| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Use make_path() instead of mkpath(), and remove_tree() instead of
rmtree().
|
|
Match the ld.so path resolution logic, by preferring multiarch paths
over multilib or cross paths.
[guillem@debian.org:
- Adapt to current code. ]
Closes: #772184
Signed-off-by: Guillem Jover <guillem@debian.org>
|
|
Split them so that we can mark the toolchain cross-compilation libraries
as deprecated, and in the future possibly just hide them by default.
|
|
Assign directly to @librarypaths when adding the cross-compilation
library paths.
|
|
This will allow in the future to hide these by default and avoid
polluting the default paths.
|
|
This was a temporary workaround for very old gcc toolchains.
See: #453267
|
|
This avoids inflicting the cost of the setup on any code loading the
module. It also makes it easier to unit test the module.
|
|
This will allow to initialize the library paths to be an empty array,
instead of just resetting it after the fact, which means the new name
will make more sense.
|
|
|
|
|
|
The loading is always the same, the difference is just in the parsing
logic.
|
|
Switch the conditional loading of the abitable to expect it to always
exist, and bump libdpkg-perl depends on dpkg to 1.16.3, the version
introducing the file.
This will allow some refactoring.
|
|
|
|
|
|
This will centralize common testsuite functions.
|
|
|
|
Fixes ValuesAndExpressions::RequireQuotedHeredocTerminator and
ValuesAndExpressions::ProhibitInterpolationOfLiterals.
|
|
This allows to easily replace the shebang at installation time to use
the system perl interpreter and will make it easier to analyze all
perl scripts in the source tree.
|
|
|
|
The old function name was inconsistent with the other gettext family
of short aliases which has already caused some code typos, and functions
starting with underscore are considered by convention private in Perl.
|
|
These allow to tag symbols as being only relevant on architectures with
specific bits or endianness. The current supported values are 32 and 64
for arch-bits, and little and big for arch-endian.
They can also be combined by using multiple tags on the same symbol.
Closes: #630342
|
|
Stop abusing the Dpkg::Deps object just to get the architecture
computation from it.
|
|
Factor out from Dpkg::Deps::Simple::arch_is_concerned().
|
|
These have a suboptimal license and implementation. Just get rid of
them.
|
|
|
|
These do not get affected by the user locale, nor by the character
encodings. For dpkg control data, we always want to use ASCII.
|
|
|
|
Some of these where accidentally dropped, reinstate them for now.
|
|
We are using these on arrays, so we end up having to compute the total
size from the number of members and the member size. Just delegate the
computation to the underlaying call which should ideally also handle
overflows and similar. And we can always pass nmemb=1 from the call
sites if needed.
|
|
|
|
Add how the pattern is interpreted as a substring match, or as a literal
path lookup. And explain how some metacharacters behave.
Closes: #775379
|
|
We should not be trimming those characters from substring matches, or we
might get very confusing results.
This was actually the intention behind commit
af9e264518c4cab8e70788d0724d362ef25534d6, but somehow failed at it.
|
|
|
|
|
|
|
|
Clarify how the package name gets generated when empty or omitted, and
when it sould be arch-qualified or not depending on whether the package
is Multi-Arch:same.
Closes: #776072
|
|
Add missing actions, and describe the startup messages.
Closes: #773398
|
|
|
|
|
|
Closes: #768842
Reported-by: Johannes Schauer <j.schauer@email.de>
|
|
|
|
|
|
|
|
|
|
There are two groups of packages, ones where the old version of the
packages produced trigger cycles as they were using awaiting trigger
directives, but are fixed now, which get a << versioned Breaks on the
fixed version. And packages that are still not fixed that get a <=
versioned Breaks on the broken version.
|
|
Limit the buffer accesses to the size of the buffer being accessed. This
affects reads done when parsing field and trigger names, or checking the
package ownership of conffiles and directories.
Use a new length member for struct fieldinfo and nickname to avoid
recomputing the same known length over and over again, but use strlen()
instead for arbitrary fields, conffiles and directories to avoid
increaseing the memory footprint too much.
Reported-by: Joshua Rogers <megamansec@gmail.com>
|
|
If we are not going to process triggers due to an unsatisfied
dependency, we should reset the trigger cycle tracking information
so that we do not end up reporting bogus cycle errors.
Closes: #771730
|
|
This allows building on older systems.
|
|
It seems that this is not supported or does not work in perl 5.14.0,
although it should be since 5.10.0. Switch to a group matching to
allow using older perl version from stable.
|
|
Signed-off-by: Guillem Jover <guillem@debian.org>
|
