Status of xpm.

[osiris]

  Ian, are you working on an elf version?  If not, I would be willing
to work on it if it's not too complicated.  I've never handled a
package before, so I'd like to start out easy.

I only brought this up because I've been doing some Motif work, and
without an elf xpm, I have to link in the *huge* motif static lib
every time (yuk).

Thanks
--
Rob

New lib{db,gdbm,readline}

[J.H.M.Dassen]

Here are new versions of libdb, libgdbm and libreadline.
Please note that this version of libgdbm will break the current perl
and man pages. Please wait for new version of these packages before
upgrading.

Feedback is appreciated.

Regards,
Ray

Date: 14 Dec 95 18:33 UT
Source: libdb
Binary: libdb1 libdb1-dev 
Version: 1.85.2-3
Description: 
 libdb1: the Berkeley database routines (runtime version).
 libdb1-dev: the Berkeley database routines (development files).
Priority: Low
Changes: 
 * Several small fixes: dependencies, ldconfig, libdb.so link.
   Suggested by David Engel <[EMAIL PROTECTED]>.
Files:
 -rw-r--r--   1 root root   273693 Dec 14 19:31 libdb-1.85.2-3.tar.gz
 -rw-r--r--   1 root root 3979 Dec 14 19:33 libdb-1.85.2-3.diff.gz
 -rw-r--r--   1 root root33675 Dec 14 19:31 libdb1-1.85.2-3.deb
 -rw-r--r--   1 root root   164871 Dec 14 19:31 libdb1-dev-1.85.2-3.deb
 fd8f1ac0203b6eb4d6b7013080a1ed53  libdb-1.85.2-3.tar.gz
 d8c37605d0d3243c904d1b481785dab0  libdb-1.85.2-3.diff.gz
 8a23d783cd7ee6dfc36f4919c3e0eee1  libdb1-1.85.2-3.deb
 6235ff4dcc4102590b5a87a3d93f0f6b  libdb1-dev-1.85.2-3.deb

Date: 14 Dec 95 18:35 UT
Source: libgdbm
Binary: libgdbm1 libgdbm1-dev 
Version: 1.7.3-6
Description: 
 libgdbm1: GNU dbm database routines (runtime version).
 libgdbm1-dev: GNU dbm database routines (development files)
Priority: Low
Changes: 
 * Several small fixes: dependencies, ldconfig, libgdbm.so link.
   Install info documentation via install-info.
   Removed backward compatibility link. Requires perl and man to be
   recompiled.
   Suggested by David Engel <[EMAIL PROTECTED]>.
Files:
 -rw-r--r--   1 root root82979 Dec 14 19:35 libgdbm-1.7.3-6.tar.gz
 -rw-r--r--   1 root root12234 Dec 14 19:35 libgdbm-1.7.3-6.diff.gz
 -rw-r--r--   1 root root13388 Dec 14 19:35 libgdbm1-1.7.3-6.deb
 -rw-r--r--   1 root root28368 Dec 14 19:35 libgdbm1-dev-1.7.3-6.deb
 6e031ad261914656596a426d9640dc52  libgdbm-1.7.3-6.tar.gz
 541a36ae06350342baac917843b67fd2  libgdbm-1.7.3-6.diff.gz
 0a05cec70b3396409205573f68959d67  libgdbm1-1.7.3-6.deb
 5cd6c00571508f49109c9ca5b8053064  libgdbm1-dev-1.7.3-6.deb

Date: 14 Dec 95 19:47 UT
Source: libreadline
Binary: libreadline2 libreadline2-dev 
Version: 2.0-10
Description: 
 libreadline2: GNU readline and history libraries, runtime versions.
 libreadline2-dev: GNU readline and history libraries. Development versions.
Priority: Low
Changes: 
 * libreadline.so link now in /usr/lib (where libreadline.a is).
 * Fixed extended description. Noted by Erick Branderhorst
   <[EMAIL PROTECTED]>.
 * Several small fixes: dependencies, symlinks, ldconfig.
   Suggested by David Engel <[EMAIL PROTECTED]>.
Files:
 -rw-r--r--   1 root root   408204 Dec 14 20:46 
libreadline-2.0-10.tar.gz
 -rw-r--r--   1 root root 3771 Dec 14 20:47 
libreadline-2.0-10.diff.gz
 -rw-r--r--   1 root root64284 Dec 14 20:45 libreadline2-2.0-10.deb
 -rw-r--r--   1 root root   113048 Dec 14 20:46 
libreadline2-dev-2.0-10.deb
 933f2e7ff54c45fa90197cbdb364b438  libreadline-2.0-10.tar.gz
 2594e726d0c74097f1d0f30a807e2578  libreadline-2.0-10.diff.gz
 f278a578539d993a6b0b6684612b3148  libreadline2-2.0-10.deb
 7f373812375f01dcf59272c427e91717  libreadline2-dev-2.0-10.deb
-- 
Cyberspace, a final frontier. These are the voyages of my messages, 
on a lightspeed mission to explore strange new systems and to boldly go
where no data has gone before. 

