.gitlab-ci.yml

 variables:
   CONTAINER_REGISTRY: $CI_REGISTRY/georg/mat2-ci-images
+  GIT_DEPTH: "5"
+  GIT_STRATEGY: clone
 
 stages:
   - linting
@@ -9,70 +11,53 @@ stages:
   before_script:  # This is needed to not run the testsuite as root
     - useradd --home-dir ${CI_PROJECT_DIR} mat2
     - chown -R mat2 .
-
-linting:bandit:
-  image: $CONTAINER_REGISTRY:linting 
-  stage: linting
-  script:  # TODO: remove B405 and B314
-    - bandit ./mat2 --format txt --skip B101
-    - bandit -r ./nautilus/ --format txt --skip B101
-    - bandit -r ./libmat2 --format txt --skip B101,B404,B603,B405,B314
-
-linting:codespell:
-  image: $CONTAINER_REGISTRY:linting
-  stage: linting
-  script:
-    # Run codespell to check for spelling errors; ignore errors about binary
-    # files, use a config with ignored words and exclude the git directory,
-    # which might contain false positives
-    - codespell -q 2 -I utils/ci/codespell/ignored_words.txt -S .git
   
-linting:pylint:
-  image: $CONTAINER_REGISTRY:linting
-  stage: linting
-  script:
-    - pylint3 --disable=no-else-return --extension-pkg-whitelist=cairo,gi ./libmat2 ./mat2
-    # Once nautilus-python is in Debian, decomment it form the line below
-    - pylint3 --disable=no-else-return --extension-pkg-whitelist=Nautilus,GObject,Gtk,Gio,GLib,gi ./nautilus/mat2.py
-
-linting:pyflakes:
+linting:ruff:
   image: $CONTAINER_REGISTRY:linting
   stage: linting
   script:
-    - pyflakes3 ./libmat2 ./mat2 ./tests/ ./nautilus
+    - apt update
+    - apt install -qqy --no-install-recommends python3-venv
+    - python3 -m venv venv
+    - source venv/bin/activate
+    - pip3 install ruff
+    - ruff check .
 
 linting:mypy:
   image: $CONTAINER_REGISTRY:linting
   stage: linting
   script:
-    - mypy --ignore-missing-imports mat2 libmat2/*.py ./nautilus/mat2.py
+    - mypy --ignore-missing-imports mat2 libmat2/*.py
 
 tests:archlinux:
   image: $CONTAINER_REGISTRY:archlinux
   stage: test
   script:
-    - python3 setup.py test
+    - python3 -m unittest discover -v
   
 tests:debian:
   image: $CONTAINER_REGISTRY:debian
   stage: test
+  <<: *prepare_env
   script:
     - apt-get -qqy purge bubblewrap
-    - python3 setup.py test
+    - su - mat2 -c "python3-coverage run --branch -m unittest discover -s tests/"
+    - su - mat2 -c "python3-coverage report --fail-under=95 -m --include 'libmat2/*'"
 
 tests:debian_with_bubblewrap:
   image: $CONTAINER_REGISTRY:debian
   stage: test
+  allow_failure: true
   <<: *prepare_env
   script:
-    - su - mat2 -c "python3-coverage run --branch -m unittest discover -s tests/"
-    - su - mat2 -c "python3-coverage report --fail-under=100 -m --include 'libmat2/*'"
+    - apt-get -qqy install bubblewrap
+    - python3 -m unittest discover -v
 
 tests:fedora:
   image: $CONTAINER_REGISTRY:fedora
   stage: test
   script:
-    - python3 setup.py test
+    - python3 -m unittest discover -v
 
 tests:gentoo:
   image: $CONTAINER_REGISTRY:gentoo
@@ -80,3 +65,51 @@ tests:gentoo:
   <<: *prepare_env
   script:
     - su - mat2 -c "python3 -m unittest discover -v"
+
+tests:python3.7:
+  image: $CONTAINER_REGISTRY:python3.7
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.8:
+  image: $CONTAINER_REGISTRY:python3.8
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.9:
+  image: $CONTAINER_REGISTRY:python3.9
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.10:
+  image: $CONTAINER_REGISTRY:python3.10
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.11:
+  image: $CONTAINER_REGISTRY:python3.11
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.12:
+  image: $CONTAINER_REGISTRY:python3.12
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.13:
+  image: $CONTAINER_REGISTRY:python3.13
+  stage: test
+  script:
+    - python3 -m unittest discover -v
+
+tests:python3.14:
+  image: $CONTAINER_REGISTRY:python3.14
+  stage: test
+  script:
+    - python3 -m unittest discover -v

.pylintrc

@@ -14,4 +14,5 @@ disable=
     catching-non-exception,
     cell-var-from-loop,
     locally-disabled,
+		raise-missing-from,
     invalid-sequence-index,  # pylint doesn't like things like `Tuple[int, bytes]` in type annotation

