Re: can't kill a PID
On Sun, Nov 03, 2002 at 06:34:55AM -0500, Kevin Coyner wrote: > I always thought that with 'kill -9 PID' you could clean up just about > any process, but I've run into one that just won't go ... > kosuke9026 0.0 0.9 14460 4932 ? D 00:16 0:00 xmms The state ``D'' means uninterruptible. > kosuke9027 0.0 0.0 00 ? Z 00:16 0:00 [xmms ] This process is a child of the uninterruptible process. Since you cannot kill the process in D state, the process in Z state will not reparent to init (by losing it's current parent) and be cleaned up by the kernel when init waits on it. -- Matthew Gregan |/ /|[EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Exim config problems
Greetings... I've tried searching the debian-user list archive already, and found some messages which probably contain the answer I'm after, but it seems that the actual messages aren't available, so I can't read them... (It was working a week ago when I last tried, but not yesterday or today...) Anyway, this should be an easy problem to solve for somebody. I've played around with eximconfig a bit, but I can't find how to solve the problem. I'm using exim and fetchmail for my mail delivery. Exim is setup using option 2 in eximconfig (recommended for dialup systems). Everything is working fine, I can receive email and send it fine, except if I try and send email to people on my ISP. The problem is that outgoing emails have their addresses rewritten as @ihug.co.nz, and my mail system classes these emails as local, so it tries to deliver them on my machine and fails. I thought I'd fixed this by stopping ihug.co.nz being considered local, but then email generated by cronjobs and such were finding their way to my provider ([EMAIL PROTECTED] - since the outgoing mails were having their addresses rewritten). I think it should be a simple fix to solve this, but I don't know what it is... I've checked howtos but they only seem to cover other MTAs. :-/ Thanks in advance. -- Matthew Gregan[EMAIL PROTECTED]
Re: Exim config problems
On Tue, Jun 08, 1999 at 11:01:39AM -0500, Marc Mongeon wrote: > These three settings are important in exim.conf: > > qualify_domain = ihug.co.nz > qualify_recipient = localhost > local_domains = localhost Thanks, that seems to have done the trick... I had wondered if 'qualify_recipient' was part of what I needed to change, but I decided it was best to find out instead of my ISP having to forward mail for my root account back and tell me my mail system was still broken. :-) Thanks again. -- Matthew Gregan[EMAIL PROTECTED]
Re: where is xemacs21?
On Sat, Jun 12, 1999 at 03:07:41AM +0200, Hartmut Figge wrote: > [EMAIL PROTECTED] wrote: > > Anyone know where to get the debian package for xemacs21.1? > > hm, not to be found on debian.org, but use > > http://ftpsearch.lycos.com/?form=medium > > and search for xemacs21, sorted by date/host. i´ve just done that and > found the packages. Has anybody tried installing these packages yet, and had much success? I grabbed the bin, support and nomule packages (version 21.1.2-1) and installed them. There seemed to be quite a few failures in byte-compiling the .el files (which I don't remember seeing with xemacs20) and some functionality appears to be missing which is working in xemacs20 (and emacs20) (e.g. M-x spook is unavailable in xemacs21). I installed the same set of packages for xemacs21 as for xemacs20, so I don't know what the problem is, I guess the packages are not quite ready yet, or I messed up somewhere. Cheers. -- Matthew Gregan[EMAIL PROTECTED]
Re: matrox g400
On Tue, Jun 15, 1999 at 07:36:32PM -0500, Aaron Solochek wrote: > I realize this isn't a great place to ask this... but I'm considering > getting the g400. I was wonding if anyone knows, or can make an > educated guess as to weather or not this will work with X at all. > I assume that the drivers for this will take a while to come out, but > enough to get by would be fine, since I play my games in windows > anyways. Thanks. Well, don't take this as gospel, but I have heard from a reasonably reliable source that the X server for the G200 (XFree86's SVGA for example) will work with the G400 as well, since they are very similar (2D wise). As far as the 3D (GLX) drivers go, I've heard that they won't work without Matrox releasing some specs on them, but that remains to be seen, I guess. -- Matthew Gregan[EMAIL PROTECTED]
(X)Emacs in the console
Hi. I've been playing around with configuring Emacs20, and XEmacs20. I've got it all working out I want under X, but under the console, XEmacs thinks that my backspace key is delete. I would've thought this had something to do with the keyboard settings for the console, but Emacs works correctly, with backspace calling backward-delete-char and delete calling delete-char. I tried rebinding backspace in XEmacs while in the console, but when I hit the backspace key to tell it what key to rebind, it shows 'delete' as the key it read, so I'm stuck... Can someone tell me what I need to do to fix this, or point me to something which will help? Thanks. -- Matthew Gregan[EMAIL PROTECTED]
Re: emacs' delete behaviour changed?
On Sat, Jun 26, 1999 at 09:00:12PM +0200, Pierfrancesco Caci wrote: > Did someone else notice that emacs in potato doesn't differentiate > delete and backspace keys any more, as it did in slink? Yep. Take a look in the changelog in /usr/doc/emacs20 and you'll see why. I can't remember the exact reason, but it was because the old behavior was breaking parts of Emacs. -- Matthew Gregan [EMAIL PROTECTED]
Re: ip masquerading/port forwarding
On Sun, Jun 27, 1999 at 12:24:03AM -0400, Paul Miller wrote: > I'm using Debian/unstable and kernel v2.2.10. I have a ip masquerading > Linux box setup and working. I'd like to configure ports 137 to 139 of an > internal machine to act as ports 20137 to 20139 on the external interface of > the Linux box. I.e., the internal ip address sent from the internal machine > is replaced with the external interface's address and anything sent to the > external interface on ports 20137 to 20139, the external interface ip > address is replaced with the internal machines ip addressed and forwarded to > the internal machine on ports 137 to 139. Okay, I'm assuming here that you have portfw compiled into the kernel, or available as a module, and that you have ipmasqadm installed. ipmasqadm portfw -a -P tcp -L external.ip 20137 -R internal.ip 137 ipmasqadm portfw -a -P tcp -L external.ip 20139 -R internal.ip 139 ...replacing external.ip and internal.ip with the obvious things :-) Now this works for requests coming in - requests to port 20137 are rewritten to port 137 and forwarded to the internal machine... I think you also want something to rewrite the outgoing stuff from the internal machine using port 137 to 20137 as well, right? This should do it: ipmasqadm portfw -a -P tcp -L internal.ip 137 -R external.ip 20137 ipmasqadm portfw -a -P tcp -L internal.ip 139 -R external.ip 20139 ...again replacing internal.ip and external.ip with the obvious. However, this time you want to use the internal.ip of the masquerading machine (I think, try it both ways). If Samba needs UDP as well (I don't think it does...) then double up the entries, replacing 'tcp' with 'udp' for the second ones. I haven't actually tried this, but it should work fine. Let me know... -- Matthew Gregan [EMAIL PROTECTED]
Re: Exim Problem (I think....)