PCMCIA support in boot kernel?

[Bjoern Stabell]

hi,

as far as i can tell, PCMCIA support is not present in the boot kernel
-- i've tried 0.93R6 and 1.0 (from InfoMagic) -- of the bootstrapping
disks.  this is a drag for people with Notebooks and PCMCIA ethernet
cards or SCSI+CDROMs, and it seems other distributions (Slackware and
RedHat that i know of) have support for PCMCIA in their boot kernels.

if i'm right, would you please consider adding PCMCIA support;  if
i'm wrong, please let me know :)


how do i get access to the development directory at ftp.debian.org?


bye,
-- 
Bjørn Stabell 
  
  

Bug#2030: fileutils should depend on libc5

[Michael Alan Dorman]

Package: fileutils-3.12
Revision: 3

The package seems to contain ELF executables, yet does not depend upon libc5.

Mike.
--
"I'm a dinosaur.  Somebody's digging my bones."

Unanswered problem reports by date

[iwj10]

The following problem reports have not yet been marked as `taken up' by a
message to [EMAIL PROTECTED] or or `forwarded' by a
message to [EMAIL PROTECTED]

OVER 10 MONTHS OLD - ATTENTION IS REQUIRED:
 Ref  PackageKeywords/Subject   Package maintainer
  416 wenglish   perl doesn't flush output auto [EMAIL PROTECTED] (Robinson,

OVER 9 MONTHS OLD - ATTENTION IS REQUIRED:
 Ref  PackageKeywords/Subject   Package maintainer
  563 tartar -x fails to overwrite or c Ian Murdock <[EMAIL PROTECTED]
  579 image (?)  missing /usr/man/man8 manpages Bruce Perens <[EMAIL PROTECTED]>

OVER 8 MONTHS OLD - ATTENTION IS REQUIRED:
 Ref  PackageKeywords/Subject   Package maintainer
  660 gdbGDB gets address of structure  Ian Murdock <[EMAIL PROTECTED]
  662 procps top doesn't behave sensibly if Ian Murdock <[EMAIL PROTECTED]
  691 textutils  textutils package, fmt(1) prog Ian Murdock <[EMAIL PROTECTED]
  702 findutils  locate crash with large db Ian Murdock <[EMAIL PROTECTED]
  710 xs3X server problem with hardware (unknown -- `xs')
  713 mh mh should pause after printing Jim Robinson <[EMAIL PROTECTED]
  725 xbase  twm places windows incorrectly Stephen Early <[EMAIL PROTECTED]
  729 procps Bizarre corrupted output from  Ian Murdock <[EMAIL PROTECTED]
  731 ncursesncurses wgetnstr doesn't work  (unknown -- `ncurses')
  740 xbase  xclock leaves `droppings' in i Stephen Early <[EMAIL PROTECTED]
  746 cpio   mt doesn't support setblk (and Ian Murdock <[EMAIL PROTECTED]
  759 kbd, xbase /usr/bin/X11/showfont conflict Ian Murdock <[EMAIL PROTECTED]
  773 xbase  xmh falls over if mh is not in Stephen Early <[EMAIL PROTECTED]
  775 xbase  twm reports errors on incorrec Stephen Early <[EMAIL PROTECTED]
  783 tartar --same-order doesn't work  Ian Murdock <[EMAIL PROTECTED]
  785 cpio   mt problemsIan Murdock <[EMAIL PROTECTED]
  786 syslogdsyslogd gone awol  Martin Schulze <[EMAIL 
PROTECTED]
  797 (base) /etc/termcap console keydefs f (unknown)
  798 svgalibsvgalib gets control key mucke Ted Hajek <[EMAIL PROTECTED]
  808 emacs  Info anchors not active in ema Ian Murdock <[EMAIL PROTECTED]