CHANGELOG.md

+# 0.13.5 - 2025-01-09
+- Keep orientation metadata on jpeg and tiff files
+- Improve cairo-related error/exceptions handling
+- Improve the logging
+- Improve the sandboxing
+- Improve Python3.12 support
+- Improve MSOffice documents handling
+
+# 0.13.4 - 2023-08-02
+
+- Add documentation about mat2 on OSX
+- Make use of python3.7 constructs to simplify code
+- Use moderner type annotations
+- Harden get_meta in archive.py against variants of CVE-2022-35410 
+- Improve MSOffice document support
+- Package the manpage on pypi 
+
+# 0.13.3 - 2023-02-23
+
+- Fix a decorator argument
+
+# 0.13.2 - 2023-01-28
+
+- Fix a crash on some python versions
+
+# 0.13.1 - 2023-01-07
+
+- Improve xlsx support
+- Remove the Nautilus extension
+
+# 0.13.0 - 2022-07-06
+
+- Fix an arbitrary file read (CVE-2022-35410)
+- Add support for heic files 
+
+# 0.12.4 - 2022-04-30
+
+- Fix possible errors/crashes when processing multiple files
+  via the command line interface
+- Use a fixed PDF version for the output
+- Improve compatibility with modern versions of rsvg
+- Improve the robustness of the command line interface with
+  regard to control characters
+
+# 0.12.3 - 2022-01-06
+
+- Implement code for internationalization
+- Keep individual files compression type in zip files
+- Increase the robustness of mat2 against weird/corrupted files
+- Fix the dolphin integration
+- Add a fuzzer
+
+# 0.12.2 - 2021-08-29
+
+- Add support for aiff files
+- Improve MS Office support
+- Improve compatibility with newer/older version of mat2's dependencies
+- Fix possible issues with the resolution of processed pdf
+
+# 0.12.1 - 2021-03-19
+
+- Improve epub support
+- Improve MS Office support
+
+# 0.12.0 - 2020-12-18
+
+- Improve significantly MS Office formats support
+- Fix some typos in the Nautilus extension
+- Improve reliability of the mp3, pdf and svg parsers
+- Improve compatibility with ffmpeg when sandboxing is used
+- Improve the dolphin extension usability
+- libmat2 now raises a ValueError on malformed files while trying to 
+  find the right parser, instead of returning None
+
+# 0.11.0 - 2020-03-29
+
+- Improve significantly MS Office formats support
+- Refactor how mat2 looks for executables
+
+# 0.10.1 - 2020-02-09
+
+- Improve the documentation and the manpage
+- Improve the robustness of css, html, png, gdk-based, exiftool-based parsers
+- Future-proof a bit the testsuite
+- Handle tiff files with a .tif extension
+- Improve the sandbox' usability
+- Add support for wav files
+
+# 0.10.0 - 2019-11-30
+
+- Make mat2 work on Python3.8
+- Minor improvement of ppt handling
+- Minor improvement of odt handling
+- Add an integration KDE's file manager: Dolphin
+- mat2 now copies file permissions on the cleaned files
+- Add a flag to disable sandboxing
+- Tighten a bit the sandboxing
+- Improve handling of MSOffice documents
+- Add support for inplace cleaning
+- Better handling of mutually-exclusive arguments in the command line
+- Add support for svg
+- Add support for ppm
+- Cleaned zip files are compressed by default
+- Minor performances improvement when dealing with images/video files
+- Better handling of optional dependencies
+
 # 0.9.0 - 2019-05-10
 
 - Add tar/tar.gz/tar.bz2/tar.zx archives support
@@ -67,12 +173,12 @@
 
 # 0.3.1 - 2018-09-01
 
-- Document how to install MAT2 for various distributions
+- Document how to install mat2 for various distributions
 - Fix various typos in the documentation/comments
-- Add ArchLinux to the CI to ensure that MAT2 is running on it
+- Add ArchLinux to the CI to ensure that mat2 is running on it
 - Fix the handling of files with a name ending in `.JPG`
 - Improve the detection of unsupported extensions in upper-case
-- Streamline MAT2's logging
+- Streamline mat2's logging
 
 
 # 0.3.0 - 2018-08-03
@@ -92,14 +198,14 @@
 - Simplify various code-paths
 - Remove superfluous debug message
 - Remove the `--check` option that never was implemented anyway
-- Add a `-c` option to check for MAT2's dependencies
+- Add a `-c` option to check for mat2's dependencies
 
 
 # 0.1.3 - 2018-07-06
 
-- Improve MAT2 resilience against corrupted images
+- Improve mat2 resilience against corrupted images
 - Check that the minimal version of Poppler is available
-- Simplify how MAT2 deals with office files
+- Simplify how mat2 deals with office files
 - Improve cleaning of office files
 	- Thumbnails are removed
 	- Revisions are removed
@@ -111,8 +217,8 @@
 - Rename some files to ease the packaging
 - Add linters to the CI (mypy, bandit and pyflakes)
 - Prevent exitftool-related parameters injections
-- Improve MAT2's resilience against corrupted files
-- Make MAT2 work on fedora, thanks to @atenart
+- Improve mat2's resilience against corrupted files
+- Make mat2 work on fedora, thanks to @atenart
 - Tighten the threat model
 - Simplify and improve how office files are handled
 

CONTRIBUTING.md

-# Contributing to MAT2
+# Contributing to mat2
 
-The main repository for MAT2 is on [0xacab]( https://0xacab.org/jvoisin/mat2 ),
+The main repository for mat2 is on [0xacab]( https://0xacab.org/jvoisin/mat2 ),
 but you can send patches to jvoisin by [email](https://dustri.org/) if you prefer.
 
 Do feel free to pick up [an issue]( https://0xacab.org/jvoisin/mat2/issues )
-and to send a pull-request. Please do check that everything is fine by running the
-testsuite with `python3 -m unittest discover -v` before submitting one :)
+and to send a pull-request.
+
+Before sending the pull-request, please do check that everything is fine by
+running the full test suite in GitLab. To do that, after forking mat2 in GitLab,
+you need to go in Settings -> CI/CD -> Runner and there enable shared runners.
+
+Mat2 also has unit tests (that are also run in the full test suite). You can run
+them with `python3 -m unittest discover -v`.
 
 If you're fixing a bug or adding a new feature, please add tests accordingly,
 this will greatly improve the odds of your merge-request getting merged.
@@ -16,7 +22,7 @@ If you're adding a new fileformat, please add tests for:
 2. Cleaning metadata
 3. Raising `ValueError` upon a corrupted file
 
-Since MAT2 is written in Python3, please conform as much as possible to the
+Since mat2 is written in Python3, please conform as much as possible to the
 [pep8]( https://pep8.org/ ) style; except where it makes no sense of course.
 
 # Doing a release
@@ -24,15 +30,16 @@ Since MAT2 is written in Python3, please conform as much as possible to the
 1. Update the [changelog](https://0xacab.org/jvoisin/mat2/blob/master/CHANGELOG.md)
 2. Update the version in the [mat2](https://0xacab.org/jvoisin/mat2/blob/master/mat2) file
 3. Update the version in the [setup.py](https://0xacab.org/jvoisin/mat2/blob/master/setup.py) file
-4. Update the version and date in the [man page](https://0xacab.org/jvoisin/mat2/blob/master/doc/mat2.1)
-5. Commit the changelog, man page, mat2 and setup.py files
-6. Create a tag with `git tag -s $VERSION`
-7. Push the commit with `git push origin master`
-8. Push the tag with `git push --tags`
-9. Download the gitlab archive of the release
-10. Diff it against the local copy
-11. If there is no difference, sign the archive with `gpg --armor --detach-sign mat2-$VERSION.tar.xz`
-12. Upload the signature on Gitlab's [tag page](https://0xacab.org/jvoisin/mat2/tags) and add the changelog there
-13. Announce the release on the [mailing list](https://mailman.boum.org/listinfo/mat-dev)
-14. Sign'n'upload the new version on pypi with `python3 setup.py sdist bdist_wheel` then `twine upload -s dist/*`
-15. Do the secret release dance
+4. Update the version in the [pyproject.toml](https://0xacab.org/jvoisin/mat2/blob/master/yproject.toml) file
+5. Update the version and date in the [man page](https://0xacab.org/jvoisin/mat2/blob/master/doc/mat2.1)
+6. Commit the modified files
+7. Create a tag with `git tag -s $VERSION`
+8. Push the commit with `git push origin master`
+9. Push the tag with `git push --tags`
+10. Download the gitlab archive of the release
+11. Diff it against the local copy
+12. If there is no difference, sign the archive with `gpg --armor --detach-sign mat2-$VERSION.tar.xz`
+13. Upload the signature on Gitlab's [tag page](https://0xacab.org/jvoisin/mat2/tags) and add the changelog there
+14. Announce the release on the [mailing list](https://mailman.boum.org/listinfo/mat-dev)
+15. Sign'n'upload the new version on pypi with `python3 setup.py sdist bdist_wheel` then `twine upload -s dist/*`
+16. Do the secret release dance