On Wed, Jun 30, 1999 at 06:47:38AM -0400, Didi Damian wrote: > Well, it works for me. I'm using kernel 2.2.10. When setting up > fetchmail, I couldn't get it to work so, after reading these posts, > I downgraded to slink exim (2.05). Then I realized I had fetchmailrc > wrong and re-installed exim 3.02-1. It's working fine. My kernel is > compiled without IPV6 support. I can back this up. I'm also running kernel 2.2.10(-ac5) with no IPv6 support, and Exim (3.02-1) is working fine here also. Having seen the recent problems with Exim being posted to the list, I thought that was my problem when my outgoing mail stopped working the other day, however after some poking around I discovered it was in fact Mutt looking for sendmail (and failing) that was my problem. This is a known bug in Mutt, and the solution is listed in the Debian bug database with the bug report. -- Matthew Gregan [EMAIL PROTECTED]
X modeline conversions
Hi everyone. I'm trying to work out a good modeline for X for 1152x864. The reason I'm doing this is because I was installing some new drivers for my video card under Windows and discovered that my monitor can do 1152x864 at an acceptable refresh rate... I've got the figures from Windows, but I only vaguely understand how they relate to the X modelines, and I can't figure out how to convert them over. Hopefully someone with more experience can point me in the right direction... Under Windows, this is the info I have about the mode: Pixel clock: 112320kHz Horizontal: Frequency: 70kHz Front porch: 64 pixels Sync: 128 pixels Back porch: 256 pixels Negative sync polarity Vertical: Frequency: 78Hz Front porch: 1 line Sync: 3 lines Back porch: 2 lines Negative sync polarity I've tried reading through ESR's XF86 video timings howto, but it wasn't really helpful to my problem. From what I understand, the front/back porch is the time of the rise and fall of the signal, but I don't know how to relate the numbers I have back to anything useful for X. I've found a modeline already existing in the XF86 config which is very close to what I want (1152x864 at 78Hz, but it's 70.8kHz horizontal frequency is just too much for my monitor to handle). If anybody can help, or point me in the right direction, I would be very happy. Thanks. :-) Oh, those Windows figures are from the Matrox monitor customization utility, if that matters... -- Matthew Gregan [EMAIL PROTECTED]
Re: gnome session-panel & X problems
On Sun, Jul 04, 1999 at 12:35:52AM -0400, Didi Damian wrote: > Whenever I try to start the gnome panel and/or gmc they fail with this error: > > [EMAIL PROTECTED]:~/.enlightenment/cached/pager/ > panel > ** WARNING **: Could not get name service! > ** ERROR **: file goad.c: line 606 (real_goad_server_activate): assertion > failed: (name_service != CORBA_OBJECT_NIL) > aborting... > Aborted Try starting gnome-name-service first, that will keep them happy. This is what I do when I'm not running my X session with gnome-session, and I want to start a Gnome app. > Does anybody know what's wrong and how can I fix it ? What you might want to try is deleting or renaming the .gnome dir in your $HOME. I've found that the old configs can cause problems when you upgrade Gnome. The gnome-name-service *should* be started by gnome-session, so if you're running your X session using gnome-session then it's probably your old configs causing the problems... Otherwise you can always start gnome-name-service by hand when you need it. -- Matthew Gregan [EMAIL PROTECTED]
Re: gnome session-panel & X problems
On Sat, Jul 03, 1999 at 05:59:29PM -0700, Brendon Baumgartner wrote: > I installed gnome and e. The session panel doesn't seem to load up. All i > can do is right click and middle click. > > Also, another problem i'm having with X: > When I Log out, the window manager closes, (the borders and such > disapear) but the system doesn't return to gdm. It sits in X till i press > ctrl-alt-bkspc. Anyone know why this might be? I can't figure it out. Firstly, are you starting Gnome with something like 'exec gnome-session' as the last thing in your .xsession (or .xinitrc if you're starting X by hand)? If so, then when you 'log out', Enlightenment is quitting, but not Gnome, so your X session is still alive (just without a window manager). This probably means that something is pretty wrong with Gnome, and it's not being told that E wants your X session to end (since gnome-session is in control, not E). If you have a .gnome directory in your home dir, remove or rename that and try again. It may be you have some stale configs messing things up. Otherwise, something is probably pretty wrong with your Gnome install... -- Matthew Gregan [EMAIL PROTECTED]
Netscape and X resources
Hi everyone. Does anybody know where I can find a list of the X resources that Netscape (4.61) supports? I've had a good look around and come up with nothing (except one .Xdefaults file which doesn't do what I want). Specifically, I'd like to change the font Netscape uses for it's toolbars. If anyone has a link or an example .Xdefaults, let me know. Thanks. -- Matthew Gregan [EMAIL PROTECTED]
Re: Alternastive to using squid
On Wed, Jul 07, 1999 at 12:35:49PM +0100, Patrick Kirk wrote: > I do not need the caching, etc. that one gets with squid. What do I need to > do just to have internet access via this machine? Take a look at tinyproxy... It should do what you're after (it's just a proxy, no caching). Or you could try something like junkbuster, which is also a proxy, but has the ability to filter stuff from sites (such as advertisements). If you want to do more than just HTTP, you might want to look into IP masquerading, which will let you do (almost) everything as if the machine were connected directly to the internet... -- Matthew Gregan [EMAIL PROTECTED]
PGP5, mutt, and signed messages
Greetings everyone. After seeing quite a few postings about Mutt and PGP in the past few days, I decided I'd finally get around to setting it up myself. I've got it working fine with GnuPG and PGP2, but I have a question about PGP5... I don't know if this is really specific to Mutt, though. I sent a test message to myself, and signed it with PGP5. When I receive the message, Mutt checks the signature and reports that it's 'good'. Good signature made 1999-07-08 00:39 GMT by key: 1024 bits, Key ID C22BCF6F, Created 1999-06-28 "Matthew Gregan <[EMAIL PROTECTED]>" This signature applies to another message That last line has me a little worried. Is that the way PGP5 is supposed to work (since the signature is attached to the message using MIME, rather than being part of it (such as the older method of PGP signing messages used). That's the only thing I can think of, I just thought I'd check and see what experiences others have had. One other thing, I saw references to 'language.txt' and 'language50.txt' (for PGP2 and PGP5, respectively) in the Mutt docs. However, it appears that the Debian package doesn't come with these files. Would this be considered a bug, and if so, should I file a report? Or is there a Debian policy which doesn't include such things? Thanks. -- Matthew Gregan [EMAIL PROTECTED]
Re: How to fix Netscape in the interim
On Sun, Jul 11, 1999 at 04:58:28PM -0500, Stephen Pitts wrote: > To fix netscape, do "apt-get source liburi-perl", then edit > debian/control to depend on perl5 instead of perl. Do a > "debian/rules binary" and install the resulting package. Voila! > Netscape will install Does this not break liburi-perl? If not, I wonder why it's taking so long for it to be fixed in potato... -- Matthew Gregan [EMAIL PROTECTED]
Re: Is IRC #debian used often?
