Re: Finding left-over libraries

[Keith Harbaugh]

On Tue, 1999-12-07 at 23:13:32 +0100, peter karlsson wrote:
> Is there any way in Debian to find out what packages no other packages
> depend on?
> 
> When I install a couple of packages, all the libraries they depend on are
> installed as well, which is quite nice, but the reverse doesn't hold - when
> I remove packages, unnecessary libraries are not removed. Because of this, I
> would like to get a list of packages that no packages depend on (restricted,
> for instance, to packages starting with lib).

You might want to check out the package pkg-order,
in particular, its programs pkg-deptree and pkg-revdeps.

Keith

Re: Trying to install X-windows

[Keith Harbaugh]

On Sat, 1999-12-04 at 17:57:00 -0700, csager wrote:

== snip ==

> I have run XFR86Config, but when I type the command:
> startx
> I get so many error messages that scroll by my screen so
> fast that I don't know what all the problems are.
> Is there a way to get all of the error message into a file
> so that I can make a reasonable attempt at solving my
> problem? Or possibly a way to step through the errors so
> that I can at least write them down?  Are there any packages
> I missed?

They may not be all error messages.
The X server, when it starts up, spits out a large number of status messages.
The precise number depends on your X configuration (e.g., number of modes);
I get roughly 75 when I start up, with only 1 mode.

If you use startx or xinit these messages go to the originatng virtual console;
if you use xdm, they go to /var/log/xdm.log
(this is set in the configuration file /etc/X11/xdm/xdm-config).

Keith

Re: Clock problems

[Keith Harbaugh]

On Mon, 1999-11-29 at 15:28:57 -0600, Marc Mongeon wrote:

> I think this is caused by the file /etc/adjtime, which is supposed to
> adjust for clock drift, but gets skewed when you first set the hard-
> ware clock.  Remove the file, then re-set the clock.  It will be re-
> created as needed.
> 
> Marc
> 
> --
> Marc Mongeon <[EMAIL PROTECTED]>
> Unix Specialist
> Ban-Koe Systems
> 9100 W Bloomington Fwy
> Bloomington, MN 55431-2200
> (612)888-0123, x417 | FAX: (612)888-3344
> --
> "It's such a fine line between clever and stupid."
>-- David St. Hubbins and Nigel Tufnel of "Spinal Tap"
> 
> 
> >>> Timothy Bedding <[EMAIL PROTECTED]> 11/29 3:05 PM >>>
> I am having problems with the clock losing time after
> I have shut my system down overnight.
> 
> Is this to be expected?
> 
> I thought that there would some way round this.
> Or is the only way simply to keep the machine running
> all the time?
> 
> Cordially
> Tim
> 

Yes, but as well you *must* comment out the line in /etc/init.d/hwclock.sh
which starts
   hwclock --adjust

See also bug report 41263.

Keith

X font and chown problems in potato

[Keith Harbaugh]

This basically reprises a message sent on 1999-11-07 at 06:10:18 +,
Subject: Cannot chown /dev/pts/0 to 0,0; likewise for /dev/ptyxx,
which received no response.

On starting xinit, the xterm it brings up only displays characters
as solid blocks of color the foreground color.

When the xterm is exited,
the normal X server status messages are visible on the virtual terminal
from which the xinit was launched, followed by these lines:

   Cannot chown /dev/pts/0 to 0,0: No such file or directory
   Cannot chown /dev/ptyxx to 0,0: No such file or directory

   waiting for X server to shut down 



On the other hand, when xdm is started,
lines in the xconsole which extend all the way to its right hand side
display legibly in the correct font,
while the remaining lines which stop before the right edge
have their "characters" displayed as solid blocks.
Likewise, the login display shows only solid blocks for its characters.

Please see the original message for configuration details;
but these problems have persisted without change from X 3.3.5-1 to 3.3.5-2.

Keith

Re: [potato] broke ps printing (still)

[Keith Harbaugh]

On Sat, 1999-11-20 at 17:31:37 -0800, Ron Farrer wrote:
> Hello all;
> 
> I upgraded to gs-aladdin from gs as one person suggested, but I still
> can't print ps. Basically nothing happens, lpq shows:
> 
> Printer: [EMAIL PROTECTED]  'Epson Stylus Color Pro'
>  Queue: no printable jobs in queue
>  Status: job '[EMAIL PROTECTED]' removed at 17:31:15.040
> 
> Any ideas?

Are you using lprng?
If so, it maintains a log file in /var/spool/lpd/lp.
Check and see if it has a JABRT (whatever that is) error message.
When I switched to potato I encountered both your problem and that error.
Deinstalling lprng and installing lpr cured both.

Keith

ptys and /dev/pts

[Keith Harbaugh]

I've just switched to potato, and still don't have it entirely working
(mainly in that X can't find its fonts),
but while I am waiting to figure out a solution to that,
here's two small questions:

Can anyone explain the following behavior:

$ tty
/dev/tty0
$ script
Script started, file is typescript
$ tty
/dev/pts/0


And, what is a good reference to all this pty and pts stuff?

Thanks,
Keith

Clock mini-HOWTO

[Keith Harbaugh]

The subject of "how to adjust your clock" is a real FAQ on this list.

I just ran across a "Clock mini-HOWTO" at .
It seems to do a really excellent job of explaining the relations
of the various hardware and software components
concerned with time maintenance.
I think it may shed some light into what still often seem
dark corners of this subject.

