without problem. Why ? -- Francois Gouget [EMAIL PROTECTED] -- http://www.mygale.org/~fgouget/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: [Hamm Install] package dependency problems
On Wednesday, June 10, 1998 3:49 PM, Johnie Ingram [SMTP:[EMAIL PROTECTED] wrote: Francois == Francois Gouget?= us-ascii writes: Francois 5.004.04-2)'. I tried to find another package Francois that would provide 'perl' but did not find Perl itself provides perl: ./hamm/hamm/binary-i386/interpreters/perl_5.004.04-6.deb I don't know what happened on my system. I kept doing dselect/install. Sometimes it would say that some files had to be downloaded sometimes it would not. Finally it decided to download perl_base and some other stuff again (like gcc !). I had some FTP_ERRORs before but since doing dselect/install would not download anything more I assumed that the package was complete (and it seemed complete). Anyway with the last download batch it's much better although I still don't understand how the dependencies work. If I look in /var/lib/available I don't see anything that provides perl: # grep Provides: perl available Provides: perl-curses If I look at the package perl it provides 'io' (a rather bad name imho), not perl. perl-base replaces perl which I guess means it provides it (once the install worked out ok). Or maybe the package perl provides perl because it is called perl then. Francois 'libg++' not the one I expected. I could not Francois find anything that ould povide 'libg++27'. Sounds like a release-critical bug in ddd-smotif. I don't know what happened here either. I still don't see anything providing 'libg++27' but ddd-smotif is installed. Same thing here I guess that the package libg++27 provides libg++27 because that's its name. Why didn't it work before then ? Well anyway... Thanks, I only have 6 broken packages left, its getting there... -- Francois Gouget [EMAIL PROTECTED] -- http://www.mygale.org/~fgouget/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[Hamm Install] Dselect/ftp, what is this error message ?
I'm installing a new Debian 2.0 on my PC using the ftp method to download it from ftp.mi.us.debian.org. During the download I got the following error message: getting: dists/frozen/main/binary-i386/interpreters/perl_5.004.04-6.deb (3125884) FTP ERROR Do you want to retry downloading at once [y]: continue file: dists/frozen/main/binary-i386/interpreters/perl_5.004.04-6.deb (at 1484820) [y]: [...] getting: dists/frozen/main/binary-i386/interpreters/perl_5.004.04-6.deb (1641064/3125884) As I understand it the first line tells me that I only got 1484820 bytes out of the 3125884. So why does the download restarts at byte 1641064 as the second line seems to indicate ? Then it failed again and I had: FTP ERROR Do you want to retry downloading at once [y]: continue file: dists/frozen/main/binary-i386/interpreters/perl_5.004.04-6.deb (at 2242560) [y]: [...] getting: dists/frozen/main/binary-i386/interpreters/perl_5.004.04-6.deb (883324/3125884) Did it restart from scratch ? I finally managed to download the entire file but it is only 262680 bytes long. So what are the numbers displayed above ? Does the ftp method support reget ? -- Francois Gouget [EMAIL PROTECTED] -- http://www.mygale.org/~fgouget/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Some scripts cause ksh to core dump
Hi,
I'm using ksh to run some scripts that generate java code. INormally I
would run these scripts on a Solaris (thus the ksh) but it's so much
faster on my Linux box... :-)
Anyway, today I got one script which is crashing ksh !!!
~/javaSDK/generate sdk SQLViewEntity
SQLViewEntity...
Memory fault (core dumped)
Now I don't publish this script here because I can't (it's
copyrighted), because when on it's own this script does not crash ksh,
it does when invoked by the other scripts, and because I could not
isolate the cause of the crash. It involves a for loop with an empty
list of parameters, accessing ${12}. Well, maybe I could send the
scripts to someone interested in debugging this though.
But basically it seems to execute some code inside a loop where it
should not go since there is nothing to iterate on (for i in ; do ;
done). I say that because by changing what's inside the loop I get it
not to crash anymore. Now I cannot pinpoint what I have to remove. T
here are many unrelated things I can remove that fix that (but my
script no longer does what it's supposed to).