OVER 7 MONTHS OLD - ATTENTION IS REQUIRED:
 Ref  PackageKeywords/Subject   Package maintainer
  817 tartar -T /dev/null extracts whol Ian Murdock <[EMAIL PROTECTED]
  818 bash   bash builtin `echo' doesn't ch Ian Murdock <[EMAIL PROTECTED]
  819 tartar should have null-separated Ian Murdock <[EMAIL PROTECTED]
  820 tcsh   tcsh builtin `echo' doesn't ch Andrew Howell <[EMAIL PROTECTED]
  821 shellutils /bin/echo doesn't check write  Ian Murdock <[EMAIL PROTECTED]
  822 tartar -t doesn't check write err Ian Murdock <[EMAIL PROTECTED]
  824 cpio   cpio should have non-verbose,  Ian Murdock <[EMAIL PROTECTED]
  825 trntrn warning messages corrupt t Ian Jackson <[EMAIL PROTECTED]
  827 libc or sh who reports wrong hostname (wa Ian Murdock <[EMAIL PROTECTED]
  835 sysklogd   syslogd dies, leaves system un (unknown -- `sysklogd')
  836 (base) Possible bugs in termcap syste (unknown)
  841 ncursesdselect from dpkg 0.93.34 says (unknown -- `ncurses')
  844 manpages   readdir(3) should document str Ian Murdock <[EMAIL PROTECTED]
  845 manpages   access(2) is ambiguous Ian Murdock <[EMAIL PROTECTED]
  850 indent [indent] option mentioned in d [EMAIL PROTECTED] (Bill
  853 shellutils `nice' does not do anythingIan Murdock <[EMAIL PROTECTED]
  857 gs_bothgs (2.6.1pl4-4) doesn't use /e (unknown -- `gs_both')
  864 make   make gets MAKEFLAGS wrong  Ian Murdock <[EMAIL PROTECTED]
  887 xarchieR6  xarchie barfs when ftp closes  (unknown -- `xarchier')
  889 info   Info 3.1-6 Ian Jackson <[EMAIL PROTECTED]

OVER 6 MONTHS OLD - ATTENTION IS REQUIRED:
 Ref  PackageKeywords/Subject   Package maintainer
  902 lprlpr can't print a PostScript f Peter Tobias <[EMAIL PROTECTED]
  903 (base) /dev miscellaney   (unknown)
  911 libc   libc causes rsh to fail on com Ian Murdock <[EMAIL PROTECTED]
  918 miscutils  mkboot and image packages  Bruce Perens <[EMAIL PROTECTED]>
  923 xbase  xdm failed with `unknown sessi Stephen Early <[EMAIL PROTECTED]
  927 ncurses?   dselect display bug(unknown -- `ncurses')
  932 pine   Pine over-encodes files and au Ted Hajek <[EMAIL PROTECTED]
  933 pine   Pine wants to post my email re Ted Hajek <[EMAIL PROTECTED]
  934 pine   Pine `Full Header Mode' when r Ted Hajek <[EMAIL PROTECTED]
  944 sysvinit   clock (-u) questionBruce Perens <[EMAIL PROTECTED]>
  945 (base) clock (-u) question(unknown)
  957 dpkg   dpkg should automatically log  Ian Jackson <[EMAIL PROTECTED]
  981 ircii  irc package should ask for def Carl Streeter <[EMAIL PROTECTED

xpilot-3.4.2-1

[Stephen Early]

-BEGIN PGP SIGNED MESSAGE-

xpilot (3.4.2-1); priority=LOW

Package: xpilot
Version: 3.4.2
Package_Revision: 1
Maintainer: Stephen Early <[EMAIL PROTECTED]>

Description: Multi-player tactical game
 XPilot is a multi-player tactical manoeuvering game for X and Unix
 workstations.  Players have a fighter which they move along in an
 artificial world and shoot each other using various kinds of weapons
 like bullets, mines, smart missiles, heat seekers and so on.  It is a
 fast paced game with a lot of tactics.  There are also robots flying
 around shooting players and other robots.  Players can pick up special
 bonuses to improve the possibilities of their ship like more engine
 power or special weapons.  The aim of the game is to score points and
 to have a lot of fun.

Changes:
  * compiled as ELF

0e831bfcc788ab8cbc554b7bb54a8f09  xpilot-3.4.2-1.deb
94510ba0a7e091012ac6c4f78e1c7843  xpilot-3.4.2-1.diff.gz
97daa8e6a92bfab6f20dd383459afe68  xpilot-3.4.2-1.tar.gz
- -rw-r--r--   1 root root   329580 Dec 15 17:38 xpilot-3.4.2-1.deb
- -rw-rw-r--   1 sde1000  sde1000 24508 Dec 15 17:39 xpilot-3.4.2-1.diff.gz
- -rw-rw-r--   1 sde1000  sde1000739309 Dec 15 17:38 xpilot-3.4.2-1.tar.gz

-BEGIN PGP SIGNATURE-
Version: 2.6.2i

iQCVAwUBMNG1Soli3Xs+7S0lAQEJ3wP+MDzS175FVHFkq5Dx9kiVhTgN66bV+fjQ
rMRU86UUdNis4Qlg9FLOOoOzHnRBV/DQ8ATXNRl7tnC5G3zhvMfWdbYERmkjtPNv
oz2NnLCxmoOnqtDMWSd2aXzvZerjGSqzVSgTPpz5NfwJbSlpwVPQADaDmmHrNgJE
dPxQW8OShXg=
=SeOn
-END PGP SIGNATURE-

Packages/Contents under 0.93

[Karl Ferguson]

Hi...