(BTW, apparently slink's xntp3 has been split into potato's ntp and ntpdate.)

Experts feel free to jump in and criticize...

Keith

Re: What the hell happened to X on Monday?

[Keith Harbaugh]

On Tue, 1999-11-09 at 23:18:03 -0600, Erick Kinnee wrote:
> I seem to have lost the 'fixed' font or whatever it's aliased to. X no
> longer starts for me, and I have heard reports from other users of this.
> Anybody got an idea?
> 

I've had no luck with X under potato: see post late on 1999-11-06,
Subject: Cannot chown /dev/pts/0 to 0:0 ...

What happens when you just try xinit
(assuming you've gotten xdm out of the way, with /etc/init.d/xdm stop
or otherwise)?

Keith

dselect logging

[Keith Harbaugh]

One may easily log an apt-get session,
by starting a script beforehand, then using the --quiet option to apt-get
to suppress the periodic progress reports on downloads.

But when I try to operate dselect within a typescript situation,
all the cursor control characters (ncurses or whatever),
while they do great things on a video terminal,
just garbage up the potentially printed typescript;
also, there still are the unwanted progress reports.

Am I missing some simple dselect option that would enable a nice log
of all the configuration action that takes place during an install?

And, the same question for console-apt (capt).

Thanks,
Keith

Re: Can't find a valid termcap file at .../Readline.pm

[Keith Harbaugh]

On Tue, 1999-11-09 at 13:44:06 -0800, Joey Hess wrote:
> Keith Harbaugh wrote:
> > Can't find a valid termcap file at /usr/lib/perl5/5.005/Term/ReadLine.pm 
> > line305
> > 
> > What do I need to (re)install to cure this,
> > without (further) breaking my system?
> 
> Your system isn't broken at all, but installing libterm-readline-gnu-perl
> will make this warning go away.

YES!!!
That one simple install *does* cure the problem.

Thanks for your all your good work,

Keith

Re: printing specific pages

[Keith Harbaugh]

On Wed, 1999-10-20 at 23:29:36 +0200, Jean-Yves BARBIER wrote:
> On Wed, Oct 20, 1999 at 10:43:12PM +0300, Paul Huygen wrote:
> > Jean-Yves BARBIER <[EMAIL PROTECTED]> wrote:
> > 
> > > I did not found anything about printing only certain pages. Is it 
> > > possible?
> > 
> > Probably. How to do it depends on the format of the pages that you
> > want to be printed. Are the pages in Postscript, TeX, Troff, Word,
> > Wordperfect, or something else?
> > 
> > Paul Huygen
> 
> I'd like to print in all these formats!


For TeX, dvips has -A and -B options for printing odd and even pages resp,
and -pp first-last for page ranges.
Just typing `dvips' gives a quick summary, while the whole story is in a
info/texinfo document.

For Postscript, psselect was already mentioned,
but the interactive viewer gv also supports both odd and even
markings/selections, as well as marking for printing individual pages.


Keith

Cannot chown /dev/pts/0 to 0,0; likewise for /dev/ptyxx

[Keith Harbaugh]

The slink -> potato saga continued (Chapter X):

Doing xinit on my upgraded potato system yields an xterm without fonts,
that is, where the characters should be, there are just colored boxes.
However, the messages from the X server all look reasonable;
the font path is shown as:
(**) FontPath set to 
"/usr/X11R6/lib/X11/fonts/local,/usr/X11R6/lib/X11/fonts/misc:unscaled,/usr/X11R6/lib/X11/fonts/75dpi:unscaled,/usr/X11R6/lib/X11/fonts/100dpi:unscaled,/usr/X11R6/lib/X11/fonts/Type1,/usr/X11R6/lib/X11/fonts/Speedo,/usr/X11R6/lib/X11/fonts/misc,/usr/X11R6/lib/X11/fonts/75dpi,/usr/X11R6/lib/X11/fonts/100dpi"

Then, after a Ctrl-D exits the xterm and kills the X server,
these messages appear:

Cannot chown /dev/pts/0 to 0,0: No such file or directory
Cannot chown /dev/ptyxx to 0,0: No such file or directory


Doing ls -l /dev/pts/0 yields:
crw--w1 root tty  136,   0 Nov  7 05:45 /dev/pts/0
while ls -l /dev/ptyxx yields:
ls: /dev/ptyxx: No such file or directory


The system context for these problems is a potato system upgraded as of today,
with X, xfs, xterm, et al, of vintage 3.3.5-1,
and a Linux kernel 2.2.13, whose .config includes the entries:

CONFIG_UNIX98_PTYS=y
CONFIG_UNIX98_PTY_COUNT=256
CONFIG_DEVPTS_FS=y

Before upgrading, the system was a slink one,
using X fonts 3.3.2.3a-11 and active X software 3.3.3.1-2 from netgod;
all that worked fine.

Any thoughts?

Keith

Can't find a valid termcap file at .../Readline.pm

[Keith Harbaugh]

I used to have a slink system which was running basically just fine,
then in a moment of Neanderthal adventurism
decided to try to "upgrade" to potato in one fell swoop.
Well, I got swooped all right, and have been trying to dig myself out
from a hole of unknown depth ever since.
I'll spare the boring or exciting (depending on your mood) details,
and focus on one very specific remaining problem:
I had trouble running debconf in "gtk" mode,
so decided to back off and settle for just "text" mode in debconf.
But now whenever I do an 'apt-get install whatever',
I get the error message

Can't find a valid termcap file at /usr/lib/perl5/5.005/Term/ReadLine.pm line305

What do I need to (re)install to cure this,
without (further) breaking my system?