On Mon, Jul 19, 1999 at 09:49:30PM +, Brian Greenfield wrote: > Just out of interest, I took a look at #debian on undernet > (I tried a few servers, details upon request). Rather, I > *tried* to join #debian, but there wasn't a #debian to > join:( > > So, is it used -- it is undernet isn't it? Try irc.debian.org, in channel #debian. That seems like quite an active place, with a lot of friendly people. :-) I think that the irc.debian.org server is part of a network of IRC servers (OpenNetworks or something), which links a few similar interest servers together. -- Matthew Gregan [EMAIL PROTECTED]
Re: apt source file semantics
On Sat, Jul 24, 1999 at 10:14:00PM -0600, Eduardo Alvarez wrote: > I have a small question: Precisely, what is the syntax to be used when > addling a source in sources.list? Specifically Speaking, when do I stop > using slashes and leave a space between folders? If you want a verbose description, try (if you haven't already): $ man sources.list The short answer is that it depends on what the type is (the first word of the line, e.g. 'deb'). It's much better explained in the man page than I can do (without basically repeating the man page). -- Matthew Gregan [EMAIL PROTECTED]
Re: bash scripting
On Sun, Jul 25, 1999 at 12:53:58AM -0500, [EMAIL PROTECTED] wrote: > Is there a way to put a 30 second pause in your bash script? Use: sleep 30s 'man sleep' for more info. :-) -- Matthew Gregan [EMAIL PROTECTED]
Re: bash scripting
On Mon, Jul 26, 1999 at 08:18:48AM +0200, Gerhard Kroder wrote: > [EMAIL PROTECTED] wrote: > > > > Is there a way to put a 30 second pause in your bash script? > > > ever RTFM for bash? it says: > >wait [n] Erm, how exactly can you use this to wait for 30 seconds? I think he wants something along the lines of sleep. -- Matthew Gregan [EMAIL PROTECTED]
Re: Gnomeicu problems
On Thu, Jul 29, 1999 at 06:37:10PM -0500, Oleg Krivosheev wrote: > just installed gnomeicu and while trying to > run it i got > > > reboot ~ $ gnomeicu > type = 0 exid = (null) > > ** CRITICAL **: file applet-widget.c: line 655 > (gnome_panel_applet_corba_init): assertion > `panel_client != ((void *)0)' failed. > > ** CRITICAL **: file applet-widget.c: line 699 (applet_widget_new): > assertion `c orbadat!=NULL' failed. > > ** ERROR **: Can't create applet! > > aborting... > Aborted Are you running the Gnome panel? If you aren't, that's probably the problem. GnomeICU is trying to create the applet in the panel, failing, and dying. Try starting it with the '-a' argument to stop the applet creation, e.g. gnomeicu -a That will work whether the panel is running or not. -- Matthew Gregan [EMAIL PROTECTED]
Problems installing Debian on a 486
Hi everyone. I've got 2 old 486 machines here that I'm trying to install slink onto. One of them is a Compaq Deskpro 486/33M and the other is built out of off the shelf parts. I've managed to install onto the Compaq machine with no problems whatsoever, but when I try to boot the rescue/install disk on the other machine, it brings up the boot: prompt, and then starts to load the root FS from root.bin.. After loading for a while it fails with the message "A20 gate not responding!"... I don't know what the problem is, so I was hoping someone could help me out. I've tried turning off both processors caches in the BIOS, but that made no difference... I can't see any BIOS entries relating to the A20 gate, so I haven't been able to experiment with that... The problem machine is a 486DX33 with 8MB of RAM, if you need more info about the hardware, please let me know. Thanks in advance. -- Matthew Gregan[EMAIL PROTECTED]
RE: Problems installing Debian on a 486
At 15:03 1999-03-23 -0800, Alan Bailward wrote: >> other machine, it brings up the boot: prompt, and then starts to load the >> root FS from root.bin.. After loading for a while it fails with the message >> "A20 gate not responding!"... >[snip] >> The problem machine is a 486DX33 with 8MB of RAM, if you need more info >> about the hardware, please let me know. > >Could it be that the kernel on the rescue disk was compiled for 586+? Could >that be causing it? I'm grasping I know, but... :) These are the rescue disks from the debian-slink disk-i386 dir... I'm pretty sure they're not using a kernel for 586+ architecture... I should also mention that someone suggested I try the tecra rescue disk. I did that, and had no luck... I get the same message, so now I'm really stumped, having read the readme for the tecra disk I would have though that'd fix the problem. Thanks again... -- Matthew Gregan[EMAIL PROTECTED]
Questions regarding kernel 2.2.1 and masquerading.
Hi everybody. I've got a machine here I want to use as a firewall/masquerade box. I've compiled and installed kernel 2.2.1 using kernel-package and then replaced the things listed on the page dealing with Debian 2.1 and kernel 2.2.x issues... Namely, I've installed the newer netbase from potato. However, I may have gone about this completely the wrong way (I'm new to all of this), what I did was: After figuring out what packages I had that broke when using kernel 2.2.1, I went about grabbing them from potato via ftp. The only package I required was netbase, but on attempting to install that, I found I needed to upgrade my libc, which then led me to find out I needed to upgrade apt. So I grabbed those via ftp and installed them in the order: apt, libc6, netbase. Now, everything seems to be running fine, but I have a bad feeling that I've done this completely the wrong way. What makes me think this, is that if I run dselect, it tries to install the timezone package that was removed when I updated libc6, and it also sees libc6 as installed, but it wants to install the older one... Was this the correct install procedure, and if not, what should I have done? I have one other question, when reading through the IP Masq howto, it refers to an /etc/rc.d/rc.firewall (and other /etc/rc.d/ scripts such as rc.local)... since I know Debian has a different layout to that, where do those files go? I've read through the Debian policy, but it didn't really help.. what's the story with these Thanks in advance for your help to these relatively stupid questions. :-) -- Matthew Gregan[EMAIL PROTECTED]
Re: [OT] How fast is a PIII?
On Fri, Aug 06, 1999 at 12:27:25PM +0200, Nils Rennebarth wrote: > On Fri, Aug 06, 1999 at 09:48:05AM +0200, E.L. Meijer Eric" wrote: > > The only difference between a PII and a PIII at the same clock speed > > that I know of are the extra `SSE' instructions that are mainly useful > > for 3D stuff and maybe some other floating point intensive software. > IIRC, the 2nd level cache of the Pentium III is running at the same speed as > the CPU, wheras the 2nd level cache of Pentium II's is running at half the > CPU speed. Actually, the P3 has the same speed L2 cache as the P2 (that is, half the CPU speed). The Celeron-As and the Xeons are the ones with full-speed L2 cache on them. > This should give a performance difference (though not enough for me, > compared to the price difference) I haven't benchmarked them myselves, but I have heard that the P3 is *slightly* slower at the same clock speed due to some changes made to the core to enable the clock rate to climb later. (i.e. a 450MHz P2 might be slightly quicker than a 450MHz P3, but they can't build a 600MHz P2, but they can build a P3). The above statement doesn't include the benefits you would receieve in a application optimized for and using the new P3 SSE instructions, obviously. Then there's quite a large difference in speed. -- Matthew Gregan [EMAIL PROTECTED]
Re: xemacs21 questions
On Fri, Aug 13, 1999 at 08:41:21AM +0300, virtanen wrote: > 1) Do you get somewhere deb-package for xemacs21? If so, where? They are in potato (unstable) right now, they've been there for a month or so now... If you're running slink (stable) then you'll have to build it from the debianized sources, or hunt down a slink package for it. > 2) Where do you get gtzpackages for it, if there aren't deb-packages? Not sure about this, xemacs.org and/or mirrors maybe? > 3) Which kind of mule support does it have? > 4) Did that mule-support work in your case? Can't answer these sorry, since I don't use mule. -- Matthew Gregan [EMAIL PROTECTED]
Re: xemacs21 questions
On Fri, Aug 13, 1999 at 12:07:18AM +0300, Micha Feigin wrote: > I just installed xemacs21. It's still running together with xemacs20 > untill it will work fine. (I Installed it because I hoped for some new > mule support which apperently isn't present yet). > Anyway, It doesn't recognise any languages. > When i open c/c++ or tex files for example, it doesn't know how to > highlight them. I get a mesage (which goes by to fast to read exactly) > which reads something like can't find c++ major mode This is because all of the extra features/packages of xemacs have been split into individual packages as of xemacs21. Go to www.xemacs.org and read up the section on using and installing packages. I've just gone through this myself, since the xemacs21 in potato seems a bit broken in some respects (but I don't know a great deal about xemacs or emacs in general). If you need some help installing packages, let me know and I'll tell you what I had to do to get it all going... > Also, how do i start deferent modes? > I ran into the x-symbols package for latex under xemacs and i was > wondering how to use it. > > I also get a massage when it xemacs starts that it can't turn some font > strings to font struct. Don't really know about these, sorry. I'm still learning the finer side of (x)emacs use atm. I'm trying to figure out why my xemacs21 tells me a bunch of warnings such as: (2) (warning/warning) Autoload error in: /usr/lib/xemacs/xemacs-packages/lisp/xemacs-devel/auto-autoloads: Already loaded When I start it and try to do something (e.g. I hit M-x and it opens another buffer with a bunch of those warnings).. -- Matthew Gregan [EMAIL PROTECTED]
Re: fetchmail: SMTP connect to localhost failed: Connection refused
On Thu, Aug 19, 1999 at 04:57:43PM -0500, Stephen Pitts wrote:
> On Thu, Aug 19, 1999 at 08:52:48AM -0700, root wrote:
> > I'm doing a fresh slink install on a friend's old '486. I've
> > installed fetchmail, but keep getting the error msg: fetchmail:
> > SMTP connect to localhost failed: Connection refused I edited
> > /etc/exim.conf and added 'localhost' and my hostname to
> > 'local_domains=' and restarted exim, but no change.
