Parent Directory
|
Revision Log
| Sticky Revision: |
releasing version 1.5.3
* Slightly optimised the postinst script while leaving old transition handling code in it by moving old code into blocks with a single check for really old versions of debconf.
move from cvs import
* Move from build-depends-indep to build-depends, to match current policy.
* Updated Japanese template po file from Kenshi Muto. Closes: #210374
* Removed the outdated debconf tutorial which was aimed at converting
pre-debconf packages to debconf. Use debconf-devel(7) for all your debconf
development needs.
* Got rid of all the xml stuff, trimmed build deps down.
* Lots of other doc reference updates.
* Removed the pre 0.9 downgrade warning from prerm.
* Removed the debconf.cfg removal code from postinst.
* Removed the showold question, and all showold support except what's
necessary for dpkg-reconfigure. Closes: #129666, #184142
* Fixed stdin/stdout inversion in call to open3 in dialog frontend.
Closes: #155682
* Fixed transition_db.pl to pass the new extra type argiments to
Question->new and Template->new. Thanks to the help of Pre and Cliph on
irc. Closes: #147932
* Don't run all the db mangling code in the whole postinst on fresh
installs.
unreleased new version
* The "bite the bullet" release.
* Enhanced fix_db.pl to detect and deal with every debconf db corruption
scenario that has been reported to me. Run it on upgrade from versions
prior to this one. I suspect that all instances of inconsistent and
corrupt debconf db's are due to past bugs in debconf and especially the
transition from the crufty old data::dumper db, and the "fix" for the
missing template problem, plus possibly some unclean shutdown problems.
So fix them all now, and either the problems go away for good or I prove
my theories wrong if they pop back up later.
* Closes: #128707, #128265, #99786
unreleased
* Looks like new shared templates have been ignored ever since 0.9.10!
Symtoms that installed a second package that shared a template
with an already installed first did not add the second owner to the
list of owners. It's possible that this was also responsible for
sporadic reports of db corruption; the question that should belong
to a template going away when the only recorded other owner was purged.
* Added a call in the postinst to a program to clean up after this
problem.
* Added debconf-show, which displays all questinos belonging to a
package, their values, and indicates if they have been seen or not, all
in a compact format handy for bug reports. Now why didn't I think of
this before?
* To keep the bit scales balanced, removed an unused template.
* Remove /var/lib/debconf on upgrade. Nothing uses it anymore,
and it contains some large old database files, and on most systems,
a large amount of cruft (temporary files, editor backup files, backups
of the db, sockets, the list goes on and on).
* Expunged all remaining traces of the directory from debconf and its
documentation.
* debconf-mergetemplates: ignore locale settings, as they should not take
effect for this program. Closes: #91860
* Completed and updated Finnish translation by Jaakko Kangasharju
<ashar@iki.fi>
* Added code in the postinst to delete any files that inexplicably linger
in /usr/lib/perl5/Debconf/, since since files break the current debconf
badly. I don't know why two people have reported files there; it's
either user error or some crazy dpkg bug. Closes: #89471
* Check to make sure a template has owners besides just existing before
returning it in Template->new. This should not be necessary, but
it fixes a problem with the templates db sometimes having templates
that lack any owners. I hypothesize that the problem was caused by
one of the early 0.9x releases, but I cannot reproduce it, so this will
have to do for a workaround. Closes: #89155
* Added Debug driver and simplified some other debug output.
* Added Pipe db driver. If Craig really wants, he can use this to do
exactly what he was complaining that debconf could not do. I think,
though, that there are really better ways of accomplishing the same thing.
* Added versioning to some 'use base'es, to detect people who have an old
base.pm floating around somewhere. See bug #89050
* Workaround a bug
that showed up in fresh installs
or ancient upgrades. Closes: #88682, #88676
The fix is simple:
just don't abort if it fails
to unregister.
builds debs now, I'm letting people test it
Use new apt's apt.conf.d directory to enable preconfiguration. Other cleanups to the maint scripts.
implemented much saner tmp file handling /var/lib/debconf will be mode 755 and will have 2 subdirs
* Clean up after bogus foo/bar template I accidentially released. * Added German template translation by Michael Bramer <grisu@debian.org>, Closes: #82914
* Clean up after bogus foo/bar template I accidentially released. * Added German template translation by Michael Bramer <grisu@debian.org>, Closes: #82914
* The stuff the postinst adds to apt.conf now doesn't return a error
code and make the apt run fail even if peices of the system like perl
are broken, as they are now for so many people. I had held off on this
change for a long time, but enough is enough.