Keith

Re: updating of CMOS clock

[Keith Harbaugh]

On Sat, 1999-11-06 at 14:36:25 -0500, Salman Ahmed wrote:
> Whenever I reboot/shutdown my debian system, one of the message I see on
> the console during shutdown is :
> 
>   CMOS clock updated to Sat Nov  6 14:25:38 EST 1999
> 
> This alongwith the fact that APM support (which I have compiled into my
> 2.2.13 kernel) seems to reset the time on every reboot is likely the
> reason why my clock seems to be losing time gradually and every couple
> of weeks totally loses it!!!
> 
> I checked my RedHat5.2 installation for this CMOS clock updating on
> reboot/shutdown and RH5.2 is not doing this. Prior to using Debian-2.1,
> I had been using RH-5.2 with APM support compiled into 2.2.10 and I
> never had any problems with my clock losing time. Only after switching
> to 2.2.10 (with APM support) on Debian-2.1 did I notice this constant
> and consistent problem with my CMOS clock.
> 
> My questions are:
> 
> (1) is it possible to configure Debian so that it DOESN'T update the
> CMOS clock whenever the system is shutdown/rebooted ?
> 
> (2) are there any tools for Debian similar to timetool and timeconfig on
> RedHat which allow the setting of system time from a GUI and ncurses
> interface respectively ?
> 
> (3) How do I change my system's configuration so that the CMOS clock
> stores time in GMT ? This is supposed to be a better way to store time
> in the CMOS clock but I can't seem to find the option in my system's
> BIOS. (FYI, I am using an ABIT BH6 MB).
> 
> A number of people on this list have already recommended the use of
> packages like ntpdate, xntp, etc. but I would first prefer to solve this
> problem by having Debian not update my CMOS clock on every
> reboot/shutdown. After that, I will look into setting up xntp, etc. to
> sync time/date whenever I dial out to my ISP.
> 
> My system is a slink/potato salad.

(1) Edit /etc/init.d/hwclock.sh to comment out the command
hwclock --adjust.

(3) Edit /etc/defaults/rcS as appropriate (it changed a little from
slink to potato).

These solved, in my case, the problems you outlined, but of course YMMV.
I have no idea re (2).

Keith

System halted (Linux 2.0) versus Power down (Linux 2.2)

[Keith Harbaugh]

Can any kernel gurus out there explain why the final message the kernel
gives upon system shutdown changed from
System halted
in Linux 2.0 to
Power down
in Linux 2.2?

For specificity here is the relevant code from kernel/sys.c for 2.2.12
(pardon me for sending this through the mail, but it will help to make
the discussion more concrete):

/*
 * Reboot system call: for obvious reasons only root may call it,
 * and even root needs to set up some magic numbers in the registers
 * so that some mistake won't make this reboot the whole machine.
 * You can also set the meaning of the ctrl-alt-del-key here.
 *
 * reboot doesn't sync: do that yourself before calling this.
 */
asmlinkage int sys_reboot(int magic1, int magic2, int cmd, void * arg)
{
char buffer[256];

/* We only trust the superuser with rebooting the system. */
if (!capable(CAP_SYS_BOOT))
return -EPERM;

/* For safety, we require "magic" arguments. */
if (magic1 != LINUX_REBOOT_MAGIC1 ||
(magic2 != LINUX_REBOOT_MAGIC2 && magic2 != LINUX_REBOOT_MAGIC2A &&
magic2 != LINUX_REBOOT_MAGIC2B))
return -EINVAL;

lock_kernel();
switch (cmd) {
case LINUX_REBOOT_CMD_RESTART:
notifier_call_chain(&reboot_notifier_list, SYS_RESTART, NULL);
printk(KERN_EMERG "Restarting system.\n");
machine_restart(NULL);
break;

case LINUX_REBOOT_CMD_CAD_ON:
C_A_D = 1;
break;

case LINUX_REBOOT_CMD_CAD_OFF:
C_A_D = 0;
break;

case LINUX_REBOOT_CMD_HALT:
notifier_call_chain(&reboot_notifier_list, SYS_HALT, NULL);
printk(KERN_EMERG "System halted.\n");
machine_halt();
do_exit(0);
break;

case LINUX_REBOOT_CMD_POWER_OFF:
notifier_call_chain(&reboot_notifier_list, SYS_POWER_OFF, NULL);
printk(KERN_EMERG "Power down.\n");
machine_power_off();
do_exit(0);
break;

case LINUX_REBOOT_CMD_RESTART2:
if (strncpy_from_user(&buffer[0], (char *)arg, sizeof(buffer) - 
1) < 0) {
unlock_kernel();
return -EFAULT;
}
buffer[sizeof(buffer) - 1] = '\0';

notifier_call_chain(&reboot_notifier_list, SYS_RESTART, buffer);
printk(KERN_EMERG "Restarting system with command '%s'.\n", 
buffer);
machine_restart(buffer);
break;

default:
unlock_kernel();
return -EINVAL;
break;
};
unlock_kernel();
return 0;
}

So why does shutdown branch through ...POWER_OFF rather than ...HALT?
At least under a Debian 2.1 (slink)/Linux 2.2.12 combination.

Keith

Re: Netscape 4.71 Is Rock Solid & Fast! [gets rid of libc5 too!]