So I downloaded the source code, compiled it, and run my script with
the compiled code and... it does not crash ! So there is nothing I can
debug unfortunately. Now the ksh I get (after stripping) is different
from what I have installed on my machine, although they are both from
the same package version (maybe the compiler is slightly different).
I feel the ksh in the distribution should be fixed. But how ?
I can send the core file to anyone who asks, all gdb tells me is:
Core was generated by `ksh ./450s.MIRXxx.ulinks.java'.
Program terminated with signal 11, Segmentation fault.
#0 0x400516c3 in ?? ()
A simple recompilation probably is not enough. I believe the core dump
is caused by writing or reading past the end of some array/string so
the bug might seem to go away with a recompilation but would come back.
I have:
- a recent hamm installation (less than 1 week old)
- ii pdksh 5.2.13-3 A public domain version of the
Korn shell
- pdksh_5.2.13.orig.tar.gz - the source code
- pdksh_5.2.13-3.diff.gz - the debian patches
--
Francois Gouget
[EMAIL PROTECTED] -- http://www.mygale.org/~fgouget/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
ksh core dumps
Hi,
I'm using ksh to run some scripts that generate java code. INormally I
would run these scripts on a Solaris (thus the ksh) but it's so much
faster on my Linux box... :-)
Anyway, today I got one script which is crashing ksh !!!
~/javaSDK/generate sdk SQLViewEntity
SQLViewEntity...
Memory fault (core dumped)
Now I don't publish this script here because I can't (it's
copyrighted), because when on it's own this script does not crash ksh
but it does when invoked by the other scripts, and because I could not
isolate the cause of the crash. It involves a for loop with an empty
list of parameters, accessing ${12}. Well, maybe I could send the
scripts to someone interested in debugging this though.
But basically it seems to execute some code inside a loop where it
should not go since there is nothing to iterate on (for i in ; do ;
done). I say that because by changing what's inside the loop I get it
not to crash anymore. Now I cannot pinpoint what I have to remove. T
here are many unrelated things I can remove that fix that (but my
script no longer does what it's supposed to).
So I downloaded the source code, compiled it, and run my script with
the compiled code and... it does not crash ! So there is nothing I can
debug unfortunately. Now the ksh I get (after stripping) is different
from what I have installed on my machine, although they are both from
the same package version (maybe the compiler is slightly different).
I feel the ksh in the distribution should be fixed. But how ?
I can send the core file to anyone who asks, all gdb tells me is:
Core was generated by `ksh ./450s.MIRXxx.ulinks.java'.
Program terminated with signal 11, Segmentation fault.
#0 0x400516c3 in ?? ()
A simple recompilation probably is not enough. I believe the core dump
is caused by writing or reading past the end of some array/string so
the bug might seem to go away with a recompilation but would come back.
I have:
- a hamm installation
- ii pdksh 5.2.13-3 A public domain version of the
Korn shell
- pdksh_5.2.13.orig.tar.gz - the source code
- pdksh_5.2.13-3.diff.gz - the debian patches
--
Francois Gouget
[EMAIL PROTECTED] -- http://www.mygale.org/~fgouget/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
some scripts cause ksh to core dump
Hi,
I'm using ksh to run some scripts that generate java code. INormally I
would run these scripts on a Solaris (thus the ksh) but it's so much
faster on my Linux box... :-)
Anyway, today I got one script which is crashing ksh !!!
~/javaSDK/generate sdk SQLViewEntity
SQLViewEntity...
Memory fault (core dumped)
Now I don't publish this script here because I can't (it's
copyrighted), because when on it's own this script does not crash ksh,
it does when invoked by the other scripts, and because I could not
isolate the cause of the crash. It involves a for loop with an empty
list of parameters, accessing ${12}. Well, maybe I could send the
scripts to someone interested in debugging this though.
But basically it seems to execute some code inside a loop where it
should not go since there is nothing to iterate on (for i in ; do ;
done). I say that because by changing what's inside the loop I get it
not to crash anymore. Now I cannot pinpoint what I have to remove. T
here are many unrelated things I can remove that fix that (but my
script no longer does what it's supposed to).
So I downloaded the source code, compiled it, and run my script with
the compiled code and... it does not crash ! So there is nothing I can
debug unfortunately. Now the ksh I get (after stripping) is different
from what I have installed on my machine, although they are both from
the same package version (maybe the compiler is slightly different).