I'm just wondering - does the Packages* and Contents* files really need
to be updated every day?  There are no new packages being put in the
0.93 area, and therefor I'm no so sure we have to...  Just saves the
mirrors downloading the same files over and over and over again etc.

...Karl

--
 
 | PO Box 828   Office: (09)316-3036 Fax: (09)381-3909
 |OWER INTERNET SERVICES   Canning Bridge   After Hours:  015-779-828
   WA, 6153 Sales Support: [EMAIL PROTECTED]
Internet Service Providers   
 and Networking Solutions   

Re: X package shared library problem

[David Engel]

> The problem is caused by ldconfig; when it is run in the xlib postrm, it
> deletes the .so link. It doesn't replace this link when it's run in the
> xlib postinst. 

This is why I suggested a couple of days ago that ldconfig should not
be run from any postrm scripts.  It should only be run from postinst
scripts.

> Can this be considered to be a bug in ldconfig?

Which part, deleting the link in the first place or not recreating it?

The latter is not a bug.  Ldconfig will never create the links needed
by ld as long as I'm maintaining it.

I'm open to suggestions on how to correct the former problem.  What
I'm leaning towards is to have ldconfig continue to remove dangling
libfoo.so.* links but not libfoo.so links.

David
-- 
David EngelOptical Data Systems, Inc.
[EMAIL PROTECTED]  1101 E. Arapaho Road
(214) 234-6400 Richardson, TX  75081

Re: X package shared library problem

[Stephen Early]

On Fri, 15 Dec 1995, David Engel wrote:

> > Can this be considered to be a bug in ldconfig?
> 
> Which part, deleting the link in the first place or not recreating it?
> 
> The latter is not a bug.  Ldconfig will never create the links needed
> by ld as long as I'm maintaining it.

Deleting something that you are not responsible for creating is probably 
not a good idea.

> I'm open to suggestions on how to correct the former problem.  What
> I'm leaning towards is to have ldconfig continue to remove dangling
> libfoo.so.* links but not libfoo.so links.

That sounds sensible.

Steve Early
[EMAIL PROTECTED]

Re: X package shared library problem

[David Engel]

> > Which part, deleting the link in the first place or not recreating it?
> > 
> > The latter is not a bug.  Ldconfig will never create the links needed
> > by ld as long as I'm maintaining it.
> 
> Deleting something that you are not responsible for creating is probably 
> not a good idea.

Agreed, but the problem is that it isn't always clear for which links
ldconfig is responsible and for which ones it isn't.

David
-- 
David EngelOptical Data Systems, Inc.
[EMAIL PROTECTED]  1101 E. Arapaho Road
(214) 234-6400 Richardson, TX  75081

X package shared library problem

[Stephen Early]

If the new xlib and xdevel packages are installed in the wrong order then 
the various *.so symlinks for the new libraries won't be made correctly. 
Check that the /usr/X11R6/lib/libX11.so symlink exists and points to 
libX11.so.6, and if it doesn't then re-install xdevel.

This is quite a nasty problem; no error messages are produced. Binaries 
just get linked statically, and unless you're looking out for this then 
you might not notice. I almost released a statically linked xpilot...

The problem is caused by ldconfig; when it is run in the xlib postrm, it
deletes the .so link. It doesn't replace this link when it's run in the
xlib postinst. 

Can this be considered to be a bug in ldconfig?

Steve Early
[EMAIL PROTECTED]

Bug#2031: ical doesn't work with new tcl/tk libs

[brian (b.c.) white]

The package "ical" no longer runs with the new tk4.  It seems to rely
on the older filename.

Brian
 ( [EMAIL PROTECTED] )

---
In theory, theory and practice are the same.  In practice, they're not.

Bug#2032: Printer stuck until the COMPUTER'S Reset button is pushed!

[Eddie Maddox]

Using: Debian 0.93R6 from ftp.debian.org.
On: 80386 DX 40
With: I/O card with a parallel port attached to...
And: HP Deskjet 500.

The printer works fine on DOS.

After booting Debian GNU/Linux the printer will not print,
the printer's own RESET button does nothing,
and the Power On Reset doesn't happen when cycling the printer power
switch.


The Linux command: ls > /dev/lp0
does nothing.

I have copied /usr/doc/HOWTO/Printing-HOWTO to /mnt... (my DOS disk),
pressed CTRL-ALT-DEL,
booted MS-DOS 6.20, and try to print, reset, cycle printer power,
all with no motion of any kind from the printer.

Only by pressing the RESET button on the COMPUTER,
and booting DOS then, of course, does the printer act normal and
do the RESET stuff and print things, such as the Printing-HOWTO.