[Keith Harbaugh]

On Tue, 1999-10-12 at 17:50:14 -0500, John Foster wrote:
> I just installed the new Netscape 4.71 version on my Debian Linux
> server. Just a report- It installs nicely with the Netscape4 installer
> from Debian if you rename it to the proper convention. It seems MUCH
> faster and more stable that all of the previous versions I have used.

Sorry, one other item of interest I should have added to earlier post
on [details of how to get] was that 4.7 as opposed to 4.51 only requires
the current slink glibc-2.0 libraries!  Note the following:

p1 $ ./netscape.old --version
Netscape 4.51/Export, 27-Feb-99; (c) 1995-1998 Netscape Communications Corp.
p1 $ ldd ./netscape.old
libXt.so.6 => /usr/lib/libc5-compat/libXt.so.6 (0x4000b000)
libSM.so.6 => /usr/lib/libc5-compat/libSM.so.6 (0x4004d000)
libICE.so.6 => /usr/lib/libc5-compat/libICE.so.6 (0x40056000)
libXmu.so.6 => /usr/lib/libc5-compat/libXmu.so.6 (0x4006b000)
libXpm.so.4 => /usr/X11R6/lib/libXpm.so.4 (0x4007d000)
libXext.so.6 => /usr/lib/libc5-compat/libXext.so.6 (0x4008c000)
libX11.so.6 => /usr/lib/libc5-compat/libX11.so.6 (0x40097000)
libdl.so.1 => /lib/libdl.so.1 (0x40135000)
libc.so.5 => /lib/libc.so.5 (0x40138000)
libg++.so.27 => /usr/lib/libc5-compat/libg++.so.27 (0x401f6000)
libstdc++.so.27 => /usr/lib/libc5-compat/libstdc++.so.27 (0x4022e000)
libm.so.5 => /lib/libm.so.5 (0x4025f000)
p1 $   
p1 $ ./netscape --version
Netscape 4.7/Export, 15-Sep-99; (c) 1995-1998 Netscape Communications Corp.
p1 $ ldd ./netscape
libBrokenLocale.so.1 => /lib/libBrokenLocale.so.1 (0x4000f000)
libXt.so.6 => /usr/X11R6/lib/libXt.so.6 (0x40011000)
libSM.so.6 => /usr/X11R6/lib/libSM.so.6 (0x4005b000)
libICE.so.6 => /usr/X11R6/lib/libICE.so.6 (0x40064000)
libXmu.so.6 => /usr/X11R6/lib/libXmu.so.6 (0x40079000)
libXpm.so.4 => /usr/X11R6/lib/libXpm.so.4 (0x4008b000)
libXext.so.6 => /usr/X11R6/lib/libXext.so.6 (0x4009a000)
libX11.so.6 => /usr/X11R6/lib/libX11.so.6 (0x400a6000)
libdl.so.2 => /lib/libdl.so.2 (0x4014a000)
libstdc++.so.2.8 => /usr/lib/libstdc++.so.2.8 (0x4014d000)
libm.so.6 => /lib/libm.so.6 (0x40192000)
libc.so.6 => /lib/libc.so.6 (0x401ab000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x4000)

This is really good news at cleaning up some of the old libraries:
I could get rid of oldlibs/libc5, oldlibs/libg++27, and oldlibs/xlib6.

-- 
Keith

Re: Netscape 4.71 Is Rock Solid & Fast! [Details of how to get]

[Keith Harbaugh]

On Tue, 1999-10-12 at 17:50:14 -0500, John Foster wrote:
> I just installed the new Netscape 4.71 version on my Debian Linux
> server. Just a report- It installs nicely with the Netscape4 installer
> from Debian if you rename it to the proper convention. It seems MUCH
> faster and more stable that all of the previous versions I have used.

Just thought it might be worthwhile to document a little more precisely
some experiences with Netscape 4.7 [not 4.71, as John later corrected].

The following command, or its ftp equivalent, fetches the file in question:

wget 
ftp://ftp.netscape.com/pub/communicator/english/4.7/unix/supported/linux20_glibc2/complete_install/communicator-v47-export.x86-unknown-linuxglibc2.0.tar.gz

To install it, an approach that works fine for me is:

tar xzf it, then cd and ./ns-install;
add
   export MOZILLA_HOME=/usr/local/netscape/
   alias   netscape=$MOZILLA_HOME/netscape
to your environment, and you're "home" :-) free.

I tried using the Debian installation method, using .deb's of netscape
from netgod.net, back circa March, but for some reason that didn't work for me,
so I switched to this non-Debian method.

Running slink.

debian-security: another new mailing list

[Keith Harbaugh]

This is to announce the establishment of a new debian mailing list:

debian-security,

for the discussion of all aspects of security
significant to the Debian system, including cryptography.

Why have a list dedicated to security?

The primary reason is to facilitate future reference and research.
When, in the future, questions arise about security issues,
e.g., what was the rationale for establishing
the security policies that will be developed,
it will be far easier to research those issues
if the discussion is brought together in one place,
rather than being intermixed with all the usability and developmental
issues that necessarily dominate debian-user and debian-devel.
Why have to wade through the masses of, say, release-oriented posts on
debian-devel if one is only interested in a security-related matter?

Conversely, by removing some of the traffic from debian-user and debian-devel,
it should make those lists a little more compact.

As to the invited inclusion of both users and developers on this list,
it is hoped that that may make the list a better learning vehicle and
communication channel for all parties.

