debian.debfile module

Representation of Debian binary package (.deb) files

Debfile Classes

class debian.debfile.DebControl(member: ArMember)

Bases: DebPart

debcontrol() → Deb822

Return the debian/control as a Deb822 (a Debian-specific dict-like class) object.

For a string representation of debian/control try .get_content(‘control’)

md5sums(encoding=None, errors=None)
md5sums(encoding, errors=None)

Return a dictionary mapping filenames (of the data part) to md5sums. Fails if the control part does not contain a ‘md5sum’ file.

Keys of the returned dictionary are the left-hand side values of lines in the md5sums member of control.tar.gz, usually file names relative to the file system root (without heading ‘/’ or ‘./’).

The returned keys are Unicode objects if an encoding is specified, otherwise binary. The returned values are always Unicode.

scripts() → Dict[str, bytes]

Return a dictionary of maintainer scripts (postinst, prerm, …) mapping script names to script text.

class debian.debfile.DebData(member: ArMember)

Bases: DebPart

exception debian.debfile.DebError

Bases: ArError

class debian.debfile.DebFile(filename: str | Path | None = None, mode: str = 'r', fileobj: BinaryIO | None = None)

Bases: ArFile

Representation of a .deb file (a Debian binary package)

DebFile objects have the following (read-only) properties:
  • version debian .deb file format version (not related with the

    contained package version), 2.0 at the time of writing for all .deb packages in the Debian archive

  • data DebPart object corresponding to the data.tar.gz (or

    other compressed or uncompressed tar) archive contained in the .deb file

  • control DebPart object corresponding to the control.tar.gz (or

    other compressed tar) archive contained in the .deb file

__updatePkgName() → None
changelog() → Changelog | None

Return a Changelog object for the changelog.Debian.gz of the present .deb package. Return None if no changelog can be found.

close() → None
property control: DebControl
property data: DebData
debcontrol() → Deb822

See .control.debcontrol()

md5sums(encoding=None, errors=None)
md5sums(encoding, errors=None)

See .control.md5sums()

scripts() → Dict[str, bytes]

See .control.scripts()

property version: bytes
class debian.debfile.DebPart(member: ArMember)

Bases: object

‘Part’ of a .deb binary package.

A .deb package is considered as made of 2 parts: a ‘data’ part (corresponding to the possibly compressed ‘data.tar’ archive embedded in a .deb) and a ‘control’ part (the ‘control.tar.gz’ archive). Each of them is represented by an instance of this class. Each archive should be a compressed tar archive although an uncompressed data.tar is permitted; supported compression formats are: .tar.gz, .tar.bz2, .tar.xz .

When referring to file members of the underlying .tar.gz archive, file names can be specified in one of 3 formats “file”, “./file”, “/file”. In all cases the file is considered relative to the root of the archive. For the control part the preferred mechanism is the first one (as in deb.control.get_content(‘control’) ); for the data part the preferred mechanism is the third one (as in deb.data.get_file(‘/etc/vim/vimrc’) ).

__normalize_member() → str

try (not so hard) to obtain a member file name in a form that is stored in the .tar.gz, i.e. starting with ./

__resolve_symlinks(path: str) → str | None

walk the path following symlinks

returns:

resolved_path, info

if the path is not found even after following symlinks within the archive, then None is returned.

close() → None
get_content(fname, encoding=None, errors=None, follow_symlinks=False)
get_content(fname, encoding, errors=None, follow_symlinks=False)

Return the string content of a given file, or None (e.g. for directories).

If encoding is given, then the content will be a Unicode object; otherwise, it will contain binary data.

If follow_symlinks is True, then symlinks within the archive will be followed.

get_file(fname, encoding=None, errors=None, follow_symlinks=False)
get_file(fname, encoding, errors=None, follow_symlinks=False)

Return a file object corresponding to a given file name.

If encoding is given, then the file object will return Unicode data; otherwise, it will return binary data.

If follow_symlinks is True, then symlinks within the archive will be followed.

has_file(fname: str | Path, follow_symlinks: bool = False) → bool

Check if this part contains a given file name.

Symlinks within the archive can be followed.

tgz() → TarFile

Return a TarFile object corresponding to this part of a .deb package.

Despite the name, this method gives access to various kind of compressed tar archives, not only gzipped ones.

class debian.debfile._NormedFilePath

Bases: str

File path class for squashing ./ at the start of files for TarFile

This class only replaces the __eq__ method of the file path comparison so that “./” is always added to the start of the file path meaning that “is this the right file to extract” will be true when the code asks for “./control” but the control.tar member contains the file “control”

class debian.debfile._NormedTarInfo(name='')

Bases: TarInfo

TarInfo replacement that uses file path comparison that squashes ./

This class uses the above _NormedFilePath class for file paths when iterating or inspecting the members of a TarFile. The _NormedFilePath normalises the file path when inspecting or extracting from a .deb file.

The class replaces the ‘name’ member that is normally a str with a property that addresses a _NormedFilePath.

The control tarball members of the .deb may contain an optional entry for “.” and prefix all entries in the tarball with “./”. Archives generated by dpkg have this prefix but third party tools might not. The published spec for the .deb file format, deb(5) from dpkg-dev, explicitly permits this difference and thus it should be supported here.

See https://bugs.debian.org/1031674 for more detail.

_get_name() → str | None
_name: str | None = None
_set_name(name: str) → None
_sparse_structs
chksum

Header checksum.

devmajor

Device major number.

devminor

Device minor number.

gid

Group ID of the user who originally stored this member.

gname

Group name.

linkname

Name of the target file name, which is only present in TarInfo objects of type LNKTYPE and SYMTYPE.

mode

Permission bits.

mtime

Time of last modification.

name

Name of the archive member.

offset

The tar header starts here.

offset_data

The file’s data starts here.

pax_headers

A dictionary containing key-value pairs of an associated pax extended header.

size

Size in bytes.

sparse

Sparse member information.

tarfile
type

File type. type is usually one of these constants: REGTYPE, AREGTYPE, LNKTYPE, SYMTYPE, DIRTYPE, FIFOTYPE, CONTTYPE, CHRTYPE, BLKTYPE, GNUTYPE_SPARSE.

uid

User ID of the user who originally stored this member.

uname

User name.