> Are you sure that exim is running? The default /etc/init.d/exim has a
> big 'exit 0' at the very top of the script?
Correct me if I'm wrong, but Exim is supposed to be run from inetd by
default (thus the 'big exit 0'). There are only certain cases where
you would want/need Exim running all the time...
> What does "ps aux | grep 'exim'" say?
On a system where Exim is run from inetd, it won't show up as running
in the process table unless there's a connection in progress...
In response to the original question, check your /etc/inetd.conf and
your /etc/hosts.{allow,deny} to make sure everything looks in
order. Try telnetting to the smtp port to check if it's Exim/inetd not
letting you connect, or whether fetchmail is misconfigured.
--
Matthew Gregan [EMAIL PROTECTED]
Re: Eterm - menus not usable.
On Sat, Aug 21, 1999 at 07:56:58PM +0100, Phillip Deackes wrote: > I am trying to use Eterm with Window Maker. I am using wmaker v. > 0.60.0-4 (and wmaker-kde, same version number). The menus look normal > when the window is unfocussed, but as soon as I move the mouse into the > window, the menus turn black and are un-delectable. > > I've just tried icewm and the same thing happens. It looks like the > problems is not with Window Maker. Anyone else seen this? It's probably turning dark grey/brown. It's actually the configuration of Eterm which is causing this 'problem'. I think it's quite a bizarre default to have (with inactive being quite bright, and selected being quite dark), but I have menus turned off, so I haven't bothered to correct it. However, the menu colours are completely configurable (and in newer versions, even support pixmaps). Check out the Eterm manpage for info about changing the colours... Hope this helps. -- Matthew Gregan [EMAIL PROTECTED]
Re: E! Enlightenment for newbie?
On Thu, Aug 26, 1999 at 05:18:39AM +0100, John Gay wrote: > 1: Most .deb's I've installed went rather well. Sometimes I have to > look for other requirement first. Are there any requirements I > should get BEFORE installing the Enlightenment.deb's? The easiest way is to just try to install the .deb. If you're missing anything it requires, it'll let you know... However, the version of E! I have installed depends on the following: Depends: freetype2 imlib1 (>= 1.9.4-1) libaudiofile0 libc6 (>= 2.1) libesd0 (>= 0.2.10-0.19990424.6.2) libfnlib0 libjpeg62 libpng2 libtiff3g libungif3g (>= 3.0-2) | giflib3g (>= 3.0-5.2) xlib6g (>= 3.3.3.1-1) zlib1g (>= 1:1.1.3) enlightenment-docs That's for the latest version of E! in the unstable Debian distribution (0.16.devel.3). The demands for earlier versions are probably quite similar. > 2: Will E configure as my default WM, or do I have to go and edit > something? It shouldn't do. Check out the manpage for window-managers(5) for information about changing the default WM in Debian. > 3: Is E fairly stable? I know what E says, I'd like some independent > opinions. The answer you get for this will really differ depending on who you ask. Personally, I've never had Enlightenment crash on me (I run the latest versions from unstable), except when I've done things to cause it to crash myself. So for day-to-day use, I've found it very stable. > 4: Is the configuration fairly straight forward? I get the > impression that there is a lot of GUI configuring tools for E. Are > they easy to find and use? Some of it is quite easy if you use something like E-Conf, or with the >0.16 releases of E!, the built-in GUI configuration tools. However, the theming side of things, and some of the configuration can be quite messy compared to the other WMs. > 5: Would this be an easier WM for my daughter? I'd like to provide > her with as much access as possible to encourage her to learn how to > explore the PC and make the most of it. It depends how you set it up... It can act quite a few different ways, so it's hard to answer that. :-) > 6: I've looked at KDE and GNOME, but right now GNOME seems a little > TOO new for my liking, and I've heard many complaints that KDE is > somewhat bloated. As a newbie, I'm just not ready to tackle these > yet. How would you rate E in relation to these? There is no comparison. E! is a window-manager, whereas KDE and GNOME are desktop enviroments. They don't manage windows (although KDE comes with the KDE window-manager KWM, and GNOME normally is used in conjuction with E!). > 7: At the moment, my daughter's PC is limited to 8 bit colour. Can E > be configured to reduce any problems this causes? Right now, with > fvwm95, if I open one app with lots of colours, than another one, > the second complains there are not enough colours left and the > screen keep switching different colours as the two app's borrow from > each other. Is there some way to make the apps use the same colours? That's really dependant on the X server configuration. Most of the themes for E! tend to use a *lot* of colours, so you'd probably find the problem you described was worse using E!, but that's not to say you can't use a low-colour theme... But the way apps handle colour allocation is out of the hands of the window-manager. Well, that's my thoughts on it. -- Matthew Gregan [EMAIL PROTECTED]
Re: E! Enlightenment for newbie?
On Fri, Aug 27, 1999 at 02:28:25AM -0400, Robbie Huffman wrote: > Hmmm...I though E and most of GNOME used Raster's imlib, which means they do > an excellent job of dithering, even in 8 bit color. At least that has been my > expereince. It's the next best thing to re-writing X code to get applications > to share colors nicely. Of course, a low color theme will free up color > cells. And you can set imlib to use a very small color pallete, thus saving > colors for other applications at the cost of appearance. Just run > imlib_config. This is true, I forgot to mention this (well, it entered my mind, but since I only run X in 16/32bit, I don't know too much about the handling of colour shortages). The few times I've run X in 8bit mode, (using E!) it has looked quite nice (much better than trying a similar thing in Windows), though. -- Matthew Gregan [EMAIL PROTECTED]
Re: E! Enlightenment for newbie?