Your cooperation in subscribing to and using the list for its stated purpose
will enable those goals to be achieved.

To subscribe, please visit the Debian mailing list subscription page
<http://www.debian.org/MailingLists/subscribe> or send mail to
<[EMAIL PROTECTED]> with the single word
'subscribe' in the message body.

Thank you for your cooperation.

-- 
Keith Harbaugh

Re: Potato frozen?

[Keith Harbaugh]

On Thu, 1999-10-07 at 20:05:51 -0400, Carl Fink wrote:
> I know that as a matter of policy, Debian doesn't announce release
> dates in advance, but can someone give a *hint*?  Is the potato
> release even frozen yet?  I mean, I heard that the only problem
> remaining was a Perl tweak months ago, so surely it's getting close to
> done, right?

Apropos your thoughts:

There was a huge debate in August on how to migrate legacy documentation
from /usr/doc to the FHS-mandated /usr/share/doc;
the problem was that some key packages had already been migrated,
while others were evidently a long way away from being ready to migrate,
so the choices were
a) to make some key developers undo some hard work,
b) to wait a long time for all packages to be migrated, or
c) (the final choice) to figure out a way (using symbolic links)
   to allow both possibilities on your system while preserving
   some semblance of consistency for those who don't want to play
   guessing games on where documentation should be found.
(This is my personal, totally unauthorized, interpretation of what happened;
see Debian Weekly News, September 7 and earlier, for a more official view.)
Both the debate and the implementation have surely delayed things.

One good way to track release status, without wading through the >2000 messages
per month on debian-devel, seems to be to follow the mailinglist debian-release.
And then of course Debian Weekly News gives an excellent overview of things.

A relevant statistic is the number of Release-critical Bugs identified
in the Release-critical Bugreports (see debian-devel-announce).
Here are the last four numbers:
09-17   271
09-24   263
10-01   266
10-08   265
In this context, Adam Di Carlo's message on debian-release of 1999-10-08
at 21:43:49 -0400 makes for interesting and somber reading,
as well as giving a good and concise summary of status.
(This is said in an effort not to be critical, but just realistic.}

There may well be a tradeoff between the timely and the bug-free,
as the expected freeze date approaches.

All thoughts above have absolutely NO official status,
and are purely the unexpurgated opinions of the author.

Re: GNU Emacs-20.4 debs

[Keith Harbaugh]

On Fri, 1999-10-01 at 18:48:03 +0400, Alexander Zhuckov wrote:
> Hi!
> 
> Tell me, ple-e-e-e-ase, where I can find
> GNU Emacs 20.4 Debian packages?

This doesn't answer your question directly, but it may help keep you
up to date without having to wait for deb's:
If you have the a gcc compiler (I use gcc-2.7.2.3, but presumably egcs
would work as well) emacs-20.x is REALLY easy to build for yourself
from the source, which you may get from

Once you untar it (I use tar xzf ...),
just cd to emacs-20.4, then:
   ./configure
   make
   make install
which will install all the emacs goodies into /usr/local/
You can override /usr/local with --prefix=... at ./configure time.

Keith

printing debian web page using html2ps

[Keith Harbaugh]

The design spec for dpkgv2, aka the Herring Package Management Library (HPML),
is available for our browsing pleasure at
   http://www.debian.org/~bcollins/hpml;
specifying precisely that URL does bring up the proper web page
on my web browser.

If I am interpreting the html2ps documentation correctly,
giving the shell command
   html2ps www.debian.org/~bcollins/hpml,
perhaps with an argument of -W b, should produce
a postscript rendition of that html code.
But when I try that, I get the error message
   *** Error opening www.debian.org/~bcollins/hpml,
   Error: /nocurrentpoint in --currentpoint--
   [followed by gs stack trace data]
and various plausible variations on the URL just give
corresponding variations on the error message.

Does anyone know how to make html2ps work on that web page,
including its hyperlinked children?

TIA
Keith

Re: Root password

[Keith Harbaugh]

On Fri, 1999-09-24 at 08:32:28 -0500, Andrei Ivanov wrote:
> Here  is how you do it:
> reboot and at lilo prompt type 
> linux init=/bin/sh
> 
> This will rop you into shell. In there, 
> mount -n -o remount rw
> 
> Then edit the /etc/passwd file and blank out root password field. (The
> second field). Or do that to shadow file, if you have one.
> Then 
> sync
> and reboot. You should be able to get in without password so you can
> change it then
> Andrew

I tried the commands as written,
but the mount command produced the error message
   mount: can't find rw in /etc/fstab or /etc/mtab
Did you mean
   mount -n -o remount,rw /dev/x
where /dev/x should be replaced by the filesystem to be mounted?
With that, everything went just fine.

Keith

Re: Does X11 ignore ~/.bash_profile?