{([|])}-{([|])}-{([|])}-{([|])}-{([|])}-{([|])}-{([|])}-{([|])}-{([|])}-{([|])}

The above reflects the observations of myself only, unless stated otherwise.

Eddie Maddox, Amateur Lobbyist  "Great spirits have always encountered
[EMAIL PROTECTED] violent opposition from mediocre minds."
P.O. Box 75321
St Paul MN 55175-0321   Albert Einstein
USA

Re: Packages/Contents under 0.93

[Bruce Perens]

I suppose it would be trivial for the script to "cmp" the old file and the
new and not replace the old file if they match.

Bruce
--
Bruce Perens <[EMAIL PROTECTED]> Pixar Animation Studios

Bug#2032: Printer stuck until the COMPUTER'S Reset button is pushed!

[Bruce Perens]

From: "Eddie Maddox" <[EMAIL PROTECTED]>
> The printer works fine on DOS.

Eddie,

You must have the "lp" module in your /etc/modules file.
Try "insmod lp" and see if the printer wakes up.
I know my printer doesn't move until that module gets loaded, so I'd guess
that's what is wrong with your system.

If that's the problem, please tell us so that we can close out the bug.

We don't have all of the drivers in the kernel when it starts because they
don't fit.

Thanks

Bruce
--
Bruce Perens <[EMAIL PROTECTED]> Pixar Animation Studios

Re: ncurses build options...

[Ian Jackson]

Michael Alan Dorman writes ("Re: ncurses build options..."):
> If you mean "during the time it takes them to do the other upgrade"  when
> you say "for a bit", then I think you're right.  I would like to confirm
> some dpkg behavior, though: 
> 
> Take ncurses-base-1 and ncurses-term-1.  The decision is made to move
> Wyse50 terminals from term to base.  New versions (xxx-2) are released. 
> User installs base-2, then term-2.  This would leave the user without
> Wyse50, unless dpkg moved the wyse50 from term.list to base.list. 
> 
> Does it do this (you said something earlier today that made it sound like
> it does)? 

Yes, it does do that.  If you install a package (A) containing a file
(foo) which is listed as being part of another already-installed
package (B) then the entry for foo in B's listing will be removed, and
A will end up the sole owner.

Directories are handled somewhat differently.  This is documented
somewhere, I think - try project/standards.

Ian.

Re: FTP Installation & Package Naming Conventions

[Ian Jackson]

Raul Miller writes in email to me:
> I presume you're familiar with the gnu conventions for labelling
> target/host architectures?  [e.g.  i486-unknown-linux]
> 
> If not, I'd like to draw your attention to them.

Yes, I am, thanks.

Do people think the architecture names as handled by dpkg should be in
this form ?  We'd probably end up with an Architecture control file
field which was a shell globbing pattern.

Ian.

Re: Comments...

[Ian Jackson]

Raul Miller writes ("Re: Comments..."):
> Raul Miller:
>> It would be even nicer if there were also a facility to register files
>> as part of a certain package, especially if it also had options to
>> register config files.
> 
> Ian Jackson:
>I'm somewhat confused.  Can you give a short example ?
> 
> Sure, say you want compressed man pages.  So, you register a hook for
> everything that compresses any files under /usr/man/man?/, and
> /usr/X11R6/man/man?/

Hmm.  Presumably you want this facility to work while dpkg is
running.  This is difficult, because dpkg has the whole database of
which file is in which package loaded into core.  I've added it to the
wishlist and I'll think about it, but don't hold your breath.

Ian.

Bug#2033: efax installation problems

[Robert Leslie]

Package: efax
Version: 07a-2

1. The postinst script says to modify /usr/bin/fax, but it should probably
   now say /etc/efax.rc instead.

2. /usr/bin/fax should probably no longer be listed in conffiles.

3. The software looks for spool files in /usr/spool/fax, but:
   a. It should probably look in /var/spool/fax instead, and
   b. This directory (and various subdirectories) are not created during
  installation (I have no idea what mode they should have?)

--
Robert Leslie
[EMAIL PROTECTED]

Re: PCMCIA support in boot kernel?

[Simon Shapiro]

Hi Bjoern,

I have been entertaining this thought.  Maybe as a followup to 1.3.47
(due out tonight or tomorrow).  Last I checked it barfed on 1.3.xx
kernels but maybe I am wrong.  I also have no way to test it as my
PCMCIA adaprtor is out of the system (no slots) and my only PCMCIA
devices are a modem (collides with my serial ports) and a 130MB
MiniScribe disk (no support I know of).  I wrote UnixWare support for
the MiniScribe but am too lazy to re-write it for Linux.  Any takers?

I'll integrate PCMCIA support (provides it compiles correctly) and
please tell me you will test it...



