|
Size: 2581
Comment: add
|
Size: 4431
Comment: clarify that we only care about source tarballs
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 1: | Line 1: |
| = Embedded source code copies = This file collects cases, where a source package embeds code from other projects, without linking dynamically: |
[[https://www.debian.org/doc/debian-policy/#convenience-copies-of-code|Debian Policy 4.13]] states that Debian packages should not use convenience copies. |
| Line 5: | Line 3: |
| xpdf code: (some use xpdf 2, some xpdf 3) gpdf pdftohtml kdegraphics/kpdf tetex-bin (the very latest tetex-bin started to use poppler) cupsys (only older releases, recent ones use xpdf-utils, it's still present in the src, though) poppler koffice libextractor |
Embedded code and data copies should be removed from the upstream VCS and source tarballs. Upstream might want to only embed the copies in the binary packages they distribute, script the install of their dependencies and or bundle the dependencies into a single but separate source tarball rather than embedding copies of them. Once upstream has fixed the issue, the Debian package can then be updated to the fixed version. If upstream refuse to remove the embedded copies, then Debian should either repack the upstream tarball using Files-Excluded (if there is a DFSG or size issue) or remove the files in `debian/rules clean` and very early in `debian/rules build`, so that there is no chance of them being used by the build process. |
| Line 15: | Line 5: |
| The list of packages that embed copies (including unused ones) of other projects is maintained in the security-tracker git repository. | |
| Line 16: | Line 7: |
| zlib code: (lots of apps embed a copy, but link dynamically, but there are a few exceptions) dpkg rsync (somehow derived code base) mozilla(?) Linux kernels pvpgn (links dynamically since 1.7.8-2) |
https://salsa.debian.org/security-tracker-team/security-tracker/raw/master/data/embedded-code-copies |
| Line 23: | Line 9: |
| This list also contains information about code forks so that the security team can check if all forks contain the same vulnerabilities. | |
| Line 24: | Line 11: |
| libgadu/ekg: centericq gaim kopete (ships the code, but links dynamically in the Debian package) kadu (not packaged in Debian) GNU gadu (not yet packaged in Debian) |
All Debian members have commit access to the security-tracker repository and others can send suggestions or additions to the [[DebianList:debian-security-tracker|debian-security-tracker mailing list]]. |
| Line 31: | Line 13: |
| Lintian detects embedding of [[https://lintian.debian.org/tags/embedded-feedparser-library.html|feedparser]], common [[https://lintian.debian.org/tags/embedded-javascript-library.html|JavaScript]]/[[https://lintian.debian.org/tags/embedded-library.html|C/C++]]/[[https://lintian.debian.org/tags/embedded-pear-module.html|PEAR]]/[[https://lintian.debian.org/tags/embedded-php-library.html|PHP]] libraries and !PostScript fragments ([[https://lintian.debian.org/tags/license-problem-font-adobe-copyrighted-fragment.html|1]] [[https://lintian.debian.org/tags/license-problem-font-adobe-copyrighted-fragment-no-credit.html|2]]). | |
| Line 32: | Line 15: |
| xmlrpc: (which package is the "origin" of this code?) drupal phpgroupware egroupware phpwiki php4 (php-pear, IIRC this was reorganized some weeks ago?) tikiwiki (not packaged in Debian) |
These wiki pages mention embedded code copies: [[arc4random]] |
| Line 40: | Line 17: |
| The [[dedup.debian.net|Debian duplication detector]] detects duplicate files in binary packages and may be useful for detecting verbatim duplication of interpreted code and data. | |
| Line 41: | Line 19: |
| shtool: (affects build-time only) mysql-ocaml php4 |
[[https://github.com/silviocesare/Clonewise|Clonewise]] is a tool not yet in Debian that [[https://lists.debian.org/debian-security/2012/07/msg00000.html|could be used to find unfixed vulnerabilities because of embedded code copies]]. [[https://github.com/Mondego/SourcererCC|SourcererCC]] is another tool for detecting embedded code copies. |
| Line 45: | Line 21: |
| The [[https://sources.debian.org/|Debian Sources website]] collects hashes and ctags of all Debian source code and allows [[https://sources.debian.org/advancedsearch/|searching]] for specific hashes and ctags, which may be useful for detecting duplication of source code and data. | |
| Line 46: | Line 23: |
| mozilla: mozilla-firefox mozilla-thunderbird nvu |
If you have a particular piece of code with some interesting aspect (security issue etc) you can likely find other copies using the [[DebianCodeSearch|Debian code search site]] or external code search engines such as [[https://code.ohloh.net/|Ohloh code]], [[https://searchcode.com/|searchcode]] and [[https://github.com/|GitHub]]. |
| Line 51: | Line 25: |
| If a file containing some code has a fairly unique name, you can often find copies of that file by searching the contents of Debian binary or source packages using apt-file: | |
| Line 52: | Line 27: |
| xli: xloadimage |