INSTALL.md

@@ -14,38 +14,57 @@ pip3 install mat2
 ## Optional dependencies
 
 When [bubblewrap](https://github.com/projectatomic/bubblewrap) is
-installed, MAT2 uses it to sandbox any external processes it invokes.
+installed, mat2 uses it to sandbox any external processes it invokes.
 
 ## Arch Linux
 
-Thanks to [Francois_B](https://www.sciunto.org/), there is an package available on
-[Arch linux's AUR](https://aur.archlinux.org/packages/mat2/).
+Thanks to [kpcyrd](https://archlinux.org/packages/?maintainer=kpcyrd), there is an package available on
+[Arch linux's AUR](https://archlinux.org/packages/extra/any/mat2/).
 
 ## Debian
 
-There is a package available in [Debian](https://packages.debian.org/search?keywords=mat2&searchon=names&section=all).
+There is a package available in [Debian](https://packages.debian.org/search?keywords=mat2&searchon=names&section=all) and you can install mat2 with:
+
+```
+apt install mat2
+```
 
 ## Fedora
 
 Thanks to [atenart](https://ack.tf/), there is a package available on
 [Fedora's copr]( https://copr.fedorainfracloud.org/coprs/atenart/mat2/ ).
 
-We use copr (cool other packages repo) as the Mat2 Nautilus plugin depends on
-python3-nautilus, which isn't available yet in Fedora (but is distributed
-through this copr).
-
-First you need to enable Mat2's copr:
+First you need to enable mat2's copr:
 
 ```
 dnf -y copr enable atenart/mat2
 ```
 
-Then you can install both the Mat2 command and Nautilus extension:
+Then you can install mat2:
 
 ```
-dnf -y install mat2 mat2-nautilus
+dnf -y install mat2
 ```
 
 ## Gentoo
 
-MAT2 is available in the [torbrowser overlay](https://github.com/MeisterP/torbrowser-overlay).
+mat2 is available in the [torbrowser overlay](https://github.com/MeisterP/torbrowser-overlay).
+
+
+# OSX
+
+## Homebrew
+
+mat2 is [available on homebrew](https://formulae.brew.sh/formula/mat2):
+
+```
+brew install mat2
+```
+
+## MacPorts
+
+mat2 is [available on MacPorts](https://ports.macports.org/port/mat2/):
+
+```
+port install mat2
+```

README.md

 ```
  _____ _____ _____ ___
 |     |  _  |_   _|_  |  Keep your data,
-| | | |     | | | |  _|     trash your meta!
-|_|_|_|__|__| |_| |___|
+| | | | |_| | | | |  _|     trash your meta!
+|_|_|_|_| |_| |_| |___|
 
 ```
 
-This software is currently in **beta**, please don't use it for anything
-critical.
-
 # Metadata and privacy
 
 Metadata consist of information that characterizes data.
@@ -22,11 +19,17 @@ camera was used. Office documents like PDF or Office automatically adds
 author and company information to documents and spreadsheets.
 Maybe you don't want to disclose those information.
 
-This is precisely the job of MAT2: getting rid, as much as possible, of
+This is precisely the job of mat2: getting rid, as much as possible, of
 metadata.
 
-MAT2 provides both a command line tool, and a graphical user interface
-via an extension for Nautilus, the default file manager of GNOME.
+mat2 provides:
+- a library called `libmat2`;
+- a command line tool called `mat2`,
+- a service menu for Dolphin, KDE's default file manager
+
+If you prefer a regular graphical user interface, you might be interested in
+[Metadata Cleaner](https://metadatacleaner.romainvigier.fr/), which is using
+`mat2` under the hood.
 
 # Requirements
 
@@ -38,7 +41,13 @@ via an extension for Nautilus, the default file manager of GNOME.
 - `libimage-exiftool-perl` for everything else
 - `bubblewrap`, optionally, for sandboxing
 
-Please note that MAT2 requires at least Python3.5.
+Please note that mat2 requires at least Python3.5.
+
+# Requirements setup on macOS (OS X) using [Homebrew](https://brew.sh/)
+
+```bash
+brew install exiftool cairo pygobject3 poppler gdk-pixbuf librsvg ffmpeg
+```
 
 # Running the test suite
 
@@ -53,11 +62,11 @@ $ python3-coverage run --branch -m unittest discover -s tests/
 $ python3-coverage report --include -m --include /libmat2/*'
 ```
 
-# How to use MAT2
+# How to use mat2
 
-```bash
-usage: mat2 [-h] [-v] [-l] [--check-dependencies] [-V]
-            [--unknown-members policy] [-s | -L]
+```
+usage: mat2 [-h] [-V] [--unknown-members policy] [--inplace] [--no-sandbox]
+            [-v] [-l] [--check-dependencies] [-L | -s]
             [files [files ...]]
 
 Metadata anonymisation toolkit 2
@@ -67,27 +76,49 @@ positional arguments:
 
 optional arguments:
   -h, --help            show this help message and exit
-  -v, --version         show program's version number and exit
-  -l, --list            list all supported fileformats
-  --check-dependencies  check if MAT2 has all the dependencies it needs
   -V, --verbose         show more verbose status information
   --unknown-members policy
                         how to handle unknown members of archive-style files
                         (policy should be one of: abort, omit, keep) [Default:
                         abort]
-  -s, --show            list harmful metadata detectable by MAT2 without
-                        removing them
+  --inplace             clean in place, without backup
+  --no-sandbox          Disable bubblewrap's sandboxing
+  -v, --version         show program's version number and exit
+  -l, --list            list all supported fileformats
+  --check-dependencies  check if mat2 has all the dependencies it needs
   -L, --lightweight     remove SOME metadata
+  -s, --show            list harmful metadata detectable by mat2 without
+                        removing them
 ```
 
-Note that MAT2 **will not** clean files in-place, but will produce, for
+Note that mat2 **will not** clean files in-place, but will produce, for
 example, with a file named "myfile.png" a cleaned version named
 "myfile.cleaned.png".
 