I feel the ksh in the distribution should be fixed. But how ?
I can send the core file to anyone who asks, all gdb tells me is:
Core was generated by `ksh ./450s.MIRXxx.ulinks.java'.
Program terminated with signal 11, Segmentation fault.
#0 0x400516c3 in ?? ()
A simple recompilation probably is not enough. I believe the core dump
is caused by writing or reading past the end of some array/string so
the bug might seem to go away with a recompilation but would come back.
I have:
- a recent hamm installation (less than 1 week old)
- ii pdksh 5.2.13-3 A public domain version of the
Korn shell
- pdksh_5.2.13.orig.tar.gz - the source code
- pdksh_5.2.13-3.diff.gz - the debian patches
--
Francois Gouget
[EMAIL PROTECTED] -- http://www.mygale.org/~fgouget/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Hamm installation problems (fdisk bug)
You're right. I used Win95's OSR2 fdisk tool. Now since my last message I have found a tool, partinfo, which is made by Power Quest, the guys that do Partition Magic. When I ran this tool on my partition table it told me this: C:WIN95FAT16B Pri,Boot 509.8 0 0 63 1044162 ExtendedXPri 8197.2 0 1 1044225 16787925 And on their Web site they talk about ExtendedX, Fat16X, Fat32X partition types as being partitions that can go past the 8GB limit. I think this is a Microsoft invention to denote partitions for which the CHS information should be ignored and the LBA used instead. But because this give different partition system ids Linux fdisk no longer recognises them as extended/fat16/fat32 partitions. The problem if I use old Dos's fdisk is that Win95 will believe that my extended partition only goes to cylinder 1023 when it goes up to 1109. I think I'll scrap what is on my hard disk for now (I'll have fun reinstalling it later) and install Linux. Unless someone knows of a version of fdisk that would support these new partition types. Anyone interested in this problem should probably have a look at the following Web pages: http://www.powerquest.com/downlwd/part.html - to download partinfo http://www.powerquest.com/support/commoncalls.html - look for Drives 8gbs winioctl.h - in this file I found the following lines: #define PARTITION_FAT32 0x0B // FAT32 #define PARTITION_FAT32_XINT13 0x0C // FAT32 using extended int13 services #define PARTITION_XINT130x0E // Win95 partition using extended int13 services #define PARTITION_XINT13_EXTENDED 0x0F // Same as type 5 but uses extended int13 services Francois -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Hamm installation problems
Hi, 123456789 123456789 123456789 123456789 123456789 123456789 123456789 To follow up on my previous message I have found the solution to the first problem. The message Kernel panic: VFS: Unable to mount root fs on 01:00 was in fact caused by a memory detection problem. I have 128M of RAM and unless I say so with a mem=128M the kernel gets confused at some point. I should have specified the amount of memory I have but since it seemed more related to the disk... For the partitionning now. I still only see two partitions. So, to tell you more about my disk, it' a Quantum SCSI VIKING II 9.1WLS. According to Linux fdisk the geometry is C=1110 H=255 S=63 so I get a warning that it has more than 1024 cylinders. But I don't see why it would prevent me from seing the logical partitions. Here is what fdisk tells me about the partitions: DeviceBoot Begin Start End Blocks Id System /dev/sda1 * 1 1 65 522081 6 DOS 16-bit =32M /dev/sda26666 1110 8393962+ f Unknown And cfdisk tells me: Name Flags Part Type FS Type LabelSize /dev/sda1 Boot PrimaryDOS FAT16 (big) [ ] 509.88 /dev/sda2PrimaryWin95 Extended (LBA)8197.23 The problem is /dev/sda2. It is an extended partition not a primay partition. But fdisk will not let me change it's type to 05 which is for Extended partition. If I go to the extended functionality and ask for the list of extended partitions I get an empty list. So maybe the solution is to change its type to 05 but how do I do that ? Now here is what Win95 fdisk tells me: Partition Status Type Volume Label Mbytes System Usage C: 1APRI DOS WIN95 510 FAT16 6% 2 EXT DOS8197 94% Thanks to Harter for telling me about /proc/kmsg. The only problem is, when I type cat /proc/kmsg cat does not show anything and blocks until I kill -9 it. But I've found that now that I can boot ok Alt+F4 will show me the last 25 lines of the boot messages. Francois -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Hamm installation problems
