The Debian signing service won't pay any attention to this, but it
sort of makes sense - you won't be able to build a source package
without dpkg-dev.
Fixes lintian error missing-dep-for-interpreter.
Currently we don't set all the architecture variables, so direct use
of these makefiles (not under dpkg-architecture or dpkg-buildpackage)
may fail.
In case ARCH and DEB_HOST_ARCH don't match in rules.real, we really
should fail, so add an explicit check for that.
This also fixes lintian warning
debian-rules-sets-dpkg-architecture-variable.
uscan now does the file removals, which makes the hook script
redundant. Not only that, but the orig tarball passed to the hook is
no longer a symlink to the upstream filename, so genorig.py doesn't
recognise the filename and fails.
Since the linux-headers-common package was changed to be
architecture-independent, it won't include any binNMU changelog
fragment that should be installed for a per-flavour linux-headers
package.
Fixes lintian warning patch-file-present-but-not-mentioned-in-series.
Also preparation for using dgit, which will remove everything except
the main patch series under debian/patches.
Rename them to genpatch-{aufs,lockdown,rt}
Fixes lintian warning patch-file-present-but-not-mentioned-in-series.
Also preparation for using dgit, which will remove everything except
the main patch series under debian/patches.
Because these command lines don't start with "$(MAKE)", we should
tell make that we are recursing. Without this, the child make
can't run more than one job in parallel (and warns about this).
These packages don't contain binaries or libraries. The dependencies
on the corresponding shared library packages are specified directly,
not generated by dpkg-shlibdeps.
This fixes build-time warnings from dpkg-gencontrol.
- asciidoctor is only used to build the linux-perf package, so !nopython
- patchutils is only used to build the linux-source package, so
!pkg.linux.nosource
Several regexes were written as normal strings. Thankfully it appears
that none of the backslash-escaped sequences in them are parsed in
normal strings, so this shouldn't make any practical difference.
linux-perf wants to build a 32-bit program perf-read-vdso32 to enable
64-bit perf builds to read the vDSO for 32-bit tasks. This is now
worth doing for MIPS since it got a vDSO in Linux 4.4.
Ideally we should not require any special profile to be enabled for a
cross-build. (Unfortunately we still will because cross-compiler
and native compiler packages are named differently.)
For a native build, the build-dependency on libssl-dev is redundant
with the previous build-dependency on libssl-dev:native, so this
shouldn't make any practical difference.
Change all the build-dependencies for perf to be qualified by the
architectures we actually build perf for, except where they are
already arch-qualified.
Closes: #908519
The drivers in pata-modules depend on libata, which depends on scsi_mod.
The corresponding package dependencies pata-modules -> ata-modules
-> scsi-core-modules are already defined, and there is no sense in
adding a direct dependency from pata-modules to scsi-core-modules.
The usb-storage driver depends on scsi_mod in scsi-core-modules, not
any of the drivers in scsi-modules.