+## Web interface
+
+It's possible to run mat2 as a web service, via
+[mat2-web](https://0xacab.org/jvoisin/mat2-web).
+
+If you're using WordPress, you might be interested in [wp-mat](https://git.autistici.org/noblogs/wp-mat)
+and [wp-mat-server](https://git.autistici.org/noblogs/wp-mat-server).
+
+## Desktop GUI
+
+For GNU/Linux desktops, it's possible to use the
+[Metadata Cleaner](https://gitlab.com/rmnvgr/metadata-cleaner) GTK application.
+
+# Supported formats
+
+The following formats are supported: avi, bmp, css, epub/ncx, flac, gif, jpeg,
+m4a/mp2/mp3/…, mp4, odc/odf/odg/odi/odp/ods/odt/…, off/opus/oga/spx/…, pdf,
+png, ppm, pptx/xlsx/docx/…, svg/svgz/…, tar/tar.gz/tar.bz2/tar.xz/…, tiff,
+torrent, wav, wmv, zip, …
+  
 # Notes about detecting metadata
 
-While MAT2 is doing its very best to display metadata when the `--show` flag is
-passed, it doesn't mean that a file is clean from any metadata if MAT2 doesn't
+While mat2 is doing its very best to display metadata when the `--show` flag is
+passed, it doesn't mean that a file is clean from any metadata if mat2 doesn't
 show any. There is no reliable way to detect every single possible metadata for
 complex file formats.
 
@@ -113,15 +144,21 @@ of the guarantee that mat2 won't modify the data of their files, there is the
 	watermarks from PDF.
 - [Scrambled Exif](https://f-droid.org/packages/com.jarsilio.android.scrambledeggsif/),
 	an open-source Android application to remove metadata from pictures.
+- [Dangerzone](https://dangerzone.rocks/), designed to sanitize harmful documents
+  into harmless ones.
 
 # Contact
 
 If possible, use the [issues system](https://0xacab.org/jvoisin/mat2/issues)
-or the [mailing list](https://mailman.boum.org/listinfo/mat-dev)
+or the [mailing list](https://www.autistici.org/mailman/listinfo/mat-dev)
 Should a more private contact be needed (eg. for reporting security issues),
 you can email Julien (jvoisin) Voisin at `julien.voisin+mat2@dustri.org`,
 using the gpg key `9FCDEE9E1A381F311EA62A7404D041E8171901CC`.
 
+# Donations
+
+If you want to donate some money, please give it to [Tails]( https://tails.boum.org/donate/?r=contribute ).
+
 # License
 
 This program is free software: you can redistribute it and/or modify
@@ -137,16 +174,20 @@ GNU General Public License for more details.
 You should have received a copy of the GNU Lesser General Public License
 along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-Copyright 2018 Julien (jvoisin) Voisin <julien.voisin+mat2@dustri.org>
-Copyright 2016 Marie Rose for MAT2's logo
+Copyright 2018 Julien (jvoisin) Voisin <julien.voisin+mat2@dustri.org>  
+Copyright 2016 Marie-Rose for mat2's logo
+
+The `tests/data/dirty_with_nsid.docx` file is licensed under GPLv3,
+and was borrowed from the Calibre project: https://calibre-ebook.com/downloads/demos/demo.docx
+
+The `narrated_powerpoint_presentation.pptx` file is in the public domain.
 
 # Thanks
 
-MAT2 wouldn't exist without:
+mat2 wouldn't exist without:
 
 - the [Google Summer of Code](https://summerofcode.withgoogle.com/);
 - the fine people from [Tails]( https://tails.boum.org);
 - friends
 
 Many thanks to them!
-

doc/implementation_notes.md

@@ -4,7 +4,7 @@ Implementation notes
 Lightweight cleaning mode
 -------------------------
 
-Due to *popular* request, MAT2 is providing a *lightweight* cleaning mode,
+Due to *popular* request, mat2 is providing a *lightweight* cleaning mode,
 that only cleans the superficial metadata of your file, but not
 the ones that might be in **embedded** resources. Like for example,
 images in a PDF or an office document.
@@ -19,7 +19,7 @@ are entirely removed.
 	deleted. For example journalists that are editing a document to erase
 	mentions sources mentions.
 
-- Or they are aware of it, and will likely not expect MAT2 to be able to keep
+- Or they are aware of it, and will likely not expect mat2 to be able to keep
 	the revisions, that are basically traces about how, when and who edited the
 	document.
 
@@ -27,15 +27,15 @@ are entirely removed.
 Race conditions
 ---------------
 
-MAT2 does its very best to avoid crashing at runtime. This is why it's checking
-if the file is valid __at parser creation__. MAT2 doesn't take any measure to
+mat2 does its very best to avoid crashing at runtime. This is why it's checking
+if the file is valid __at parser creation__. mat2 doesn't take any measure to
 ensure that the file is not changed between the time the parser is
 instantiated, and the call to clean or show the metadata.
 
 Symlink attacks
 ---------------
 
-MAT2 output predictable filenames (like yourfile.jpg.cleaned).
+mat2 output predictable filenames (like yourfile.jpg.cleaned).
 This may lead to symlink attack. Please check if you OS prevent
 against them
 
@@ -65,10 +65,10 @@ didn't remove any *deep metadata*, like the ones in embedded pictures. This was
 on of the reason MAT was abandoned: the absence of satisfying solution to
 handle PDF. But apparently, people are ok with [pdf redact
 tools](https://github.com/firstlookmedia/pdf-redact-tools), that simply
-transform the PDF into images. So this is what's MAT2 is doing too.
+transform the PDF into images. So this is what's mat2 is doing too.
 