Hi, I just got a new computer and I'm trying to install Debian on it. The computer is a Gateway E3110-333 with an Adaptec 2940UW controller and all SCSI devices. I currently have two problems: - I took the hamm install floppies but when I boot the root.bin floppy I get the following error message: scsi0 : Adaptec AHA274x/284x/294x (EISA/VLB/PCI-Fast SCSI) 4.1.1/3.2.1 scsi : 1 host. scsi0: Scanning channel A for devices. ... scsi : detected 1 SCSI cdrom 1 SCSI disk total. SCSI device sda: hdwr sector= 512 bytes. Sectors= 17836668 [8709 MB] [8.7GB] Partition check: sda: sda1 sda2 [MS-DOS FS Rel. 12,FAT 0,check=n,conv=b,uid=0,gid=0,umask=022,bmap] [me=0x0,cs=0,#f=0,fs=0,fl=0,ds=0,de=0,data=0,se=0,ts=0,ls=0,rc=0,fc= 4294967295] Transaction blocks size = 521 Kernel panic: VFS: Unable to mount root fs on 01:00 That's when I start the kernel using loadlin linux root=/dev/ram initrd=root.bin. But I get the same error message about the device 03:01 if I boot from a floppy (with VFS: Cannot open root device 03:01 on the line just before). The kernel version reads 2.0.33 ([EMAIL PROTECTED]) #2 (gcc 2.7...)Wed Apr 1 04:20:21 CST 1998 The Adaptec BIOS version reads Adaptec AHA-2940 Ultra/Ultra W SCSI BIOS v1.32S10 It seems to be recognised. At least the console correctly lists the SCSI hard drive, CD-ROM and tape unit where the ... are above. Now if I use the old Debian 1.2 rescue floppy (kernel 2.0.27) I can boot and access my hard drive (e.g. I can mount /dev/sda1 and read it). - The second problem now. When I boot with the old Debian 1.2 floppy I get to the point where I must partition my hard disk. I already made partitions for Linux. In fact I know my hard disk has a small (500MB) Primary partition and a big (8.2GB) extended partition in which I have partitions for Linux. The problem is that (c)fdisk sees the second partition as Primary with an Unknown (0F) file system. How do I get to see my extended partitions ? What I did is that I installed Win95 and NT4.0 Enterprise Server on this machine. I don't remember what I used to partition the disk, probably dos fdisk but maybe I later modified it with NT4.0 fdisk at installation time. - (Not really a problem in itself) Is there a way to get a dump of the messages that print on the screen at boot time or at least slow down the scrolling so that I have more time to read them/write them down ? Francois -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Colored ls output
On 1 Aug 1997, Stig Sandbeck Mathisen wrote: BN == Bob Nielsen [EMAIL PROTECTED] writes: BN On Thu, 31 Jul 1997, Shaleh wrote: When I installed Slackware a year ago ls's output was colored. Is there a colored version of ls for Debian or is it a switch for ls?? BN You can add the following to .bash_profile (or whatever). BN alias ls='ls --color=auto' You might also want to set the TERM variable to xterm-color if you're running Xterm and not Rxvt. If all above does not work then maybe your problem is with your Xresources file (/etc/X11/Xresources). Try adding the line below: *customization: -color -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/~fgouget/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: bind
On Mon, 28 Jul 1997, dada wrote: I instaled bind pakage to make an DNS server to my local area network (to learn how setup it). Before install it, Netscape works fine. But now,ie: when I open and web page in my hardisk,Netscape don't work fine (it wait too much time and I have to kill it,becouse it newer open that page). If I stop the deamon named and re-run Netscape,it works fine. Here is what my powerful divination talents tell me. You only have problems when you are not connected to the Internet. What happens is that Netscape reads the file, and for some reason tries to resolve an address that it contains. So it queries the local name server, i.e. your local bind server. But most probably your server does not have the necessary information in its local file. Thus it tries to contact another name server on the Internet. But since it's offline the UDP messages are not answered and eventually time out. After trying many times it will definitely give up. This is why netscape seems dead. It may even happen that if you try to use the menus you will have to kill you X session. At least that's what happened to me. Solutions: - Stop using bind. - Change the named.boot file depending on whether you are connected or not. When not connected don't use forwarders. - Set up dial on demand AND arrange for the DNS resolution packets to trigger the connection. -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/~fgouget/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
where should examples go ?
