stripper: Hack to skip files with no DWARF information

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pakfire: Reduce pool debugging

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

file: Remove debugging code

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pakfire: Don't make the reader sticky

This totally goes wrong when we want to dist a source package while the
filelist is scanning for source files. Oh well. This isn't too bad.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

repo: Free scanned archives

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

packager: Refactor how we read metadata from the file system

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

file: Simplify creating new file objects

The create function will now take the path relative to the Pakfire
instance. It will also automatically compute the absolute path.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

path: Move pakfire_path_is_absolute to the path library

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

file: Add functions to store a file descriptor

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

filelist: Change walk callback to refer to the context

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

filelist: Remove the legacy logger and some mild refactoring

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

repo: Refactor scanning for archives

It does not make too much sense to (ab)use the filelist object because
it depends on so many other things and makes the code unnecessarily
complicated.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

FHS: Ensure the checked path is absolute

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

path: Add function to make a path absolute

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Seal the file after reading it into memory

This avoids that something (either intentionally or unintentionally) can
be changed when the file is being analyzed.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Read the payload into memory

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Run FHS checks on binary packages

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

FHS: Return a bitmap with any issues

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

progress: Don't return an error if there is nothing to update

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

FHS: Anchor on the context

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jenkins: Remove broken branch check

This does not seem to work, so we remove it for now.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

logging: Drop syslog-compatibility option

It is no longer possible to build Pakfire without systemd and so we
don't need to carry around code that cannot be used.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jenkins: Run Clang Static Analyzer on Debian Trixie

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jenkins: Add LCOV pass and build Debian packages

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

debian: Fix systemd unit directory path

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

configure: Don't try to autodetect the systemd unit directory

This no longer works on Debian Trixie and so on and I don't think there
is any reason why we should not hard-code this.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

debian: Add some files to create packages for Debian

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

util: Use our custom realloc function everywhere else, too

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

util: Free pointer if it could not be reallocated

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

snapshot: Ensure that pointer it always reset

Even if something else keeps holding a reference to the snapshot.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

xfer: Sanitise format strings

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

package: Initialize r

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: progress: Remove another unused function

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jenkins: Don't build for AlmaLinux 9 any more

AlmaLinux does not support the new mount API and we currently don't
really have a user for Pakfire on this platform.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

job: Sanitise format strings when creating xfers

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jail: Make sure argv is always set

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: progress: Remove unused function

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

configure: Check for the new mount API

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

daemon: Sanitise format strings when creating xfers

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

buildservice: Don't count packages

I don't know why we are doing this...

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jail: Return the correct return code on errors

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

job: Remove unused variable

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pakfire: Don't assign r when it not being read

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: Mark the assertion function to not return

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: version: Update to new-style function

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

repo: Sanitise format strings when creating xfers

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

key: Have pakfire_key_unref return a pointer

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

dist: Sanitize format strings when creating an xfer

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: progressbar: Remove the elements array

This was a dynamically allocated array which was difficult to track as
it was filled in two passes.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

packager: Initialize tmppath

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

arch: Initialize interpreter

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

buildservice: Sanitise format strings when creating xfers

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

solutions: Initialize elements

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Check if we actually really absolutely have a buffer

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

path: Make the static analyzer happy

I don't see how path can be NULL here, but let's rather be safe and
check for it.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

repo: Check if progress could be started okay

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

build: Correct evaluate some more return codes

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

digest: Don't assign r when it is never read

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jenkins: Add a pipeline

This is going to help us to achieve better code quality and
compatibility because we will automatically build Pakfire for a couple
of common distributions and run it through Clang's Static Analyzer.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

stripper: Identify source files

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

config: Avoid updating line when it is never being read back

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: daemon: Correctly catch daemon initialization errors

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

key: Refactor return values

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

path: Fix a potential memory leak if either path is not absolute

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pwd: Initialize array before using it

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

file: Cleanup initializing files

There was a potential memory leak here if the entry could not be
allocated.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

libpakfire: Initialize some return values that could be undefined

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pakfire: Align the disk reader with the writer

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

configure: Link against libdw

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

stripper: Find all ELF files

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

repo: Improve return code handling

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Fix return code if the archive could not be opened

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

stripper: Build scaffolding for a new stripper

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Build out dummy function to scan the payload

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Keep a reference to Pakfire

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Use convenience function to open and walk through the archive

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Log the format and compression

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Simplify opening an archive

There is nothing special needed now we clone the file descriptors. So
this is mainly just to make the file sticky.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Fix passing error codes when opening the archive

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Simplify the walk/filter callbacks

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Cleanup loop when searching for files

This mainly removes the ugly label jump.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: cookie: Rename __f to f now that variable is free again

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Streamline opening and reading from the archive

We will now always create a copy of the file descriptor so that we can
independently run mutliple operations on the same archive without
breaking anything.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Fetch the writer only once

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Move extraction code back to where it belongs

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Remove unnecessary underscores from functions

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

archive: Minor code cleanups

No functional changes.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

build: Lint the archives after build

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Check if package name contains whitespace

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Add result callback to print to the CLI

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Make this entirely private

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Store any results

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

cli: pakfire-builder: Add command to call the linter

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

linter: Base this on archives rather than packages

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pty: Hack to avoid lock-up when we don't have a newline

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jail: Create helper function to capture output

That way, we can have the length as an extra argument.

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

tests: Have pipe return an error if it could not write

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pty: Replace any CRNL with NL when capturing the output

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

pty: Configure like SSH does it

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>

jail: Immediately send SIGKILL when the timer has expired

Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>