 Of course, it would be possible to detect images in PDf file, and process them
-with MAT2, but since a PDF can contain a lot of things, like images, videos,
+with mat2, but since a PDF can contain a lot of things, like images, videos,
 javascript, pdf, blobs, … this is the easiest and safest way to clean them.
 
 Images handling
@@ -81,7 +81,7 @@ XML attacks
 -----------
 
 Since our threat model conveniently excludes files crafted to specifically
-bypass MAT2, fileformats containing harmful XML are out of our scope.
-But since MAT2 is using [etree](https://docs.python.org/3/library/xml.html#xml-vulnerabilities)
+bypass mat2, fileformats containing harmful XML are out of our scope.
+But since mat2 is using [etree](https://docs.python.org/3/library/xml.html#xml-vulnerabilities)
 to process XML, it's "only" vulnerable to DoS, and not memory corruption:
 odds are that the user will notice that the cleaning didn't succeed.

doc/mat2.1

-.TH MAT2 "1" "May 2019" "MAT2 0.9.0" "User Commands"
+.TH mat2 "1" "January 2025" "mat2 0.13.5" "User Commands"
 
 .SH NAME
 mat2 \- the metadata anonymisation toolkit 2
@@ -32,7 +32,7 @@ show program's version number and exit
 list all supported fileformats
 .TP
 \fB\-\-check\-dependencies\fR
-check if MAT2 has all the dependencies it needs
+check if mat2 has all the dependencies it needs
 .TP
 \fB\-V\fR, \fB\-\-verbose\fR
 show more verbose status information
@@ -41,11 +41,16 @@ show more verbose status information
 how to handle unknown members of archive-style files (policy should be one of: abort, omit, keep)
 .TP
 \fB\-s\fR, \fB\-\-show\fR
-list harmful metadata detectable by MAT2 without
-removing them
+list harmful metadata detectable by mat2 without removing them
 .TP
 \fB\-L\fR, \fB\-\-lightweight\fR
 remove SOME metadata
+.TP
+\fB\--no-sandbox\fR
+disable bubblewrap's sandboxing
+.TP
+\fB\--inplace\fR
+clean in place, without backup
 
 .SH EXAMPLES
 To remove all the metadata from a PDF file:
@@ -57,6 +62,24 @@ mat2 ./myfile.pdf
 .fi
 .PP
 
+.SH NOTES ABOUT METADATA
+
+While mat2 is doing its very best to display metadata when the --show flag is
+passed, it doesn't mean that a file is clean from any metadata if mat2 doesn't
+show any. There is no reliable way to detect every single possible metadata for
+complex file formats.
+.PP
+This is why you shouldn't rely on metadata's presence to decide if your file must
+be cleaned or not.
+.PP
+Moreover, mat2 goes to great lengths to make sure that as much metadata as
+possible are removed. This might sometimes result in a loss of quality of the
+processed files. For example, textual based pdf file converted into image based
+one means that it'll be no longer possible to select text in them. If you're
+experiencing this, you might want to give the lightweight cleaning mode a try,
+but keep in mind by doing so, some metadata \fBwon't be cleaned\fR.
+
+
 .SH BUGS
 
 While mat2 does its very best to remove every single metadata,

doc/threat_model.md

@@ -3,7 +3,7 @@ Threat Model
 
 The Metadata Anonymisation Toolkit 2 adversary has a number
 of goals, capabilities, and counter-attack types that can be
-used to guide us towards a set of requirements for the MAT2.
+used to guide us towards a set of requirements for the mat2.
 
 This is an overhaul of MAT's (the first iteration of the software) one.
 
@@ -53,7 +53,7 @@ Adversary
     user. This is the strongest position for the adversary to
     have. In this case, the adversary is capable of inserting
     arbitrary, custom watermarks specifically for tracking
-    the user. In general, MAT2 cannot defend against this
+    the user. In general, mat2 cannot defend against this
     adversary, but we list it for completeness' sake.
 
     - The adversary created the document for a group of users.
@@ -65,7 +65,7 @@ Adversary
 		- The adversary did not create the document, the weakest
 		position for the adversary to have. The file format is
 		(most of the time) standard, nothing custom is added:
-		MAT2 must be able to remove all metadata from the file.
+		mat2 must be able to remove all metadata from the file.
 
 
 Requirements
@@ -73,28 +73,28 @@ Requirements
 
 * Processing
 
-    - MAT2 *should* avoid interactions with information.
+    - mat2 *should* avoid interactions with information.
     Its goal is to remove metadata, and the user is solely
     responsible for the information of the file.
 
-    - MAT2 *must* warn when encountering an unknown
-    format. For example, in a zipfile, if MAT2 encounters an
+    - mat2 *must* warn when encountering an unknown
+    format. For example, in a zipfile, if mat2 encounters an
     unknown format, it should warn the user, and ask if the
     file should be added to the anonymised archive that is
     produced.
 
-    - MAT2 *must* not add metadata, since its purpose is to
+    - mat2 *must* not add metadata, since its purpose is to
     anonymise files: every added items of metadata decreases
     anonymity.
 
-    - MAT2 *should* handle unknown/hidden metadata fields,
+    - mat2 *should* handle unknown/hidden metadata fields,
     like proprietary extensions of open formats.
 
-		- MAT2 *must not* fail silently. Upon failure,
-		MAT2 *must not* modify the file in any way.
+		- mat2 *must not* fail silently. Upon failure,
+		mat2 *must not* modify the file in any way.
 
-		- MAT2 *might* leak the fact that MAT2 was used on the file,
+		- mat2 *might* leak the fact that mat2 was used on the file,
 		since it might be uncommon for some file formats to come
 		without any kind of metadata, an adversary might suspect that
-		the user used MAT2 on certain files.
+		the user used mat2 on certain files.
 

