This reverts commit 99d37f9b16, which
caused most binary uploads to be rejected. dak's allows upload of
debug symbol packages not listed in the Binary field only if there is
a corresponding binary package without the -dbgsym suffix, which is
not the case on architectures where we use a -unsigned suffix.
Any packages listed in debian/control that are not installed in the
main archive will always be seen as NEW. This might be fixable by
archive configuration changes, but for now we'll generate them in a
similar way to debhelper.
- incoming.debian.org now uses pool layout
- deb.debian.org is a better default than ftp.de.debian.org
- ftp.debian-ports.org redirects to ftp.ports.debian.org, so use the
latter directly
I changed the wrapper to call gpgv instead of gpg. It is much easier
and cleaner to use local configuration this way, and it won't produce
a warning that the key isn't trusted.
I also removed used of an environment variable, as we (currently) only
pass one keyring filename here.
Include headers for all architectures that we build a kernel for.
This allows co-installation of per-flavour header packages for
multiple Debian architectures, and fixes the problem of arm64 headers
depending on arm headers that we did not include.
By default dpkg-architecture lets the current environment override the
architecture specified by the -a option. We mustn't let that happen
here as we are considering all architectures. Use the -f option to
force use of our specified architecture.
The current cross-compiler packages don't set the Multi-Arch field, so
specify that the cross-compiler package must be native, rather than any
architecture.
flex doesn't support multi-arch, and this would require splitting it
(#611230, #761449). Force use of the native package for now.
openssl doesn't support multi-arch but probably easily could (#827028).
Force use of the native package for now.
We need the native libssl-dev while building the kernel itself and the
host libssl-dev while building tools for linux-kbuild.
Document the state of cross-building in README.source.
These packages will be taken over by src:linux-signed. Still do
everything but building the packages so we find configuration
errors before building linux-signed.
This fixes some of the problems dch was causing:
- Putting the stable log in the wrong place
- Updating the date unnecessarily
Change stable-update.sh to be a wrapper for stable-update.
Delete ckt-stable-update.sh; if we need it again in future, it can be
implemented more cleanly as part of the new script.
- Enable it by default
- Disable it for armel/marvell since signature verification is not enabled.
- Disable it for mips and mipsel so linux-signed can be uploaded without
waiting for them to build
- Disable it for all architectures not in the main archive, as linux-signed
won't support them (at least, not initially).
We don't need a variable to control signing of the image, because
we should do that for all flavours that have CONFIG_EFI_STUB=y.
* Drop redundant gitignore.patch from linux-tools
* Rename linux-tools' debian/templates/control.main.in to
debian/templates/control.tools.in
* Combine changelogs, putting all entries for each upstream release
cycle in chronological order
* Combine rules and gencontrol.py code
The linux-grsec source package needs a way to explicitly disable these
binary packages which are already built by the linux source package.
We already do that when there are no actual kernels for the target
architecture. Rename the FOREIGN_KERNEL make variable and combine the
two conditions.
Based on work by Yves-Alexis Perez.
A parallel 'debian/rules build' will now invoke 'debian/rules.real
build' twice in parallel, which is disastrous.
- Add and use proper build-arch and build-indep targets in
debian/rules.gen and debian/rules.real
- Assign a separate temporary directory to each target in
debian/rules.real. Add the directories to .gitignore and
the clean rule.
- Pull installation of the lockdep wrapper (which is indep)
up into debian/rules.real so that we don't end up building
liblockdep twice in parallel.
Sort the tarball contents and set their mtimes to the upstream
release time. (In case we're given a tarball and patch, the
release time is unknown and the result is not reproducible. So
don't do that.)
Ben Hutchings
Yves-Alexis Perez