[Keith Harbaugh]

On Fri, 1999-09-24 at 00:42:35 +0200, Johann Spies wrote:
> I am trying to install real player and have the following lines in my
> ~/.bash_profile:
> 
> LD_LIBRARY_PATH=/usr/local/src/rvplayer5.0
> export LD_LIBRARY_PATH
> PATH=$PATH:/usr/local/src/rvplayer5.0
> 
> When I do an "echo $PATH" in the console, I get:
> 
> .::/home/jhspies/bin:/usr/gnu/bin:/usr/bin/X11:/use/X11R6/bin:
> /usr/local/bin:/usr/ucb:/bin:/usr/bin:/usr/etc:/etc:/sbin:
> /usr/sbin:/usr/games:/usr/local/sbin:/home/jhspies/bin:
> /usr/arx/bin:/usr/local/lib:/usr/local/pgsql/bin:
> /home/jhspies/Office51/bin:/usr/local/src/rvplayer5.0
> 
> However when I do it in xterm I get 
> 
> .::/home/jhspies/bin:/usr/gnu/bin:/usr/bin/X11:/use/X11R6/bin:
> /usr/local/bin:/usr/ucb:/bin:/usr/bin:/usr/etc:/etc:/sbin:
> /usr/sbin:/usr/games:/usr/local/sbin:/home/jhspies/bin:
> /usr/arx/bin:/usr/local/lib
> 
> Does X11 ignore ~/.bash_profile?
> 
> Johann.

The answer to your explicit question is:
"It depends on what you tell xterm you want."
as the following quote from the xterm man page describing
one of the resources available in the xterm.vt100 widget:
   loginShell (class LoginShell)
   Specifies  whether  or  not the shell to be run in
   the window should be started  as  a  login  shell.
   The default is ``false.''
hopefully makes clear.
Thus putting
xterm.vt100.loginShell: true
in your .Xresources file would make each xterm shell a login shell,
which would then read ~/.bash_profile during its initialization.
(xterm also has command line options -ls and +ls to determine this,
if you are starting xterm from a command line.
But the Xresources approach also catches all those xterms
that get started up implicitly somehow or another.)

But the common solution to your implicit question is as others have suggested:
put most all of your initialization into ~/.bashrc,
and then in ~/.bash_profile source ~/.bashrc.
The usage of ENV may be influenced by this quote from the bash FAQ:
bash reads ~/.bashrc for interactive shells, $ENV for non-interactive

Keith

Re: Strange bash prompt

[Keith Harbaugh]

On Wed, 1999-09-22 10:25:29 +0530, XRDLAB wrote:
> Hi,
> 
> I noticed a strange behaviour of bash regrding the prompt. I have set
> PS1='\h:\w$ '. With that I get both the host name and the working
> directory as my shell prompt. Yesterday I noticed a strange behaviour
> accidentally. The sequence is given below:
> 
>  mysxrd:~$ cd /var 
>  mysxrd:/var$ prompt is correct
>  mysxrd:/var$ cd ../usr
>  mysxrd://usr$    prompt has 2 slashes!
> 
> I am surprised by this behaviour. Is it expected? The bash version is:
>   GNU bash, version 2.01.1(1)-release (i486-pc-linux-gnu)
> 
> TIA,
> 
> sridhar

Here's your test run first with bash-2.03, then with bash-2.01:
euler:~$ echo $PS1
\h:\w$
euler:~$ cd /var
euler:/var$ cd ../usr
euler:/usr$ echo $BASH_VERSION
2.03.0(1)-release
euler:/usr$ ls -l /bin/bash*
-rwxr-xr-x   1 root  1582874 May 12 06:37 /bin/bash
-rwxr-xr-x   1 root   426980 Oct 20  1998 /bin/bash-2.01.1
-rwxr-xr-x   1 root  1582874 May 12 06:37 /bin/bash-2.03.0
euler:/usr$ /bin/bash-2.01.1
bash-2.01.1-2.01$  PS1='\h:\w$ '
euler:/usr$ cd
euler:~$ echo $PS1
\h:\w$
euler:~$ cd /var
euler:/var$ cd ../usr
euler://usr$ echo $BASH_VERSION
2.01.1(1)-release

The bash-2.01 was the slink release from debian;
the bash-2.03 was built from sources obtained from ftp.gnu.org,
using gcc-2.7.2.3 on a slink system.

Keith

Re: What is required for Kernel 2.2.12 ?

[Keith Harbaugh]

On Sat, 1999-09-18 14:57:50 -0400, Salman Ahmed wrote:

> Can someone tell me exactly what packages I will need to
> run the 2.2.x kernel ? I know that I will need some packages
> from unstable but that's ok.
> 
> Here is what my /etc/apt/sources.lists looks like right
> now :
> 
> # Use for a local mirror - remove the ftp1 http lines for the bits
> # your mirror contains.
> # deb file:/your/mirror/here/debian stable main contrib non-free
> # See sources.list(5) for more information, especial
> # Remember that you can only use http, ftp or file URIs
> deb http://http.us.debian.org/debian dists/proposed-updates/
> deb http://http.us.debian.org/debian stable main contrib non-free
> deb http://non-us.debian.org/debian-non-US stable non-US
> deb http://www.netgod.net/ x/
> 
> I suppose I will have to add an entry for the unstable area like :
> 
>  deb http://http.us.debian.org/debian unstable main contrib non-free
> 
> BTW, on a somewhat related note what problems should I expect
> when I start mixing up packages from unstable ? I am doing this
> on my home system which is a standalone system. I use it for
> doing some development and to browse the WWW and read this list.
> 

I did this back in July (for 2.2.10) on a slink box,
and took a rather minimal route.
My /etc/apt/sources.list only contained two lines:
   deb http://ftp.netgod.net/ x/
   deb http://http.us.debian.org/debian stable main contrib non-free
I made myself current with those sources using dselect, although
apt-get would presumably have worked equally well.
Note that neither unstable nor proposed-updates was necessary; netgod
had taken care of the required netbase upgrade, to netbase_3.12-2_i386.deb.