Simon

P.S.  Please ignore the below address and flame [EMAIL PROTECTED]
  He receives and answers mail :-)


Simon Shapiro   Bullet Technologies, Inc.
[EMAIL PROTECTED]  13130 SW Haystack St.
(503) 524-6631  Beaverton, OR 97005

Bug#1984: dpkg won't install cdtool

[Ian Jackson]

Raul Miller writes ("Re: Bug#1984: dpkg won't install cdtool"):
> begin 644 dpkg-out.gz

-chiark:~/junk> cat /var/lib/dpkg/info/workbone.postrm
#! /bin/sh
rm -f /usr/bin/workbone /usr/man/man1/workbone.1
-chiark:~/junk>

Honestly !  Why was this done ?  I suggest the package maintainer read
some of the documentation in project/standards, particularly
maintainer-script-args.txt.

This is with workbone 2.3-3.

I'm reassigning this bug to the workbone package.  I assume there is a
similar problem with cdtool.  If so that should be fixed too before
this bug is finally closed.

Ian.

Re: PCMCIA support in boot kernel?

[Simon Shapiro]

I'll try to add PCMCIA support to the follow-up release of 1.3.47.  47
is already stable and I'd rather not de-stabilize it.  It will be
released tonight or tomorrow.

Any volunteers to check it as I have no way to test it.

Also, I have a PCMCIA 130 MB disk.  I have written UnixWare drivers for
it.  Any Linux takers?



Simon

P.S.  Please ignore the below address and flame [EMAIL PROTECTED]
  He receives and answers mail :-)


Simon Shapiro   Bullet Technologies, Inc.
[EMAIL PROTECTED]  13130 SW Haystack St.
(503) 524-6631  Beaverton, OR 97005

Bug#2033: efax installation problems

[Dirk . Eddelbuettel]

  Robert Leslie writes:
  Robert>  Package: efax
  Robert>  Version: 07a-2
  Robert>
  Robert> 1. The postinst script says to modify /usr/bin/fax, but it should
  Robert> probably now say /etc/efax.rc instead.

Thanks, I will modify that for the next release.

  Robert> 2. /usr/bin/fax should probably no longer be listed in conffiles.

No! It is a conffile because /usr/bin/fax, a bash script, was the
only place to set your local fax settings as serial port, modem capabilities,
INIT strings, phone number, prinyer what have you.

Suppose you added your stuff to /usr/bin/fax. Now you install a new efax. If
it is not a conffile, dpkg will silently overwrite your settings in
/usr/bin/fax with a fresh and unmodified efax --- so that your old settings
are _lost_. Hardly desirable.

  Robert> 3. The software looks for spool files in /usr/spool/fax, but: a. It
  Robert> should probably look in /var/spool/fax instead, and b. This
  Robert> directory (and various subdirectories) are not created during
  Robert> installation (I have no idea what mode they should have?)

Part a) is not a bug because /usr/spool is a soft link to /var/spool.  Part
b) is a bug. This is due to the qfax patches for efax that I included in this
release. Qfax is an email-to-fax gateway to be used with efax. Qfax will be
available in January as it's prospective maintainer told me.

I will bugs 1. and 3b. Thanks for spotting them.

--
Dirk Eddelb|ttelhttp://qed.econ.queensu.ca/~edd

Bug#2033: efax installation problems

[Robert Leslie]

>   Robert> 2. /usr/bin/fax should probably no longer be listed in conffiles.
>
> No! It is a conffile because /usr/bin/fax, a bash script, was the
> only place to set your local fax settings as serial port, modem capabilities,
> INIT strings, phone number, prinyer what have you.

While /usr/bin/fax was the only place you could set these things previously, I
can understand this. But now that there is /etc/efax.rc, is this still
necessary? Perhaps a preinst script could detect whether configurations need
to be extracted from /usr/bin/fax (and put into /etc/efax.rc) before it is
clobbered?

If it is listed in conffiles, then it won't go away when I remove the package,
leaving (broken) executable software on my system. :(

> Part a) is not a bug because /usr/spool is a soft link to /var/spool.

True; the feeling I've been getting is that software should try to use the
"correct" path instead wherever possible, but I agree this is probably rather
insignificant.

Thanks for the quick followup.

--
Robert Leslie
[EMAIL PROTECTED]

Re: re:-O2 or -O3 ?

[Ian Jackson]

Bill Mitchell writes ("Re: re:-O2 or -O3 ? "):
> > >Does anyone disagree with Brian White ?  If not I'll change the
> > >guidelines back to recommending -O2.
> > 
> > Brian's right.  Stick with -O2.
> 
> But some verbage in the guidelines about -O3 and tradeoffs with -O2
> would probably be a good idea.

I've done this.

Ian.

