Couple of patches from Tollef ... Correct case of Origin and Bugs fields
in dpkg-scanpackages output and fix the unknown() function of controllib.pl
to output the correct field name.
Add support for the new Hurd-style i386-gnu triplets; rather than hacking
this onto the Linux support, I'm deliberately doing it separately so when
they change it again, we don't have to unbreak things.
Add support for pure amd64 in archtable and dpkg-architecture, it's
maturing sufficiently that we should probably support it now. archtable
made a bit easier to read too...
Run update.sh over the po files; this should probably get done before a
release gets tagged, otherwise if translators patch the .po in the tarball
then it makes it a bugger to merge into CVS.
* main/remove.c (removal_bulk_remove_configfiles): Don't change the
"previous pointer" pointer if we remove the node from the linked list,
ensuring that if the next node is to be removed the right thing will
happen. This corrects the bug where every second shared or diverted
conffile would be incorrectly deleted by dpkg.
Apply patch from Christian Haggstrom to fix #213038, #217286,
#213543 and #213846; there was an extra %s in the error message
printed when out of disk space, causing a strange SEGV.
Apply patch from Colin Watson to fix #139781; dselect had a
hard-coded maximum window width of 179 chars for some things,
when COLS is easily available.
Apply patch to fix #232025; the tar spec doesn't require a 100-char filename
to end with a NULL terminator, so we have to copy it out of the header
carefully.
Retro-actively apply patch for the 1.10.18.1 release:
* main/processarc.c (process_archive): Copy code from main/remove.c
to ensure that hardlinks to devices, setuid files or setgid files
cannot be stashed away in the hope that they become compromisable
in the future. This was handled when removing a package, but not
when upgrading one.
Adam Heath [Mon, 27 Oct 2003 01:43:06 +0000 (01:43 +0000)]
* scripts/dpkg-checkbuilddeps.pl: s/dpkg-checkbuild/dpkg-checkbuilddeps/
in usage.
* version-nr: Bump to 1.10.17.
* debian/changelog: Make the date UNRELEASED.
Adam Heath [Sat, 25 Oct 2003 20:03:20 +0000 (20:03 +0000)]
Almost *EVERY* damn file was including config.h in the wrong spot. After
making it the first include, then off_t/size_t/etc could actually be
redefined to be 64-bit compatible.
Adam Heath [Sun, 21 Sep 2003 01:43:39 +0000 (01:43 +0000)]
* man/*
* Imported several translated manpages, from debian-doc cvs:
* Spanish: 822-date.1 cleanup-info.8 deb-control.5 deb-old.5
deb.5 dpkg-checkbuilddeps.1 dpkg-deb.1 dpkg-divert.8 dpkg-name.1
dpkg-scanpackages.8 dpkg-split.8 dpkg-statoverride.8 dpkg.8
dselect.8 md5sum.1 start-stop-daemon.8 update-alternatives.8
* French: deb.5
* Portuguese: dpkg.8 dselect.8
* main/dpkg.8: -O had an inverted word.
* po/cz.po: Improve translation of: "dpkg - warning: ignoring request to
remove %s which isn't installed."
* po/it.po: Fix missing space in translation when listing files, and the
file is diverted.
* po/pl.po:
* Updated.
* Fix translation of 'conflicts with'
* po/ja.po: Fix typo in translation of dpkg --force-help; dowgrade ->
downgrade
* po/fr.po:
* Fix translation when listing files, and the file is diverted.
* Fix missing space in translation of: "%s - warning: downgrading %.250s
from %.250s to %.250s.\n"
* po/da.po: Fix translation of(was missing the leading space): " does not
appear to be available\n"
Adam Heath [Sat, 20 Sep 2003 23:48:18 +0000 (23:48 +0000)]
* Add a -d|--chdir option to start-stop-daemon.
* Split the background block into 2 parts: one that does the fork, and
opens /dev/tty and /dev/null, and one that does everything else. The
second block is then moved to be run right before the exec. This allows
error messages to be seen from the child(previously, they were lost), and
allows for the chroot to not require the device files.
* When --start, --startas and --pidfile are given, print 'process' instead
of '(null)' for the process name, if it's already running.
Adam Heath [Sat, 20 Sep 2003 00:57:39 +0000 (00:57 +0000)]
* lib/varbuf.c, configure.in: Fix va_copy detection.
* scripts/dpkg-buildpackage.sh: Back out debian/rules build-arch detection.
It is *not* possible *at all* to detect available targets in a rules
file. Period.
Adam Heath [Fri, 19 Sep 2003 17:29:08 +0000 (17:29 +0000)]
* scripts/{controllib,dpkg-{source,checkbuilddeps,control}}.pl:
* Rewrote the parsedep stuff, so that it wasn't done during control file
parsing. Scripts that need the internal parsed format should call
parsedep on the field's value.
* Split the substvars parsing into a separate function.
* No longer validate dependency fields when reading the control file.
Some fields may have vars in them, which breaks the validation.
* dpkg-gencontrol calls substvars after parsing the control file, and
then validates the substituted depends lines. Originally,
substitution occured only during writing of the final output file.
* Andreas Barth <aba@not.so.argh.org>:
Fix bad regex that didn't allow spaces in dependency field parsing.
It now directly matches what the c code expects. It previously was
allowing all of \S, which matched on '('. Closes: #211660.
* debian/control: Remove stale debug that would have broken installation on
s390.
* debian/rules: We were calling our own local copy of dpkg-gencontrol, but
the system-installed copy of dpkg-shlibdeps. Now, both are called from
the build directory.
Adam Heath [Wed, 17 Sep 2003 18:25:50 +0000 (18:25 +0000)]
* scripts/controllib.pl: Fix parsing of deps when both () and [] are
specified.
* scripts/controllib.pl: Set $host_arch in controllib.pl:parsedep.
* scripts/dpkg-checkbuilddeps.pl: Don't output the [] stuff when finding
invalid deps.