dolphin/README.md

+Dolphin integration
+===================
+
+Thanks to [Miguel Marco](https://riemann.unizar.es/~mmarco/), here is an neat
+integration for [Dolphin](https://kde.org/applications/system/org.kde.dolphin),
+the KDE file manager:
+
+1. Add the `mat2.desktop` file either in
+	- `/usr/share/kservices5/ServiceMenus/` to install it globally
+	- `~/.local/share/kservices5/ServiceMenus/` for a specific user
+2. Run `kbuildsycoca5` to update the corresponding database
+3. Enjoy your new contextual menu to remove metadata from your files!
+
+

dolphin/mat2.desktop

+[Desktop Entry]
+X-KDE-ServiceTypes=KonqPopupMenu/Plugin
+MimeType=application/pdf;application/vnd.oasis.opendocument.chart;application/vnd.oasis.opendocument.formula;application/vnd.oasis.opendocument.graphics;application/vnd.oasis.opendocument.image;application/vnd.oasis.opendocument.presentation;application/vnd.oasis.opendocument.spreadsheet;application/vnd.oasis.opendocument.text;application/vnd.openxmlformats-officedocument.presentationml.presentation;application/vnd.openxmlformats-officedocument.spreadsheetml.sheet;application/vnd.openxmlformats-officedocument.wordprocessingml.document;application/x-bittorrent;application/zip;audio/flac;audio/mpeg;audio/ogg;audio/x-flac;image/jpeg;image/png;image/tiff;image/x-ms-bmp;text/plain;video/mp4;video/x-msvideo;
+Actions=cleanMetadata;
+Type=Service
+
+[Desktop Action cleanMetadata]
+Name=Clean metadata
+Name[de]=Metadaten löschen
+Name[es]=Limpiar metadatos
+Icon=/usr/share/icons/hicolor/scalable/apps/mat2.svg
+Exec=kdialog --yesno  "$( mat2 -s %F )" --title "Clean Metadata?" && mat2 %U
+Exec[de]=kdialog --yesno  "$( mat2 -s %F )" --title "Metadaten löschen?" && mat2 %U

libmat2/__init__.py

@@ -2,15 +2,10 @@
 
 import enum
 import importlib
-from typing import Dict, Optional, Union
+from typing import Dict
 
 from . import exiftool, video
 
-# make pyflakes happy
-assert Dict
-assert Optional
-assert Union
-
 # A set of extension that aren't supported, despite matching a supported mimetype
 UNSUPPORTED_EXTENSIONS = {
     '.asc',
@@ -67,8 +62,9 @@ CMD_DEPENDENCIES = {
     },
 }
 
+
 def check_dependencies() -> Dict[str, Dict[str, bool]]:
-    ret = dict()  # type: Dict[str, dict]
+    ret: Dict[str, Dict] = dict()
 
     for key, value in DEPENDENCIES.items():
         ret[key] = {

libmat2/abstract.py

 import abc
 import os
 import re
-from typing import Set, Dict, Union
-
-assert Set  # make pyflakes happy
+from typing import Union, Set, Dict
 
 
 class AbstractParser(abc.ABC):
@@ -11,8 +9,8 @@ class AbstractParser(abc.ABC):
     It might yield `ValueError` on instantiation on invalid files,
     and `RuntimeError` when something went wrong in `remove_all`.
     """
-    meta_list = set()  # type: Set[str]
-    mimetypes = set()  # type: Set[str]
+    meta_list: Set[str] = set()
+    mimetypes: Set[str] = set()
 
     def __init__(self, filename: str) -> None:
         """
@@ -32,10 +30,14 @@ class AbstractParser(abc.ABC):
 
         self.output_filename = fname + '.cleaned' + extension
         self.lightweight_cleaning = False
+        self.sandbox = True
 
     @abc.abstractmethod
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
-        """Return all the metadata of the current file"""
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
+        """Return all the metadata of the current file
+
+        :raises RuntimeError: Raised if the cleaning process went wrong.
+        """
 
     @abc.abstractmethod
     def remove_all(self) -> bool:

libmat2/archive.py

@@ -7,14 +7,10 @@ import tempfile
 import os
 import logging
 import shutil
-from typing import Dict, Set, Pattern, Union, Any, List
+from typing import Pattern, Union, Any, Set, Dict, List
 
 from . import abstract, UnknownMemberPolicy, parser_factory
 
-# Make pyflakes happy
-assert Set
-assert Pattern
-
 # pylint: disable=not-callable,assignment-from-no-return,too-many-branches
 
 # An ArchiveClass is a class representing an archive,
@@ -48,22 +44,24 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
     def __init__(self, filename):
         super().__init__(filename)
         # We ignore typing here because mypy is too stupid
-        self.archive_class = None  #  type: ignore
-        self.member_class = None  #  type: ignore
+        self.archive_class = None  # type: ignore
+        self.member_class = None  # type: ignore
 
         # Those are the files that have a format that _isn't_
-        # supported by MAT2, but that we want to keep anyway.
-        self.files_to_keep = set()  # type: Set[Pattern]
+        # supported by mat2, but that we want to keep anyway.
+        self.files_to_keep: Set[Pattern] = set()
 
         # Those are the files that we _do not_ want to keep,
         # no matter if they are supported or not.
-        self.files_to_omit = set()  # type: Set[Pattern]
+        self.files_to_omit: Set[Pattern] = set()
 
         # what should the parser do if it encounters an unknown file in
         # the archive?
-        self.unknown_member_policy = UnknownMemberPolicy.ABORT  # type: UnknownMemberPolicy
+        self.unknown_member_policy: UnknownMemberPolicy = UnknownMemberPolicy.ABORT
 
-        self.is_archive_valid()
+        # The LGTM comment is to mask a false-positive,
+        # see https://lgtm.com/projects/g/jvoisin/mat2/
+        self.is_archive_valid()  # lgtm [py/init-calls-subclass]
 
     def is_archive_valid(self):
         """Raise a ValueError is the current archive isn't a valid one."""
@@ -71,35 +69,47 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
     def _specific_cleanup(self, full_path: str) -> bool:
         """ This method can be used to apply specific treatment
         to files present in the archive."""
-        # pylint: disable=unused-argument,no-self-use
+        # pylint: disable=unused-argument
         return True  # pragma: no cover
 
     def _specific_get_meta(self, full_path: str, file_path: str) -> Dict[str, Any]:
         """ This method can be used to extract specific metadata
         from files present in the archive."""
-        # pylint: disable=unused-argument,no-self-use
+        # pylint: disable=unused-argument
         return {}  # pragma: no cover
 
+    def _final_checks(self) -> bool:
+        """ This method is invoked after the file has been cleaned,
+        allowing to run final verifications.
+        """
+        # pylint: disable=unused-argument
+        return True
+
     @staticmethod
-    @abc.abstractstaticmethod
+    @abc.abstractmethod
     def _get_all_members(archive: ArchiveClass) -> List[ArchiveMember]:
         """Return all the members of the archive."""
 