Re: debian-1.0 availability

[Ian Jackson]

Bruce Perens writes ("Re: debian-1.0 availability"):
> > [wuarchive] systems/linux/tsx-11/distributions/debian/debian-1.0/sources
> > is empty !!!
> 
> I thought the GPL only required you to provide the sources
> _when_someone_asks_for_them_.
> 
> We did learn a lesson today.

My reading of the GPL2 is that it requires you to carry the sources on
an anon-ftp site which carries the binaries, unless you got just the
binaries with a written assurance that the sources would be available
or provide a written assurance yourself.

Ian.

Re: ncurses build options...

[Ian Jackson]

Bill Mitchell writes ("Re: ncurses build options..."):
> [ much deleted ]
> It seems to me that the the sequence would go something like this:
> ...
> 4.  User says something like "dpkg --remove --force-essential list_of_libs".

No, users should not need to use --force-remove-essential.  I was
worried when I put those flags in that people might think it
reasonable too make packages and structures that depend on them.  It
is not, and we should not encourage users to use any --force options.

Packages which are not essential in and of themselves, but are
essential merely because they (for example) provide shared libraries
for others should probably not be marked essential.

The dependency scheme will prevent their premature removal, if
properly used.

Ian.

Re: ncurses build options...

[Ian Jackson]

Michael Alan Dorman writes ("Re: ncurses build options..."):
> I forget who holds the keys to the virtual package stuff.  Would whoever 
> does that please register me an ncurses-dev package, in order to simplify 
> this for ncurses as it is for tcl/tk?

You can do this yourself.  Look in project/standards.

Ian.

Re: Default organization: is there a site config file?

[Ian Jackson]

Manoj Srivastava writes ("Default organization: is there a site config file?"):
>   It seems to me that this information is not news specific
>  (mailers could utilize it too), so should this file be installed
>  instead in /etc, also, which package should take responsibility? 

Mailers do (should) not usually add Organization headers.  I think it
should stay in /etc/news.

Ian.

Re: ncurses-1.9.8a ELF release

[Ian Jackson]

David Engel writes ("Re: ncurses-1.9.8a ELF release"):
> [ and earlier: ]
> > > The runtime package installs the shared libraries as lib*.so.3.0.new
> > > and then renames them to lib*.so.3.0 in the postinst script.  This is
> > > fine for not disturbing running programs when the package is being
> > > upgraded, but there is no provision for deleting the files when the
> > > package is removed.  Again, I'd like to hear Ian Jackson's thoughts on
> > > adding special installation support for shared libraries to dpkg.
> 
> Ian Jackson, are you there?  I'd *really* like to hear your opinions
> on this.

Yes, I'm here.  Sorry, I haven't been reading this mail for a few days
and I'm noow catching up on a 400-message backlog.

As far as I can see we have three kinds of shared library:

1. Packages like X or mkfs or what have you, where it doesn't matter
if the library link is missing momentarily, or even if it's absent
while the package is in a "broken" state according to dpkg.

2. Packages like the libc or ncurses, where we can't leave it broken
even for an instant without risking an unrecoverable system.

3. The shared library is part of the same package as the executables
and is version-specific - it's just there to save disk space and
memory, and furthermore this is a critical package which we can't
leave broken.

AFAIK only dpkg falls into category 3.  Lots of things fall into
category 1, but they can do without special handling.

Category 1 needs the link to be updated *with both libraries present*,
am I right ?

If you naively install the libraries under their real names in the
package, then there will only be a brief period while dpkg is running
when both libraries are present, and there's not currently any way to
get a script or anything run there.

You could possibly fudge something up with the maintainer scripts, and
putting the files in the package under different names.

Could you take a good look at maintainer-script-args.txt in
project/standards, and consider what extra functionality you'd like me
to include in dpkg ?

One possibility is an extra maintainer script called in the guts of
dpkg's processing - if you like that idea, remember to think about
error handling ...

Ian.

Bug#1995: run-parts on laptops

[Ian Jackson]

Raul Miller writes ("Bug#1995: run-parts on laptops"):
> run-parts should probably not do what it normally does, when a laptop
> doesn't have AC power.  This could be implemented with something along
> the lines of:

What does AC power have to do with run-parts ??

run-parts is just a utility to run all the scripts in a directory.

I think you should think where else this problem should be solved -
possible the answer is to modify your /etc/crontab.

Ian.

Re: newest ncurses

[Ian Jackson]

Michael Alan Dorman writes ("Re: newest ncurses"):
> On Sun, 10 Dec 1995, Michael Alan Dorman wrote:
> > I'm now uploading ncurses-1.9.8a-2 & co.  It is also available for ftp
> > from lot49.med.miami.edu:/pub/linux/ until it gets cleared at
> > ftp.debian.org.
> 
> I forgot to mention that you will almost certainly have to use 'dpkg 
> --purge --force-depends' on both ncurses-runtime and -developer, since 
> the new packages now (correctly) conflict with the above, and some base 
> packages depend on ncurses-runtime.