Compiling (using gcc-2.7.2.3) the ftp.kernel.org sources then ran fine,
using the instructions in Linus's README file.

After booting the 2.2.10 kernel, the following configuration problems
needed attention:
1) you get harmless boot-time error messages using slink's /etc/init.d/network;
   the following minimal one avoids those error messages and works fine for me:
ifconfig  lo 127.0.0.1
IPADDR=XXX.XXX.XXX.XXX
BROADCAST=YYY.YYY.YYY.YYY
GATEWAY=ZZZ.ZZZ.ZZZ.ZZZ
ifconfig  eth0 $IPADDR  broadcast $BROADCAST
route  add default  gw $GATEWAY  metric 1
   substituting appropriate values for the addresses, of course
2) in /etc/printcap, change lp1 to lp0

If memory serves correctly, that's all you need to do.

Since then I've upgraded to 2.2.12, and added
   deb http://security.debian.org/ stable updates
   deb http://non-us.debian.org/debian-non-US stable non-US
without any problem.

Hopefully this clarifies that you don't _need_ to upgrade to unstable
to enjoy the 2.2 kernel, but you do need to make some very minor configuration
file changes.
If I've missed anything, I hope somebody will notice and add/correct words
for a more accurate description.

Keith

Re: Fonts

[Keith Harbaugh]

On Tue, 1999-09-14 22:10:09 -0700, Craig B wrote:
> I really like the way X works on my Hamm system.  The only complaint I
> have is with the font quality.  Sometimes they look fine for example as
> I type this message, and other times they look really crappy.  The
> biggest problem I have is with Netscape Navigator.  Sometimes the fonts
> are too small and sometimes they are just distorted.  I have tried all
> the font settings I can find.  I have seen other posts complaining about
> this and I understand there are some ways to improve the font quality.  
> 

Maybe you've already tried this, but when I started with Netscape 4.51
on both hamm and slink I also found the fonts were much too small for an
1152x864 screen.
The solution was to follow the menu path:
  Edit -> Preference -> Fonts,
then adjust the sizes of the Variable Width (Times) and Fixed Width (Courier)
fonts to taste.  I wound up with 12.0 and 10.0; your mileage may vary.
The results seem fine in normal, italic and bold, in all sizes,
even with the standard Debian distributed fonts.

Can you cite specific Web pages that yield poor results?

Keith

Re: glimpseindex

[Keith Harbaugh]

On Mon, 1999-09-13 18:03:35 +0100, Martin Oldfield wrote:
> I thought I'd give glimpse a whirl to index a bunch of email archives, 
> but I get a segmentation fault:
> 
> 
> % glimpseindex -o -B archive
> 
> This is glimpseindex version 4.1, 1997.
> 
> Indexing "/home/mjo/adamsarch/archive" ...
> 
> 
> Size of files being indexed = 172104723 B, Total #of files = 26801
> Segmentation fault
> 
> 
> The BTS seems to have similar problems already e.g. bug 20037 but no
> obvious solution. I'm running version 4.1-2 which seems to be the most 
> recent.
> 
> Has anyone else been down this path before, and did they find a
> solution ?
> 
> Cheers,
> -- 
> Martin Oldfield,

I had success with just `glimpseindex -o'.
Caveats:
1) The working directory was ~/Mail.
2) Under that, there were a large number (~30) of subdirectories,
   each in maildir format (i.e., each had a new, cur and tmp directory,
   with the email being held one message to a file under cur).
3) The result was to place eight .glimpse... files in ~,
   but then glimpse worked just as documented to search them.
4) Also running version 4.1-2 on a slink system (but with linux 2.2.12).

Good luck,
Keith

Re: where is what? - clarification (hopefully)

[Keith Harbaugh]

Thanks to all who have replied.

Actually what motivated my question was the advocacy of the procmailer
Jari Aalto: please see his http://www.procmail.org/jari/pm-tips.html
(`pm' here and below is for `procmail'),
in particular, its
1.1 for its examples of the `@(#)' identifier (in addition to the normative
RCS `$...$'), and
1.5 for its explicit advocacy of the `what' program, *in-addition-to*
the RCS `ident' program

To provide a self-contained reference for what we are talking about,
here are excerpts from the text version `pm-tips.txt' of the html page,
first for 1.1:
.@(#) $Id: pm-tips.txt,v 1.74 1999/04/23 14:45:05 jaalto Exp $
.$Keywords: procmail sendmail formail mail UBE UCE spam filter $
.$URL: http://www.procmail.org/jari/ $
.$Contactid: <[EMAIL PROTECTED]> $
.$FileServer: send mail to Contactid with subject "send help" $
.$UrlLinksLastChecked: 1999-04-30 $
.@(#) This is a procmail tips page: a collection of procmail recipes,
.@(#) instructions, howtos. The document also contains URL pointers to
.@(#) the procmail mailing list and sites that fight against Internet
.@(#) UBE. You will also find many other interesting subjects that
.@(#) discuss about internet email: headers, mime and RFCs. There is

then for 1.5:
Please also familiarise yourself to unix what(1) and GNU RCS
ident(1), if you have those commands in your system. It is
important that you mark interesting text to these tools so that
someone can get an overview of your supplied files
% what  FILES   - Print @( # ) tags
% ident FILES   - Print $ $ keywords

Further, if one retrieves his file pm-code.zip, one finds nearly 50
source files commented using these conventions.

