Bug#20659: marked as done (dpkg: Suggest changing dselect's treatment of Recommends)
Your message dated Sun, 10 Oct 1999 02:14:10 -0400
with message-id <19991010021410.E18835@lappy.djj.state.va.us>
and subject line some dpkg bug maintainence with permission from wichert
has caused the attached Bug report to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere. Please contact me immediately.)
(administrator, Debian Bugs database)
Received: (at submit) by bugs.debian.org; 4 Apr 1998 18:36:42 +0000
Received: (qmail 11898 invoked from network); 4 Apr 1998 18:36:42 -0000
Received: from bigbang.generation.net (126.96.36.199)
by debian.novare.net with SMTP; 4 Apr 1998 18:36:42 -0000
Received: from localhost (gsstark@portD22.Generation.NET [188.8.131.52])
by bigbang.Generation.NET (8.8.7/8.8.7) with SMTP id NAA24553;
Sat, 4 Apr 1998 13:36:37 -0500 (EST)
Received: by localhost (sSMTP sendmail emulation); 4 Apr 98 13:36:35 EST
Date: 4 Apr 98 13:36:35 EST
From: "Gregory S. Stark" <email@example.com>
Subject: dpkg: Suggest changing dselect's treatment of Recommends
X-Mailer: bug 3.1.0
Currently dselect treats Recommends as equivalent to Depends, which correctly
makes it very hard for a user to install packages that violate Recommends via
dselect. However, if a user does manage to install such a package dselect
continues to insist something is wrong even though really the install is
correct, just unusual.
I would suggest that dselect continue to behave as it does for newly selected
packages, but ignore Recommends when checking dependencies on exit.
This would mean a user would be presented with the conflicts subwindow, and
would have to press `Q' to get out of it without satisfying the Recommends.
But the user would not then be forced to continually use `Q' to exit dselect
from then on. This would make dselect's dependency features more useful and
make people less likely to have to defeat them more than is necessary.
-- System Information
Debian Release: 2.0 (frozen)
Kernel Version: Linux portD22.Generation.NET 2.0.33 #2 Mon Jan 12 11:08:08 EST 1998 i586 unknown
Versions of the packages dpkg depends on:
libc6 Version: 2.0.7pre1-4
libg++272 Version: 184.108.40.206-0.1
ncurses3.4 Version: 1.9.9g-8
Received: (at 1037-done) by bugs.debian.org; 10 Oct 1999 06:12:48 +0000
Received: (qmail 18194 invoked from network); 10 Oct 1999 06:12:46 -0000
Received: from ppp28.ts1-3.newportnews.visi.net (HELO lappy.djj.state.va.us) (220.127.116.11)
by master.debian.org with SMTP; 10 Oct 1999 06:12:46 -0000
Received: from bmc by lappy.djj.state.va.us with local (Exim 3.03 #1 (Debian))
id 11aCFG-00038m-00; Sun, 10 Oct 1999 02:14:10 -0400
Date: Sun, 10 Oct 1999 02:14:10 -0400
From: Ben Collins <firstname.lastname@example.org>
To: email@example.com, firstname.lastname@example.org,
Subject: some dpkg bug maintainence with permission from wichert
Content-Type: text/plain; charset=us-ascii
1685: dpkg-split command line parsing.
My take on this, the options expected aren't in a compatible format,
so it can't be made GNU style. Also the man page shows that the
complained about format is expected. IMO, if it's documented as such,
then it should be acceptible. GNU is not the endall argument format.
16217: prompting for clearing avail on updates in dselect
dselect doesn't seem to do this anymore
10263: segfaults on upgrade
Very old bug concerning libc5->libc6 upgrades. I don't think we have any
way to really reproduce it, plus the reports show that the cause was
possibly found from some dependencies, but the original poster never
4074: conffile's left behind
Not a current problem with dpkg it seems
1037: dselect help screen
Ian said he would fix this in the report when he released the C version of
dpkg, that seems to have occured (bug report is vague on what the actual bug
is, and attempts at clarifying have failed).
Ben....more to come