     @staticmethod
-    @abc.abstractstaticmethod
+    @abc.abstractmethod
     def _clean_member(member: ArchiveMember) -> ArchiveMember:
         """Remove all the metadata for a given member."""
 
     @staticmethod
-    @abc.abstractstaticmethod
+    @abc.abstractmethod
     def _get_member_meta(member: ArchiveMember) -> Dict[str, str]:
         """Return all the metadata of a given member."""
 
     @staticmethod
-    @abc.abstractstaticmethod
+    @abc.abstractmethod
     def _get_member_name(member: ArchiveMember) -> str:
         """Return the name of the given member."""
 
+    @staticmethod
+    @abc.abstractmethod
+    def _is_dir(member: ArchiveMember) -> bool:
+        """Return true is the given member is a directory."""
+
     @abc.abstractmethod
     def _add_file_to_archive(self, archive: ArchiveClass, member: ArchiveMember,
                              full_path: str):
@@ -111,8 +121,20 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
         # pylint: disable=unused-argument
         return member
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
-        meta = dict()  # type: Dict[str, Union[str, dict]]
+    @staticmethod
+    def _get_member_compression(member: ArchiveMember):
+        """Get the compression of the archive member."""
+        # pylint: disable=unused-argument
+        return None
+
+    @staticmethod
+    def _set_member_compression(member: ArchiveMember, compression) -> ArchiveMember:
+        """Set the compression of the archive member."""
+        # pylint: disable=unused-argument
+        return member
+
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
+        meta: Dict[str, Union[str, Dict]] = dict()
 
         with self.archive_class(self.filename) as zin:
             temp_folder = tempfile.mkdtemp()
@@ -121,12 +143,20 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
                 local_meta = self._get_member_meta(item)
                 member_name = self._get_member_name(item)
 
-                if member_name[-1] == '/':  # pragma: no cover
-                    # `is_dir` is added in Python3.6
+                if self._is_dir(item):  # pragma: no cover
                     continue  # don't keep empty folders
 
-                zin.extract(member=item, path=temp_folder)
                 full_path = os.path.join(temp_folder, member_name)
+                if not os.path.abspath(full_path).startswith(temp_folder):
+                    logging.error("%s contains a file (%s) pointing outside (%s) of its root.",
+                        self.filename, member_name, full_path)
+                    break
+
+                try:
+                    zin.extract(member=item, path=temp_folder)
+                except OSError as e:
+                    logging.error("Unable to extraxt %s from %s: %s", item, self.filename, e)
+
                 os.chmod(full_path, stat.S_IRUSR)
 
                 specific_meta = self._specific_get_meta(full_path, member_name)
@@ -134,6 +164,7 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
 
                 member_parser, _ = parser_factory.get_parser(full_path)  # type: ignore
                 if member_parser:
+                    member_parser.sandbox = self.sandbox
                     local_meta = {**local_meta, **member_parser.get_meta()}
 
                 if local_meta:
@@ -153,12 +184,12 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
 
             # Sort the items to process, to reduce fingerprinting,
             # and keep them in the `items` variable.
-            items = list()  # type: List[ArchiveMember]
+            items: List[ArchiveMember] = list()
             for item in sorted(self._get_all_members(zin), key=self._get_member_name):
                 # Some fileformats do require to have the `mimetype` file
                 # as the first file in the archive.
                 if self._get_member_name(item) == 'mimetype':
-                    items = [item] + items
+                    items.insert(0, item)
                 else:
                     items.append(item)
 
@@ -166,18 +197,36 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
             # we're iterating (and thus inserting) them in lexicographic order.
             for item in items:
                 member_name = self._get_member_name(item)
-                if member_name[-1] == '/':  # `is_dir` is added in Python3.6
+                if self._is_dir(item):
                     continue  # don't keep empty folders
 
-                zin.extract(member=item, path=temp_folder)
                 full_path = os.path.join(temp_folder, member_name)
+                if not os.path.abspath(full_path).startswith(temp_folder):
+                    logging.error("%s contains a file (%s) pointing outside (%s) of its root.",
+                            self.filename, member_name, full_path)
+                    abort = True
+                    break
+
+                zin.extract(member=item, path=temp_folder)
+
+                try:
+                    original_permissions = os.stat(full_path).st_mode
+                except FileNotFoundError:
+                    logging.error("Something went wrong during processing of "
+                            "%s in %s, likely a path traversal attack.",
+                            member_name, self.filename)
+                    abort = True
+                    # we're breaking instead of continuing, because this exception
+                    # is raised in case of weird path-traversal-like atttacks.
+                    break
 
-                original_permissions = os.stat(full_path).st_mode
                 os.chmod(full_path, original_permissions | stat.S_IWUSR | stat.S_IRUSR)
 
+                original_compression = self._get_member_compression(item)
+
                 if self._specific_cleanup(full_path) is False:
-                    logging.warning("Something went wrong during deep cleaning of %s",
-                                    member_name)
+                    logging.warning("Something went wrong during deep cleaning of %s in %s",
+                                    member_name, self.filename)
                     abort = True
                     continue
 
@@ -197,12 +246,13 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
                             logging.warning("In file %s, keeping unknown element %s (format: %s)",
                                             self.filename, member_name, mtype)
                         else:
-                            logging.error("In file %s, element %s's format (%s) " \
+                            logging.error("In file %s, element %s's format (%s) "
                                           "isn't supported",
                                           self.filename, member_name, mtype)
                             abort = True
                             continue
                     else:
+                        member_parser.sandbox = self.sandbox
                         if member_parser.remove_all() is False:
                             logging.warning("In file %s, something went wrong \
                                              with the cleaning of %s \
@@ -214,6 +264,7 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
 
                 zinfo = self.member_class(member_name)  # type: ignore
                 zinfo = self._set_member_permissions(zinfo, original_permissions)
+                zinfo = self._set_member_compression(zinfo, original_compression)
                 clean_zinfo = self._clean_member(zinfo)
                 self._add_file_to_archive(zout, clean_zinfo, full_path)
 
@@ -221,11 +272,14 @@ class ArchiveBasedAbstractParser(abstract.AbstractParser):
         if abort:
             os.remove(self.output_filename)
             return False
+        if not self._final_checks():
+            return False  # pragma: no cover
         return True
 
 
 class TarParser(ArchiveBasedAbstractParser):
     mimetypes = {'application/x-tar'}
+
     def __init__(self, filename):
         super().__init__(filename)
         # yes, it's tarfile.open and not tarfile.TarFile,
@@ -335,6 +389,11 @@ class TarParser(ArchiveBasedAbstractParser):
         member.mode = permissions
         return member
 