Given this extensive use of `unix what(1)', and wondering how much else code
is out there documented in a similar fashion,
I think it is a reasonable question to ask what is the `open source status'
of a tool to take advantage of those documentation conventions.

As to whatis, whatis is to "display manual page descriptions" according to
its man page, thus something from what Jari is requiring.
As to grep, I suppose grep can do something reasonable in this regard,
but in consideration of how long grep has been around, I suspect `what'
must have had some added value to make it worth having an independent
existence.

Clearly this isn't the most critical lacuna for debian, or for anyone else,
but I thought it might be worth bringing up. Thank you for your time.

Keith

Re: xterm-menus not working

[Keith Harbaugh]

On Sun, 1999-09-05 21:50:52 +0200, Johann Spies wrote:
> Thanks Keith.  That did it.  After changing my .Xdefaults to
> Xterm.vt.geometry in stead of using * the menus are usable.  In which
> documentation did you find that?

Well, actually it wasn't so much the documentation
as the experimentation that helped focus in on what was needed.
After installing X, I immediately customized the xterm
to take advantage of the full screen that was available,
but soon encountered the same problem that you did.
After some abortive attempts to solve the problem,
I backed off to an absolutely minimal X configuration:
just starting X from a console root login with xinit(1),
without an .xinitrc file in root's home directory
(and, of course, without starting xdm).
The resulting spartan xterm, while its geometry was only 80x24,
at least had the virtue that its menus worked as advertised
in xterm's manual page.
So that yielded the insight that there was nothing wrong with
the xterm software or X itself,
but rather something was wrong with the way I had customized things.
Then it was just a matter of trial and error to find the problem,
using mainly the X(1) and xterm(1) man pages for guidance.

I might add, though, that Branden Robinson has been doing a great job
of providing some much needed supplemental documentation in the FAQ's
he's been generating within the debian package.
Another great resource for X is still the X Window System User's Guide 
published by O'Reilly. Although it only covers up through Release 5 (1993),
at the user level not a whole lot has changed.
There's really a lot to X!

Keith

Re: xterm-menus not working

[Keith Harbaugh]

On Sun, 1999-09-05 14:44:47 +0200, Johann Spies wrote:
> On Sat, 4 Sep 1999, Keith Harbaugh wrote:
> 
> > If you want to do a little further digging on this situation,
> > try the following:
> >xrdb -q | grep -i ^xterm
> > and post the results back.
> > 
> 
> Here are the results:
> $xrdb -q | grep -i ^xterm
> XTerm*termName: xterm-debian
> XTerm*geometry: 80x24
> XTerm.vt100*geometry:   80x24
> XTerm*scrollBar:true
> Xterm*saveLines:1
> XTerm*font: 8x16
> 

I had exactly the same problem you originally described, shrunken xterm-menus,
when I originally used for the geometry resource what the X people call
"loose resource specifications", with the `*',
which modifies the geometry of *all* widgets subordinate to the XTerm class.
But the problem immediately went away when I switched to a
"tight resource specification", namely, 'XTerm.vt100.geometry: whatever',
which only influences the geometry of the XTerm vt100 widget,
but not the geometry of the menu widgets.

In your case, you still have the loose specification
> XTerm*geometry: 80x24
which would likewise be impacting the menu widgets.
You might want to examine several places from which that specification
could be being set:
   /etc/X11/Xsession
   /etc/X11/Xresources/
as well as your ~/.Xresources or ~/.Xdefaults.

Then again, maybe your problem has an entirely different cause,
so I'll bow out at this point.  But it is really nice to have those
menus working, e.g. to be able to adjust the font size and thus the
overall window size to fit the situation.

HTH,
Keith

where is what?

[Keith Harbaugh]

`what' was an old unix program which would access certain identifying
lines within text files, allowing easy reading of key parts of files
without having to use a pager or editor.

I used the search features of the debian web site's Package page, searching
on `what', but to no avail (the first search yielded the empty set,
the second search more than would fit on a page, but all that was not
the "right" what) (eh, what?).

So, two questions:
1) where (i.e., in what package) is the program `what' to be found?
2) how could I effectively use the web pages search, or some other search
   capability, to have answered that question without bothering the
   debian mailing lists?

Re: xterm-menus not working

[Keith Harbaugh]

If you want to do a little further digging on this situation,
try the following:
   xrdb -q | grep -i ^xterm
and post the results back.

And yes, I am using slink, but with some addons.
Here's my /etc/apt/sources.list if you're interested
   deb http://ftp.netgod.net/ x/
   deb http://security.debian.org/ stable updates
   deb http://http.us.debian.org/debian stable main contrib non-free
   deb-src http://http.us.debian.org/debian stable main contrib non-free
   deb http://non-us.debian.org/debian-non-US stable non-US

Keith

Re: xterm-menus not working

[Keith Harbaugh]

Change
   XTerm*geometry: 80x24
   XTerm.VT100*geometry: 80x24
to
   XTerm.vt100.geometry: 80x24
Let me know if that doesn't work;
it works fine for me (although actually I use 80x63).

Re: X with kernel 2.2.12

[Keith Harbaugh]

I just compiled 2.2.12: here's the uname -a:
Linux euler 2.2.12 #1 Fri Aug 27 05:43:17 UTC 1999 i586 unknown
No problem whatsoever, it just dropped in and replaced the previous
2.2.11 I had been running.
I'm running mainly a straight slink system, but with netgod.net's
X 3.3.3.1 + glibc-2.0.7 compilation.
Good luck to you.

Keith