I noticed that some example files where stored in /usr/doc/package/examples while others are stored in /usr/doc/examples/package. Unfortunately I did not notice the second location until relatively recently. What does the standard say ? Personally I prefer the first setting and more packages seem to respect it anyway. Here are the packages that store configuration files in /usr/doc/examples on my system: deliver 2.1.12-2 ncurses-dev none samba 1.9.16p9 bison A2.6-12 fvwm2 2.0.43-BETA-0 lilo19-1 popclient 3.05-3 tar 1.11.8-5 fwtk1.3-1 mount 2.5l-1 procmail3.10-4 wu-ftpd 2.4-27 Some other packages make symbolic links to the /usr/doc/package/example directories (apache, cgi-scripts, libpaper). On my system bing still stores its Readme file in the file /usr/doc/bing ! -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/~fgouget/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: ip-up question
On Mon, 14 Jul 1997, Jim Foltz wrote: Is it safe to put these lines in the /etc/ppp/ip-up scripts? $6 is passed to the script using the ipparam option of pppd and is the name of the user initiating the ppp link. if [ -e /home/$6/.ip-up ]; then cd /home/$6 . .ip-up fi No because /etc/ppp/ip-up is executed with root permission so the script /home/user/.ip-up would be executed with root permission !!! Certainly not something to allow. Maybe doing a su before would be enough to shut the security problem though: if [ -e /home/$6/.ip-up ]; then cd /home/$6 su $6 -c .ip-up fi -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/~fgouget/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Help with IP masquerading
On Wed, 21 May 1997, Craig Sanders wrote: On Tue, 20 May 1997, Benjamin T. White wrote: **I can not do domain name resolution with my new setup** The ip [...] DNS is one of the limitations of masquerading. It doesn't work. I have the same setup as Benjamin T except that I have two Linux machines. I could not prove it right now because I have installed a DNS server on the Linux doing the masquerading but if I remember well my 486 was able to do DNS resolution before I installed the new DNS. So that was through the IP masquerading. I have the kernel 2.0.27 and I load some optional IP masquerading modules (mainly ftp). Most Linux documentation advises against running bind, saying that it's [...] get it workingit only takes a few minutes at most. I would rather say that it took me a several hours but perhaps I'm worse than average. BTW, if you're using diald you'll probably want to configure it so that it doesn't bring up the link every time you want to resolve a name. But you'll want to do that whether you're running bind or not. In fact if you're using diald having a local bind server is perhaps more trouble than it's worth. Here is why: - Either diald does not bring the connection up for DNS requests. Then applications will seem to hang if the result for their DNS query is not in the cache. They will stay blocked in some gethostbyname call until the DNS server times out which takes quite a long time. With some X applications you can completely freeze the X server (with netscape click on a menu. It does it's name lookup right here and it seems to block X). - The second problem does not depend on whether DNS bring the PPP link up. If your IP address is dynamically assigned by you ISP and you type ftp ftp.debian.org and the name lookup is returned by the local DNS cache then the first packet on the network is the first packet for the TCP conenction. But I noticed that in that case diald seems to send the packet with the wrong source IP address, i.e. that of the fake serial device instead of the one of the fresh new PPP connection. Consequence the connection will never make it, you have to abort ftp and restart it. This effectively prevents me from using diald with the DES client. -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/05/fgouget/ Wonder what to do with all your spare CPU cycles ! Participate to the DES cracking challenge with the SolNet team http://www.des.sollentuna.se/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Help with IP masquerading