On Sat, Aug 28, 1999 at 12:33:30AM +0100, John Gay wrote: > available, but the desktop is still 1024x768. At lower resolutions, I can use > the mouse to scroll around the screen. How to I get the desktop to change > with the resolution? In the XF86Config, check the size you've got for "Virtual" in each display depth setup. This determines the virtual desktop size... But what also happens is, if you have no "Virtual" entry for that display depth, XFree86 will take the largest listed resolution as the virtual desktop size, no matter what your visible resolution is. For example, if you had: Modes "800x600" "1024x768" "640x480" Then X would start up in 800x600 at that display depth, but still use a virtual desktop size of 1024x768. Which is pretty much what you described. I'm not 100% sure, but if you specify a "Virtual" size of 0 0, it should use the same virtual desktop size as the default resolution, so for the above example it would pick 800x600 as the virtual desktop size. But I'm not sure if it works this way. Alternatively, you can delete the higher resolutions (1024x768 in the example) so the largest listed resolution is the one you want to use. This is not a great solution. You could probably also try setting the virtual desktop size (via "Virtual") to the same as the resolution you want to use (e.g. "Virtual 800 600" in this example). Check out the manpage for XF86Config if my description is too badly written to make much sense. :-) -- Matthew Gregan [EMAIL PROTECTED]
Re: "import"ing screenshots?
On Sat, Sep 04, 1999 at 01:19:25AM -0400, Harlan Crystal wrote:
> I recently reinstalled my debian system, and the command which i had
> been using to take screenshots ("import thepicture.jpg") no longer
> works. I seemed to remember that import was associated with "xv" or
> "imlib" but I have both installed and "import" still doesn't work.
> Does anyone know how I could enable Import or some type of screenshot
> taking utility? I'm running all off the newest upgraded unstable with
> kernel 2.2.12.
What you're after is imagemagick.
Just watch out not to install libmagick4g-lzw, since that will break
imagemagick (I don't think the maintainer has fixed it yet). Use the
non-LZW one instead.
Have fun.
--
[ Matthew Gregan ] [ crash.ihug.co.nz/~kinetik ]
[ [EMAIL PROTECTED] ] [ PGP key ID: 2048/D1B2FF49 ]
[ PGP fingerprint: 15 CF 1F D5 7C 59 63 F5 79 52 A5 7D EB 50 E5 27 ]
Re: SoundBlaster 16 PNP WaveEffects help needed
On Sat, Sep 04, 1999 at 03:59:19PM -0700, Brian E. Lavender wrote: > Where do I find the source for modconf, and who is the maintainer? # dpkg -s modconf Package: modconf Status: install ok installed Priority: optional Section: base Installed-Size: 171 Maintainer: Enrique Zanardi <[EMAIL PROTECTED]> Version: 0.2.26 Depends: whiptail, modutils (>= 2.1.85-14) Description: Device Driver Configuration Modconf provides a GUI for installing and configuring device driver modules. The source is available at ftp.debian.org (and mirrors), or if you have apt-get set up correctly (you probably need the newer apt-get from netgod.net if you're running slink): # cd # apt-get source modconf And then hack away. If you have a new enough apt-get that supports the 'source' command but don't have it set up correctly, you need something like this added to your /etc/apt/sources.list: deb-src http://http.us.debian.org/debian stable main contrib non-free -- [ Matthew Gregan ] [ crash.ihug.co.nz/~kinetik ] [ [EMAIL PROTECTED] ] [ PGP key ID: 2048/D1B2FF49 ] [ PGP fingerprint: 15 CF 1F D5 7C 59 63 F5 79 52 A5 7D EB 50 E5 27 ] pgpc82bO6veuB.pgp Description: PGP signature
X and nice values
Greetings. I was reading somewhere recently that you should run X with a nice value of -10... I can't remember which document I saw this in now, I think it was one of the X setup HOWTOs. Well, I have a couple of questions: 1) Is this changing of the nice value considered a good idea? 2) If I use XDM/WDM, how can I set the nice value of X? I can figure out how I'd do it if I was starting X from the console, but I've no idea how to do it from WDM. Thanks in advance. -- Matthew Gregan[EMAIL PROTECTED]
Re: Netscape 4.6 crashes
On Mon, May 31, 1999 at 07:29:48PM +0200, Andreas Voegele wrote: > The faulty behaviour is most obvious if you've got more than one > Navigator window opened. Instead of closing the selected window only > the whole application crashes sometimes. I've been having that exact same problem here... Running potato with kernel 2.2.9 and Communicator 4.6. I was also having the problem with an earlier version (probably 4.51 or 4.5). > I'm using IceWM, but I don't think that this problem has something to > do with the window manager. It doesn't seem to be the window manager affecting it. I was using Enlightenment, and I switched to WindowMaker suspecting that E was the problem. I discovered that it wasn't, and since you're having the same problems with IceWM, it looks to be WM independent. -- Matthew Gregan[EMAIL PROTECTED]
Re: Anyone using Sound Blaster PCI64?
On Tue, Oct 12, 1999 at 06:27:33PM +0200, Joop Stakenborg wrote: > At the moment I am using a Sound Blaster vibra16x pnp card, > this card is not really supported by OSSFree (kernel) > or ALSA. Although the card produces sound, it gives dma > errors when using the second dma channel and it is > unable to work in full duplex. I am getting fed up with > this card I used to have a Vibra16C, which worked fine. I haven't had any experience with the 16X, but I believe they aren't well supported, and have a problem with respect to not having a 16bit DMA channel. > I am thinking about a new one. The PCI64 by Sound Blaster > seems a good candidate. Does anyone have experience using this > card? Some Questions: > > - Can I configure it with ISAPNP? If not, how do I do it? > - Does it work full duplex? I just put one of these in my machine - it's using the ens1370 chipset (some of them use the ens1371 or ens1373). I'm actually using it with the ALSA drivers, and it works fine (provided I keep the volume levels on the mixer below 90% for both the Master and PCM outputs, otherwise I get nasty clipping on the sound). You don't need to configure it with ISAPNP, because it's a PCI card, both the ALSA drivers and the kernel drivers find the card by themselves. I'm not sure about full-duplex mode, since I never use it. I'm pretty sure it works full-duplex under ALSA, but I have absolutely no idea about the kernel drivers. Hope that helps a little. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgp1bEWgo4FuP.pgp Description: PGP signature
/etc/hosts manpage
Hi everyone. I'm going to file a bug report about this, but I thought I'd try and gather a little more information first. I'm running potato, and I just noticed there is no manpage for the /etc/hosts file. I don't remember if there was one for slink (could someone please let me know if there is, and what package it comes from?) I'd guess that netbase should be supplying the manpage, but I don't know. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgptMOFQqjdN1.pgp Description: PGP signature
Re: Help
On Wed, Oct 20, 1999 at 01:44:53PM -0500, John Franck wrote: > I have a Savage4 graphics card. I am trying to run X on my computer, but > don't want to use VGA16. I downloaded the latest version of X 3.3.5, which > was supposed to support Savage4 for Linux, but was unable to find it in the > supported card list. Does anyone know what I can do to get this to work?? > --Any help or success stories would be much appreciated. Did you upgrade to 3.3.5 proper, or just install a newer X server binary? If you did the latter, the X configuration tools will still be the old ones, and not know about support for the Savage4. If you know which server the support for that card is built into, you should be able to use the old config tools and just lie to them about what card you have. As an example, when I first installed slink, I needed to upgrade my X server to 3.3.3 to get support for my G200 (slink had 3.3.2), so I just grabbed the new 3.3.3 X SVGA binary from xfree86.org and copied it over the old binary. Then I ran the config tools and told them I had a Matrox Millenium (which would use a similar setup to a G200) and everything worked fine. You will run into problems doing this trick if you choose a card which is drastically different from your Savage4 though, such as picking a card with a weird RAMDAC or something. Your best bet would be to find out if X 3.3.5 is available as a set of Debian packages from somewhere, and upgrade using apt. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpwg1YFE7N4u.pgp Description: PGP signature
Re: Purging mozilla
On Wed, Oct 20, 1999 at 11:54:16PM +0200, David Jardine wrote:
> It appears that mozilla creates a subdirectory in the user's home
> directory with the user's name, eg /home/fred/fred. I ran
> which according to my understanding of the
> manpages should eliminate all traces of its existence - it didn't
> say that exactly, but this seemed the most radical option.