* Also some not-yet-ready copletion stuff in the text frontend.
* Use fully-qualified path for dpkg-preconfigure in apt.conf,
Closes: #58469
* Added code to postinst to delete long-obsolete /etc/debconf.cfg
* Renamed dpkg-preconfig to dpkg-preconfigure, for consistency.
Closes: #53893
* Moved dpkg-preconfigure and dpkg-reconfigure to /usr/sbin.
* Debconf install now asks if you want to preconfigure packages, and
if you answer no, does not add/removes call to dpkg-preconfig in apt.conf.
* Changed user's guide to match.
* dpkg-reconfigure runs the postrm now. It's also substantially smaller.
* Added comment to apt.conf that the line was added by debconf,
Closes: #51720
* Squashed a ConfModule startup warning.
* Removed an implicit apt dependancy.
* _Really_ fixed problem with newline after owner. Tested and resteted
this time. Closes: #48450
* Now uses version 2.0 of the configuration management protocol.
- All commands in the protocol now return a numerical return code,
optionally followed by a space and a text return code.
* confmodule is a new shell library that handles this by making each
command it provides now return the numerical return code. They continue
to set $RET to the text return code. This means that you now have to
check the return codes of those commands, or the set -e script you are
running them in may exit if they return an error code.
* confmodule.sh is now deprecated, but remains for backwards compatability,
and has special compatability code in it.
* ConfModule.pm handles this by making each of its commands, when called
in list contect, return a list consiting of the numeric return code, and
the string return code. When called in scalar context, it behaves
in a backwards compatable way.
* Deprecated the VISIBLE command. Check to see if INPUT returns 30 instead.
* Deprecated the EXIST command. Check for return code 10 from commands that
try to use the question, instead.
* The GO command no longer returns "back"; instead, it returns 30.
* Documented all this.
* Hey, a state machine is the way to go if you want to support back buttons!
Converted the tutorial to reccommend this, and converted debconf's own
config script into a state machine.
* Used tables in several places in the tutorial where they make sense.
* Split the actual working templates and code out of the tutorial, and put
it in the samples direcotry. It is included inline so it is still
available in the tutorial, but now I can also debug it and make sure it
works..
* Added the noninteractive frontend to the list of choices you get when
configuring debconf.
* If the text frontend fails (this can really happen, if you run debconf
w/o a controlling tty in an autobuilder, say), falls back to the
noninteractive frontend. (Closes: #48644)
* The web frontend now only accepts connections from localhost.
* The web and noninteractive frontends now print out text saying they are
running.
* If a frontend fails, the failure message is always printed, not just in
debug mode.
* Fixed checkboxes in the web frontend so if they are unchecked, this
fact is noted.
* Added debconf-loadtemplate to the .deb.
* Guarded postinst code that modifies apt.conf to prevent dup entries.
* Started doing some cleanup of the gtk frontend:
- It no longer flashes the window up on the screen unless it really
has a question to ask this time around.
- Made cancel button work.
It still segfaults on exit though.
* Added password data type. Currently supported by the Text frontend
(though it has problems displaying right in an xterm), the Gtk frontend,
and the Web frontend (though you'd be insane to use it!).
* Fixed a nasty bug in the fallback code.
* Read-protected the debconf db directory.
* Added a config script and templates for debconf itself. It uses
them to configure what frontend to use, etc. /etc/debconf.cfg is no
more.
* Modified Config.pm so it contains functions that return values,
not just hard coded values. The functions now try to pull values out of
the database, and fall back on the defaults. Also, environment
DEBIAN_FRONTEND always works for specifying a frontend now, overriding
all else.
* Changed a myriad of other files that use Config.pm to call the new
functions.
* The Priority module is no longer used to set priority, Config.pm can
handle that now.
* Added showold to Config.pm, you can always see old questions now, if
you like.
* Renamed the entire Line frontend to Text. Line really doesn't make as
much sense. If you're following long in CVS, I also probably broke your
repository again; a clean checkout is reccommended. Sorry.
* Got rid of the DEBIAN_FRONTEND environment variable entirely. Instead,
/etc/debconf.cfg has a variable in it to specify the default frontend
to use.
* Also added a question priority variable to the config file.
Added postinst.
This form allows you to request diffs between any two revisions of this file. For each of the two "sides" of the diff, enter a numeric revision.
| ViewVC Help | |
| Powered by ViewVC 1.1.5 |