If you use dselect (or dpkg --remove without --force-depends) to
deselect the old packages, dpkg will automatically remove them when
the conflicting ones are installed (if they have the right Provides
lines).

> This may break those particular programs, since we have changed the 
> search structure for terminfo entries in the new packages.  You can 
> correct this by symlinking the entries under /etc/terminfo to 
> /usr/lib/terminfo.

Is it just me, or should that be /lib/terminfo ?  I know that termcap
used to be in /etc, but I think we're agreed that that (like most of
termcap) was an aberration.

terminfo doesn't contain configuration files, so I don't think it
belongs in /etc.

Ian.

Re: Package Verification

[Ian Jackson]

brian white writes ("Package Verification "):
> I'd like to suggest another field to be automatically added to the
> "Packages" files that exist at the top of each hierarchy in the
> distribution.  I'd like to see a "Checksum:" field that can be used to
> verify the correct download of these packages.  I think including both
> an 'md5sum' and a (filesize) would be best as the file size would
> allow a reasonable check on non-Debian systems and the 'md5sum' would
> allow absolute verification before installation.
> 
> Example:
> checksum: d14d384e0895986bc9f2b09f0a8b84fc (295393)
> 
> The reason for this is so programs like 'dftp' can verify that they
> retrieved the packages correctly before attempting to install them.

Eventually dpkg will have its own support for package verification.
Also, the format you propose can't be used as input to `md5sum -c'.

Ian.

Re: Packages Files

[Ian Jackson]

brian white writes ("Packages Files "):
> I notice that the "Packages" file in the debian-1.0 directory lists
> pathnames relative to ALPHA-TEST directory instead of the root
> directory of the distribution as is the case will all the
> other "Packages" files.  This makes little sense to someone who
> got to the debian-1.0 directory via the "development" symlink.
> 
> Could this be changed so they are all relative to, preferrably, the
> tree under which they reside (eg. debian-0.93, contrib, etc.) or,
> alternatively, the root directory of the distribution?

The 1.0 Packages file hasn't been updated since the tree was moved,
because it is generated by my crontab and I don't seem to be in the
appropriate group on ftp.debian.org.

If you hold off for a bit this will get sorted out when the ftp site
organisation is organised.

> Also, the "project" directory is missing the "Packages" file altogether.

The project directory shouldn't have one - it's not an ordinary
distribution tree, and isn't intended for automatic processing.

Ian.

Re: coming soon

[Ian Jackson]

Bruce Perens writes ("coming soon"):
> I am changing "init" in a few ways:
> 
>   1. The "initrunlevel" file is moving to /etc from /var/log.
>  Currently, init gets sick if /usr is unmounted. I may look
>  at an alternative to using this file, so that it can better
>  tolerate a read-only root.

/var/run, surely ?

>   3. /etc/rc[0-6].d will move to /etc/rc.d/rc[0-6].d to match the
>  practice on other Linux systems. Symbolic links will provide
>  compatibility with the old locations.

Is this really necessary ?  Real SysV's do things the way we have
done.

>   4. The /etc/init.d/functions file will no longer be used.

Please make it exist and be empty so that existing programs don't
break.  It should contain a comment saying that programs shouldn't use
it.

Ian.

Bug#1779: acknowledged by developer (was: unclutter - I need -noevents)

[Ian Jackson]

The bug tracking system writes:
> This is an automatic notification regarding your bug report.
>
> Responsibility for it has been taken by one of the developers, namely
> "D.J. Gregor" <[EMAIL PROTECTED]>.
>
> You should be hearing from them with a substantive response shortly, if
> you have not already done so.  If not, please contact them directly,
> or email [EMAIL PROTECTED] or myself.

I haven't received such a substantive response.

Ian.

Re: ncurses build options...

[Bill Mitchell]

On Sat, 16 Dec 1995, Ian Jackson wrote:

> Packages which are not essential in and of themselves, but are
> essential merely because they (for example) provide shared libraries
> for others should probably not be marked essential.
> 
> The dependency scheme will prevent their premature removal, if
> properly used.

Granted, but there are probably a lot of deficiencies in
dependency declarations.  I'm doubtful that we'll ever get
near 100% correctness on this.

Should all packages which use common unix commands provided
by essential digest packages such as cat (textutls), echo
(shellutls), mkfs (miscutls) declare explicit dependencies on
the digest packages they need, or is the fact that the digest
packages are essential sufficient to omit explicit declaration
of dependencies on them?

If essential packages depend on shared libs, should those
essential packages declare dependencies on nonessential shared
libs?  Should essential packages be dependent on nonessential
shared libs?