>
> However, the directories are still there, so:
> Should I have done it another way?
> Has mozilla left any other droppings on my system?
> If there is no other way to clean up, shouldn't there be?
dpkg has no way of knowing about files the program creates after
installation, so it can't remove them. That means that any config files
which are created and stuck in your $HOME are going to stay there even
when you remove the package that created them...
If you take a look in /var/lib/dpkg/info/, the *.list files tell dpkg
what files belong to what package, and therefore, what files dpkg can
remove when removing the associated package.
Aside from the actual difficulties of knowing about files created by a
package after installation, I can think of plenty of reasons why having
dpkg remove such files would be really evil.i It *could* be done, though
- if the package maintainer, for example, made the {pre,post}rm scripts
look for known files and kill them (e.g. the script for mozilla could
search the /home of every user for it's .mozilla dir or something).
--
[ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ]
[ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ]
pgp7p6iSlLzhT.pgp
Description: PGP signature
Re: ensonique?
On Thu, Oct 21, 1999 at 11:51:48AM +0530, T.V.Gnanasekaran wrote: > can i purchase ensonique cards? are they supported very well in 2.2.x > kernels? what other 32-bit cards are very well supported and cheap. > ;) If you mean Ensoniq, then the answer is "probably". I don't know the full range of cards that they make, but the ES1370/1371 based cards are supported in Linux, both in the kernel sound drivers, and in ALSA. The ES1370 card is called Ensoniq AudioPCI. The ES1371 is the Ensoniq AudioPCI'97. I actually thought that Creative Labs bought Ensoniq, and were making their cards under different names now (or does that only apply to the AudioPCI range?). The Creative Labs cards using the ES1370/1371 chips are the Soundblaster PCI 64 and the PCI 128. Your best bet is to find out more about the card, and check out www.alsa-project.org's list of supported cards. If you want well supported cards that are made by Linux friendly people, I think the Trident cards are what you're after, if I remember correctly. Personally, I'm using ALSA and a Soundblaster PCI 64 (ES1370 chipset). It's adequate for what I use it for, and it was cheap. It doesn't sound as good as my Yamaha Waveforce YMF724 did (which was only half the price, too), but Yamaha aren't giving out the documentation for that chipset, so we won't be seeing support for it anytime soon. :-( -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpVhAth7xIqF.pgp Description: PGP signature
Re: New Kernel: make-kpkg vs. make bzLilo
On Wed, Oct 20, 1999 at 07:41:20PM -0400, Salman Ahmed wrote: > One other thing: are there any problems compiling 2.2.12 or 2.2.13 with gcc > 2.95.2 ?? I read today in the 2.2.13 kernel's release notes at > <http://www.linux.org.uk/VERSION/relnotes.2213.html> that gcc 2.95 should > not be used to compile 2.2.13. > > Has anyone tried compiling the recent kernels (ie 2.2.10-12) with gcc 2.95 > or 2.95.2 ? Any problems ?? I build all of my kernels with whatever the latest EGCS/GCC in potato is, so I'm currently using gcc 2.95.2. I haven't found any major problems doing this. For the record, I also have gcc 2.7.2.3 installed for building stable/bug-report kernels. Basically, you can use whatever you like to compile your kernels, but if you have problems and want to send a bug report to the kernel developers, don't bother unless the kernel was built with gcc 2.7.2. The developer will most likely tell you to go away and rebuild the kernel with that version of gcc and retest it before they'll listen to your problems. There is definately a problem building kernels older than 2.2.10/2.2.11, and you'll need to add -fno-strict-aliasing to the CFLAGS in the makefile to get the kernel to build with gcc 2.95. This is because between egcs and the changeover to gcc 2.95, the default option for strict aliasing was changed from off to on, and the kernel won't built with it on. > I would like to try compiling the new 2.2.13 on my Celeron using gcc > 2.95.2. Go for it. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpOQ6Bzx0N1v.pgp Description: PGP signature
Re: Why isn't my ftp working?
On Sat, Oct 23, 1999 at 10:02:54AM -0400, Aaron Solochek wrote: > When I put together my potato partition, I noticed that I was no longer > able to ftp to my machine. I actually wasn't able to do much. But > I copied my slink config files over, and that helped most services. So > could someone tell me where ftpd starts from? And, is there any > encrypted ftpd that I can use, other than kftp? About 2-3 weeks ago, ftpd was split out of netbase. If you upgraded your install of potato sometime after then, you won't have ftpd anymore. You can reinstall that particular ftpd (the one from netbase) by installing the ftpd package. Or you can go with something else, like wu-ftpd* or proftpd, depending on your needs. As for where ftpd starts from, it's normally started by inetd when it's needed. Check out /etc/inetd.conf and look for an entry containing ftpd. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpb7Vk3vMcgd.pgp Description: PGP signature
Re: segmentation faults
On Sat, Oct 23, 1999 at 08:39:19PM +, John wrote: > Have been unable to find any definition and write-up on > 'segmentation faults'. I have encountered this whilst trying > a package installation as part of the learning process - > the matter is not critical, but nontheless I would like to > understand what is involved in case I meet this error > when it is important. Basically it means the program tried to access memory it doesn't have access to. The kernel will see this happen, and kill the program (by sending it a SIGSEGV signal (SEGV = segment violation)). I'm sure someone else can give you a better description that mine. :-) > Could anyone please point me towards a source of > information suitable to a novice. I can't think of anything off-hand, sorry... -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpyL9W7vL3nu.pgp Description: PGP signature
Re: hdparm settings
On Tue, Oct 26, 1999 at 04:02:20PM +1300, [EMAIL PROTECTED] wrote: > i have set a few hdparm settings to tune up my ide drive.. i however > do not know how to make those settings permanent as setting switch -k1 > does not seem to work. Should i just set the settings everytime my > machine is rebooted in an init script or something?? The -k1 switch will only keep settings over a reset of the drive controller, not of the machine. To make hdparm settings "permanent", write a little shell script, put it in /etc/init.d and use update-rc.d to have it symlinked to the correct places in the /etc/rc.* directories. You'll want to check out the man page for update-rc.d first, if you haven't used it before. An example: # update-rc.d -n hdparm defaults This will show you what changes it's going to make, assuming your script in init.d is called 'hdparm'. If it the settings look good, remove the '-n' and run again to actually make the links. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint:FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpGRMmxLKhYG.pgp Description: PGP signature
Re: manual kernel recompile vs make-kpkg