On 21 May 1997, Rob Browning wrote: Francois Gouget [EMAIL PROTECTED] writes: I have the same setup as Benjamin T except that I have two Linux machines. I could not prove it right now because I have installed a DNS server on the Linux doing the masquerading but if I remember well my 486 was able to do DNS resolution before I installed the new DNS. So that was through the IP masquerading. Yep, I have the same situation, and name resolution works fine. The only things I've found that don't work are ftp (dir listings only, file gets by wget and netscape work fine (which I don't understand)) and ping. This must be related to masquerading (i.e. not diald). For ftp to work you must load a specific module: ip_masq_ftp. I think this module also does the icmp masquerading (for ping). This is because ftp sends a port number and has the server call you back at that port. There are other specific modules for some other protocols. The modules that I have are: ip_masq_raudio, ip_masq_vdolive, ip_masq_cuseeme, ip_masq_irc. -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/05/fgouget/ Wonder what to do with all your spare CPU cycles ! Participate to the DES cracking challenge with the SolNet team http://www.des.sollentuna.se/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: INN 1.5.1 problem
On Sun, 18 May 1997, Remco van de Meent wrote: On Sun, 18 May 1997, Francois Gouget wrote: [...] 441 Can't set system NNTP-Posting-Host header [...] You're getting these errors because the NNTP-Posting-Host header isn't filtered out. Just write your own script to do that, instead of the post-news script, and it should work again. At least it did for me ;) Okay, I had a look along these lines and solved the problem. In fact the format of suck's get-news.conf changed between 3.3.2-1 and 3.4.1-2 but the installation script of suck did not issue any warning (just told me that I already had a customised get-news.conf file). It now stores the regular expression to be matched for removing the NNTP header. At least squid-1.1.1-1 issued a warning when upgrading my squid-1.0.20-1 because of the incompatible format and diald_0.16.1-4 even proposed to update my existing configuration file (from diald_0.14-9) which it did properly. Clearly the suck-3.4.1-2 installation script should be enhanced. -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/05/fgouget/ Wonder what to do with all your spare CPU cycles ! Participate to the DES cracking challenge with the SolNet team http://www.des.sollentuna.se/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
INN 1.5.1 problem
I recently (yesterday) upgraded my inn and suck packages from the bo debian packages and I find that now when I try to post a message to a newsgroup I get the error: Official host name: news.grolier.fr Alias news.club-internet.fr Address: 194.158.97.89 Connected to news.grolier.fr 200 grolier InterNetNews NNRP server INN 1.5.1 17-Dec-1996 ready (posting ok). 340 Ok 441 Can't set system NNTP-Posting-Host header Malfunction, Unable to post Article! Closing connection to news.grolier.fr Error remote posting I have no clue about why I get this error. Can anyone help me ? I have 1.5.1-2 and suck 3.4.1-2. Previously I had inn 1.4unoff* and suck something both from Debian 1.2.14. By the way I think that the following are suck bugs: - When I installed suck get-news.inn would not work unless I changed the -s argument of testhost into -S /dev/null (was the same with the suck that came from debian 1.2.14). - Now that I can no longer post suck sees that it fails but never retries to post the articles. So they will never be posted. -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/05/fgouget/ Wonder what to do with all your spare CPU cycles ! Participate to the DES cracking challenge with the SolNet team http://www.des.sollentuna.se/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Hanging diald connections
On Mon, 12 May 1997, Christian Lynbech wrote: [...] But if I subsequently tried to connect to that particular host, the link would start up as normal, but the program would just hang and get nowhere. Killing the application and trying again, or trying the application from another shell (after the link had come up) worked fine, as did connecting to another host, most likely because the nameserver query would bring the link up before the actual communication took place. I have the same problem: if i start with ftp somehost somehost actually being in my local name server cache the link comes up but ftp just hangs there. The same goes on for the DES cracking client (which is annyoing). Are you using PPP with a dynamically assigned address ? I suspect that the first packet gets sent with the FROM IP address being that of the serial device and not the address that was assigned to the PPP link when it came up. I currently consider this to be a diald bug or at least a severe limitation. Maybe someone knows how to deal with it ? -- Francois Gouget [EMAIL PROTECTED]http://www.mygale.org/05/fgouget/ Wonder what to do with all your spare CPU cycles ! Participate to the DES cracking challenge with the SolNet team http://www.des.sollentuna.se/ -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