{{{ apt-file search uniquename.py apt-file search -I dsc uniquename.c }}} |
| Line 55: | Line 32: |
| Various Debian folks keep track of embedded code copies they found via usertags: | |
| Line 56: | Line 34: |
| lesstif: (beware: two different lesstif APIs supported in one package, MOTIF 1.2 discarded upstream) openmotif xfree86/xorg (in libxpm) |
[[https://udd.debian.org/cgi-bin/bts-usertags.cgi?tag=embedded-code-copy&user=rbrito@ime.usp.br|rbrito@ime.usp.br]] [[https://udd.debian.org/cgi-bin/bts-usertags.cgi?tag=embedded-code-copy&user=jwilk@debian.org|jwilk@debian.org]] [[https://udd.debian.org/cgi-bin/bts-usertags.cgi?tag=embed&user=pabs@debian.org|pabs@debian.org]] [[https://udd.debian.org/cgi-bin/bts-usertags.cgi?tag=embedded-synctex-parser&user=sramacher@debian.org|sramacher@debian.org]] [[https://udd.debian.org/cgi-bin/bts-usertags.cgi?tag=embed&user=dr@jones.dk|dr@jones.dk]] |
| Line 60: | Line 40: |
| = See also = | |
| Line 61: | Line 42: |
| kerberized apps with BSD origin: krb4 krb5 heimdal grip: (which pkg is the origin?) libcdaudio grip gnome-vfs (vfs2 as well?) fudforum: phpgroupware-fudforum egroupware-fudforum cvs: gcvs (at least an additional script is included, check if there's more) pcre: all pythons php4 (src included, but Debian package links dynamically) analog (src included, but Debian package links dynamically) libgoffice-1 tf5 (since 5.0beta7 the Debian package links dynamically) tiff: wxpythongtk (check, which debian pkg this is in) older kdegraphics/kpdf releases < 3.3 embedded a copy uudeview: libconvert-uulib-perl sqlite: (not affected by security vulnerabilities so far) amarok util-linux/mount: loop-aes-utils contains code from util-linux' mount in the mount-aes-udeb webmin: usermin sylpheed: sylpheed-claws phpsysinfo: egroupware phpgroupware phpldapadmin: egroupware chmlib: kchmviewer (not packaged in Debian) libavcodec/libavformat: ffmpeg xine-lib xvidcap (currently in NEW) kino(?) gst-ffmpeg xmovie (currently in NEW) mad MPEG decoding lib: mad xine-lib libdts: libdts xine-lib flac: flac xine-lib liba52: a52dec xine-lib libmpeg2: mpeg2dec xine-lib curl: wget (code for NTLM authentication) |
* [[https://fedoraproject.org/wiki/Packaging:Guidelines#Bundling_and_Duplication_of_system_libraries|Fedora policy]] * [[https://wiki.gentoo.org/wiki/Why_not_bundle_dependencies|Gentoo policy]] |
Debian Policy 4.13 states that Debian packages should not use convenience copies.
Embedded code and data copies should be removed from the upstream VCS and source tarballs. Upstream might want to only embed the copies in the binary packages they distribute, script the install of their dependencies and or bundle the dependencies into a single but separate source tarball rather than embedding copies of them. Once upstream has fixed the issue, the Debian package can then be updated to the fixed version. If upstream refuse to remove the embedded copies, then Debian should either repack the upstream tarball using Files-Excluded (if there is a DFSG or size issue) or remove the files in debian/rules clean and very early in debian/rules build, so that there is no chance of them being used by the build process.
The list of packages that embed copies (including unused ones) of other projects is maintained in the security-tracker git repository.
https://salsa.debian.org/security-tracker-team/security-tracker/raw/master/data/embedded-code-copies
This list also contains information about code forks so that the security team can check if all forks contain the same vulnerabilities.
All Debian members have commit access to the security-tracker repository and others can send suggestions or additions to the debian-security-tracker mailing list.
Lintian detects embedding of feedparser, common JavaScript/C/C++/PEAR/PHP libraries and PostScript fragments (1 2).
These wiki pages mention embedded code copies: arc4random
The Debian duplication detector detects duplicate files in binary packages and may be useful for detecting verbatim duplication of interpreted code and data.
Clonewise is a tool not yet in Debian that could be used to find unfixed vulnerabilities because of embedded code copies. SourcererCC is another tool for detecting embedded code copies.
The Debian Sources website collects hashes and ctags of all Debian source code and allows searching for specific hashes and ctags, which may be useful for detecting duplication of source code and data.
If you have a particular piece of code with some interesting aspect (security issue etc) you can likely find other copies using the Debian code search site or external code search engines such as Ohloh code, searchcode and GitHub.
If a file containing some code has a fairly unique name, you can often find copies of that file by searching the contents of Debian binary or source packages using apt-file:
apt-file search uniquename.py apt-file search -I dsc uniquename.c
Various Debian folks keep track of embedded code copies they found via usertags:
rbrito@ime.usp.br jwilk@debian.org pabs@debian.org sramacher@debian.org dr@jones.dk