On Tue, Oct 26, 1999 at 05:48:51PM -0500, Charles Lewis wrote: > Install debian kernel source using dselect (not sure how to do patches) To do a patch: # cd /usr/src/kernel-source-2.2.12 # zcat ~/patch-2.2.13.tar.gz | patch -sp1 That'll do the trick. Removing a patch is exactly the same, except you give patch the -R option as well... Something to watch out for is that some patches are incremental (e.g. 2.2.12 -> 2.2.13 -> 2.2.14) whereas others need to be removed and patched with a newer version (the ac patches are an example of this). > cd /usr/src > tar zxpvf kernel-source-2.2.12.tar.gz > cd kernel-source-2.2.12 You might want to link your current kernel source tree to /usr/src/linux: # cd /usr/src # ln -sf kernel-source-2.2.12 linux > make menuconfig (or xconfig) > make dep (make clean??) > make bzImage > make modules > make modules_install > cp /usr/src/kernel-source-2.2.12/arch/i386/boot/bzImage /boot/vmlinuz-2.2.12 > rm /vmlinuz > ln -s /boot/vmlinuz-2.2.12 /vmlinuz > (I learned the hard way that when I copied bzImage directly to /vmlinuz, I > was overwriting the previous kernel in /boot) You'll want to copy the System.map to /boot as well. # cp /usr/src/kernel-source-2.2.12/System.map /boot/System.map-2.2.12 > Now, my question is, is what I do above pretty much what everyone else does? > And what would be the equivalent steps for doing the same thing using > make-kpkg? After reading the man pages, I'm not sure what options I should > use or what the resulting deb would be. This is what I've been able to glean > so far: > > make menuconfig > make-kpkg clean?? > make-kpkg --revision 1.0 --bzimage kernel_image You might want to use something like --revision custom.1.0 or something, since you run the risk of apt overwriting your kernel package with a newer official Debian one (when one becomes available), which is probably not what you want. Unless you're doing this as root (bad), you'll need something like fakeroot to finish the kernel package making process, since you need root access to build a package. What I normally do is what you've done, and then when it fails with an error about not having root access: # fakeroot make-kpkg kernel_image And it'll build up the package for you. > dpkg -i ??? This looks right, although you don't need the --bzimage, since that's the default anyway. The .deb you install is left just outside of your kernel tree, (in your case, it would be in /usr/src). kernel-image-*.deb will be the name of it... > Another question is what exactly is the difference between bzimage and > zimage, and which one should I be using? bzImage is "big" zImage, and is able to load larger kernels (>1MB), whereas zImage is limited to something smaller (600kB?). You want to use bzImage all the time, unless your machine doesn't like booting such images (e.g. some laptops will only boot zImage kernels). -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgp1c0SRABOMN.pgp Description: PGP signature
Re: manual kernel recompile vs make-kpkg
On Wed, Oct 27, 1999 at 08:41:47AM +0200, Onno wrote: > >You'll want to copy the System.map to /boot as well. > ># cp /usr/src/kernel-source-2.2.12/System.map /boot/System.map-2.2.12 > > Would you elaborate on that please... I'll try, but I don't know much about the System.map file, so what I tell you could be wrong - you have been warned. :-) The System.map file is a "map" of the symbols in the kernel, and their location (I think, if you look at the file it has what looks like addresses and then a symbol name in the third column (I'm not sure what that other character means, the t/T in the second column)). Some programs (top, ps, etc.) can use the System.map to discover and display extra information... If you have the wrong System.map file for the current kernel, when you start those programs, they'll complain. Most/all of the programs which use the System.map now look for it in the following order: /boot/System.map- /boot/System.map I think it looks in at least one other place, but I don't remember where that is (sorry). I haven't actually come across any information about the use of the System.map so far, which is why I don't really know a great deal. Maybe it's one of those things which you know the purpose of it if you truly have a use for it. ;-) The file isn't critical to the running of a system, so if you don't bother updating it, I don't think anything Really Bad (TM) will happen. Oh, if you use make-kpkg to build kernels, when you install them with dpkg the updating of the System.map is handled as part of that, so you don't have to worry about it. If someone can post a better explanation, or a pointer to some information, I'd be very grateful... -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpM6TI9carJ8.pgp Description: PGP signature
Re: Debian Linux vs BSD
On Thu, Oct 28, 1999 at 01:30:19PM -0400, Paul McHale wrote: > There is one question. They announce openBSD ships with a secure > version of ProFTP. The version appears to be older than the bug > version(s). Is there something inherently different about BSD that it > was not affected by the bug ? Where is this announcement, on the OpenBSD website? I don't know what the story is with ProFTPD in OpenBSD in regard to security, but recently on the OpenBSD mailing list, Theo de Raadt (leader of the project), stated that ProFTPD won't be secure without a complete rewrite... I'm not sure if the version in the OpenBSD distribution has been audited by the team or not, though. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpZfoiiZB3v9.pgp Description: PGP signature
Re: Need a runlevel editor! (console ofcourse)
On Fri, Oct 29, 1999 at 04:58:50AM -0800, Ethan Benson wrote: > On 29/10/99 Martin Fluch wrote: > > >Ther is a script called update-rc.d ... > > no this is not what we are looking for, update-rc.d only works if > there is no symlinks at all for a given script, so to use it to > change a runlevel you must first rm all the symlinks then use > update-rc.d to recreate them in the configuration you want, not that > much more convenient then using mv... Look at the man page for update-rc.d: -f Force removal of symlinks even if /etc/init.d/name still exists. This sounds exactly like what you need... -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgp1qbebEuaHT.pgp Description: PGP signature
Re: Debian Linux vs BSD
On Sat, Oct 30, 1999 at 10:44:54AM -0400, Paul McHale wrote: > This was my mistake. There is a link on the openBSD site to "ports". This > link is to the xBSD general repository which I mistakenly thought was the > openBSD repository. The ProFTP program is part of the general repository. > Sorry for the confusion. That's probably the OpenBSD specific ports you're referring to. The security status of them is a little hazy to me, since I haven't seen anything which states they've been audited for security like the base OpenBSD has, but I can't say for sure they haven't been, either. > I also assume this means that openBSD is more secured as long as what you > need comes with openBSD as part of their closer reviewed distribution. > Installing anything else would presumably cause the same bugs under openBSD > as it would under freeBSD. This is true, your security will only be as strong as the weakest software you have running... OpenBSD actually comes with an FTP daemon (which is used as the basic ftpd for Debian, it used to be part of the netbase package), I don't think it's as feature-filled as ProFTPD, but it's presumably a lot more secure. > openBSD code review must have been quite an impressive effort to say the > least... I think it took them about 1.5-2 years all up, with the majority of the problems being found and fixed in the first 6 months. Very impressive indeed, I think they have a claim to "no remote root exploits" since the audit now. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpqJCIB9dNgz.pgp Description: PGP signature
Re: Need a runlevel editor! (console ofcourse)
On Sat, Oct 30, 1999 at 04:36:58AM -0800, Ethan Benson wrote: > >Look at the man page for update-rc.d: > > > >-f Force removal of symlinks even if /etc/init.d/name still > > exists. > > yes I read the man page, yes I tried this, no it didn't work :| It may be that you have to do something like: update-rc.d -f remove ...before you remake the links with update-rc.d. Not great, but still more convenient than having to rm them individually. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgplHdOCwmzs0.pgp Description: PGP signature
Re: package
On Mon, Nov 01, 1999 at 01:16:29PM +0700, Kreshna D. Armand wrote: > How do I know the size of each packages which I want to install? If I > use SuSE or RH, you can query it. If I use Debian, by what way I can > do the query of the .deb packages? If you're using potato (or you feel like recompiling) grab the console-apt package - it shows both the package size and the installed size of each package on the main screen. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgplEjrjnWhNY.pgp Description: PGP signature
Re: kernel upgrade options
On Tue, Nov 02, 1999 at 05:18:00PM +, John wrote: > Kern.log says the kernel is not compiled for ppp (I've not used any > modules). > > My limited experience indicates, therefore, it must have to do with > the kernel - but what? I enabled 'networking support' which seems the > only relevant option. I've re-run 'make config' (good experience for a > new- comer) and found nothing helpful. The other distributions connect > without difficulty, so presumably it is not hardware. You need to enable the options for PPP, either compiled in or as a module. If you're using menuconfig, you'll find them under "Networking Devices". If you compile it as a module you'll need to load it before you try and use PPP. > When I do get ppp working and go for patching, are there any apart > from 8 to 12 incl that do not need to be compiled (between 2 and 13 > that is?). You don't need to compile any of them as you patch them up, you can if you want to test each patch, but otherwise just apply all of the patches and then do your compiling. make oldconfig is useful once you've applied patches, since it'll only ask you config questions about new options... -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgp8CiTr2Ye4k.pgp Description: PGP signature
Re: kernel upgrade options
On Thu, Nov 04, 1999 at 06:07:31PM +, John wrote: > I've obtained patches 2 to 5 but now have a problem in applying them. > My /usr/src had only 'kernel-sources-2.2.1' (and the .tar.gz source which > I'm leaving there for the time being) until I moved 'patch-2.2.xx' from the > download site. > > When I do 'zcat patch2.gz | patch -p0' I get:- Cd into your kernel source dir, and use '-p1' instead. > I discover the quoted lines are the first three text lines of the patch, > and the fourth is I presume machine instructions viz, > @@-1403,6 +1403,13@@ These numbers are actually line numbers - it's telling patch the general location of where to look in the file. I think the format is like this: - look at line 1403 - 6 lines are to be removed - look at the new line 1403 - 13 lines are to be added Someone please correct me if I messed that up. :-) > Can anyone help - I feel I understand what is required generally, > but not specifically. Is it that I don't have a directory called > Linux in /usr/src? (everything I read seems to assume I have) > I don't mind messing-up things and having to reinstall, but it > wouldn't help I'd still not be able to apply patches. You could rename your kernel-source-2.2.1 to whatever the patch is expecting and run patch again with the arguments you used, but you are better off to cd into the the dir and use '-p1', since everybody seems to have different names for their top level directory. :-) -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpMSKevHIqLc.pgp Description: PGP signature
Re: kernel upgrade options
On Fri, Nov 05, 1999 at 05:57:58PM +, John wrote: > I couldn't get this to work. My patches were in /usr/src and I got 'no > such file or directory', so I moved them into the kernel source dir. > and did -p0 thinking that should be right. No luck, so I did -p1 - > there was action, but only 33 times 'Hunk xx failed at ' and > quite a few 'succeeded' at with fuzz 1 or 2 and a final line > 'patch: malformed patch at line 2121: s'. Okay, it sounds like your kernel source isn't a pure one (I'm guessing here). Patch is quite smart, and can patch files even if they're different to what it's expecting, although it'll only work if they're a little different, not a huge amount. It might be easier to download a fresh kernel source tarball from one of the kernel.org mirrors and try from there. You could download the latest, so you don't have to do any patching - but if you want to learn to patch, download an older one and apply the patches you've got to it. The only problem is a kernel source tarball is about 13MB. I'm pretty sure the source you extracted from the Debian package should be a standard kernel though, so maybe there's something messed up. rm the source dir and extract it again, then try patching it... > After looking at the .rej's, which were enlightening but not helpful > with the problem, and not finding anything else to guide me, I tried > to apply patch 3 expecting some message indicating it was out of > order. It, in fact, worked partially and said '2 out of 11 hunks > failed'. Emboldened, I did patch 4 ( '1 out of 2 failed') and then 5 > which seemed OK except for a final 'malformed patch at line 146'. The problem with patch is if it fails, generally you can consider that kernel source trashed and kill it and start over. If you really know what you're doing, the above isn't true, but for most people it's far easier to just start over than try and reverse the failed patches and find out what's gone wrong... If you've only applied one patch and it failed, then you're pretty safe to reverse it (option -R). You could probably reverse all of the patches this way if you were patient, but it's easier to just reextract the source and start over. > My problem here is that I do not know what the patch is expecting. If > I renamed to 'Linux-2.2.1' would it affect other things, maybe > creating more difficulties? Normally the patches will expect /usr/src/linux to be your kernel source, but some patches will look elsewhere instead. It all depends on the struct of the filesystem where it was built. But that's what the '-p' option is for - to tell patch your structure is a little different. > Another mystery to me arises from the fact that the command lines > suggested in all the literature I've seem do not appear to say what is > to be patched. Yet, according to O'Reilly's Linux in a Nutshell it's > based on 'patch [options] [original [patchfile]]. That's for patching single files, e.g.: # patch hello.c hello-fix.patch Which will apply hello-fix.patch to the hello.c source. With the kernel sources (and most patches) you're applying it to a whole bunch of files. Inside the patch file is where patch finds out which files it's supposed to modified, all you need to do is be in the right directory and run patch. The reason we are piping the output of zcat to patch is two-fold. Firstly, it needs to uncompress the file so patch can read it, and you need to send the patch file to patch from STDIN for patches with multiple files. If the patch wasn't compressed, you could do something like: # patch -p1 < patchfile > Any suggestions where I should stumble next? Will the patches still be > OK after what I've tried so far. My installation is appears undamaged > - seems indestructible. I'd be pretty surprised if the kernel compiles now, since some parts of the patches you applied didn't work, so the kernel source is in quite a messy state now... If you do something like zcat patchfile | patch -Rp1 in *reverse* order from the order you applied them, it will try and remov the patches you've applied. At the end you should have a copy of the 2.2.1 sources exactly like you started with. > Grateful for any help. I could, despite the online time involved, > download 2.2.13 but would learn nothing further here and would have to > face the problem later 2.4 if I want to keep up to date. You could download something a wee bit earlier, like 2.2.9 and patch it up from there. It'll be quite tedious to apply all of the patches from 2.2.1 up to 2.2.13, which is what you're going to have to do otherwise... That way you get a 'clean' kernel source, you get to learn how to use patch, and you get an up-to-date kernel at the end of it. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgp9Qiul5iHzi.pgp Description: PGP signature
Re: Applying kernel patch
On Thu, Nov 11, 1999 at 09:56:11AM -0500, [EMAIL PROTECTED] wrote: > > I'm not sure how or where to find information about applying a kernel > patch. Can someone point me in the right direction? I have the kernel > sources, and the appropriate kernel patch deb, but do I apply the > patch before configuring and recompiling the kernel, or after? Guess > I'm not sure what the patch actually does. > Check out the Kernel HOWTO, that has a section on kernel patches, using the patch program, and things to be aware of when using patches. Also check out the debian-user archives on debian.org, there are quite a few threads stored there relating to kernel patches. -- [ Matthew Gregan ] [ GPG ID: B63A1E95 ] [ [EMAIL PROTECTED] ] [ GPG fingerprint: FB83 2911 F170 B31C 9E4A E382 CA8A A2F6 B63A 1E95 ] pgpqoGnreclqS.pgp Description: PGP signature
Re: view output of process while
On Wed, Apr 24, 2002 at 01:22:18PM +, Rory Campbell-Lange wrote: > Is it possible to log back into a machine over ssh and 'attach' to a > running process so that one can see the output of, say, a kernel > compile? Take a look at the ``screen'' package. -- Matthew Gregan |/ /|[EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