+    @staticmethod
+    def _is_dir(member: ArchiveMember) -> bool:
+        assert isinstance(member, tarfile.TarInfo)  # please mypy
+        return member.isdir()
+
 
 class TarGzParser(TarParser):
     compression = ':gz'
@@ -353,16 +412,17 @@ class TarXzParser(TarParser):
 
 class ZipParser(ArchiveBasedAbstractParser):
     mimetypes = {'application/zip'}
-    def __init__(self, filename):
+
+    def __init__(self, filename: str):
         super().__init__(filename)
         self.archive_class = zipfile.ZipFile
         self.member_class = zipfile.ZipInfo
-        self.zip_compression_type = zipfile.ZIP_DEFLATED
 
     def is_archive_valid(self):
         try:
-            zipfile.ZipFile(self.filename)
-        except zipfile.BadZipFile:
+            with zipfile.ZipFile(self.filename):
+                pass
+        except (zipfile.BadZipFile, OSError):
             raise ValueError
 
     @staticmethod
@@ -398,7 +458,7 @@ class ZipParser(ArchiveBasedAbstractParser):
         assert isinstance(member, zipfile.ZipInfo)  # please mypy
         with open(full_path, 'rb') as f:
             archive.writestr(member, f.read(),
-                             compress_type=self.zip_compression_type)
+                             compress_type=member.compress_type)
 
     @staticmethod
     def _get_all_members(archive: ArchiveClass) -> List[ArchiveMember]:
@@ -409,3 +469,19 @@ class ZipParser(ArchiveBasedAbstractParser):
     def _get_member_name(member: ArchiveMember) -> str:
         assert isinstance(member, zipfile.ZipInfo)  # please mypy
         return member.filename
+
+    @staticmethod
+    def _get_member_compression(member: ArchiveMember):
+        assert isinstance(member, zipfile.ZipInfo)  # please mypy
+        return member.compress_type
+
+    @staticmethod
+    def _set_member_compression(member: ArchiveMember, compression) -> ArchiveMember:
+        assert isinstance(member, zipfile.ZipInfo)  # please mypy
+        member.compress_type = compression
+        return member
+
+    @staticmethod
+    def _is_dir(member: ArchiveMember) -> bool:
+        assert isinstance(member, zipfile.ZipInfo)  # please mypy
+        return member.is_dir()

libmat2/audio.py

@@ -2,42 +2,51 @@ import mimetypes
 import os
 import shutil
 import tempfile
-from typing import Dict, Union
+from typing import Union, Dict
 
 import mutagen
 
-from . import abstract, parser_factory
+from . import abstract, parser_factory, video
 
 
 class MutagenParser(abstract.AbstractParser):
     def __init__(self, filename):
         super().__init__(filename)
         try:
-            mutagen.File(self.filename)
+            if mutagen.File(self.filename) is None:
+                raise ValueError
         except mutagen.MutagenError:
             raise ValueError
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
         f = mutagen.File(self.filename)
         if f.tags:
-            return {k:', '.join(v) for k, v in f.tags.items()}
+            return {k: ', '.join(map(str, v)) for k, v in f.tags.items()}
         return {}
 
     def remove_all(self) -> bool:
         shutil.copy(self.filename, self.output_filename)
         f = mutagen.File(self.output_filename)
-        f.delete()
-        f.save()
+        try:
+            f.delete()
+            f.save()
+        except mutagen.MutagenError:
+            raise ValueError
         return True
 
 
 class MP3Parser(MutagenParser):
     mimetypes = {'audio/mpeg', }
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
-        metadata = {}  # type: Dict[str, Union[str, dict]]
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
+        metadata: Dict[str, Union[str, Dict]] = dict()
         meta = mutagen.File(self.filename).tags
+        if not meta:
+            return metadata
         for key in meta:
+            if isinstance(key, tuple):
+                metadata[key[0]] = key[1]
+                continue
             if not hasattr(meta[key], 'text'):  # pragma: no cover
                 continue
             metadata[key.rstrip(' \t\r\n\0')] = ', '.join(map(str, meta[key].text))
@@ -59,12 +68,12 @@ class FLACParser(MutagenParser):
         f.save(deleteid3=True)
         return True
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
         meta = super().get_meta()
         for num, picture in enumerate(mutagen.File(self.filename).pictures):
             name = picture.desc if picture.desc else 'Cover %d' % num
             extension = mimetypes.guess_extension(picture.mime)
-            if extension is None:  #  pragma: no cover
+            if extension is None: #  pragma: no cover
                 meta[name] = 'harmful data'
                 continue
 
@@ -73,7 +82,33 @@ class FLACParser(MutagenParser):
             with open(fname, 'wb') as f:
                 f.write(picture.data)
             p, _ = parser_factory.get_parser(fname)  # type: ignore
+            if p is None:
+                raise ValueError
+            p.sandbox = self.sandbox
             # Mypy chokes on ternaries :/
             meta[name] = p.get_meta() if p else 'harmful data'  # type: ignore
             os.remove(fname)
         return meta
+
+
+class WAVParser(video.AbstractFFmpegParser):
+    mimetypes = {'audio/x-wav', }
+    meta_allowlist = {'AvgBytesPerSec', 'BitsPerSample', 'Directory',
+                      'Duration', 'Encoding', 'ExifToolVersion',
+                      'FileAccessDate', 'FileInodeChangeDate',
+                      'FileModifyDate', 'FileName', 'FilePermissions',
+                      'FileSize', 'FileType', 'FileTypeExtension',
+                      'MIMEType', 'NumChannels', 'SampleRate', 'SourceFile',
+                     }
+
+
+class AIFFParser(video.AbstractFFmpegParser):
+    mimetypes = {'audio/aiff', 'audio/x-aiff'}
+    meta_allowlist = {'AvgBytesPerSec', 'BitsPerSample', 'Directory',
+                      'Duration', 'Encoding', 'ExifToolVersion',
+                      'FileAccessDate', 'FileInodeChangeDate',
+                      'FileModifyDate', 'FileName', 'FilePermissions',
+                      'FileSize', 'FileType', 'FileTypeExtension',
+                      'MIMEType', 'NumChannels', 'SampleRate', 'SourceFile',
+                      'NumSampleFrames', 'SampleSize',
+                     }

libmat2/subprocess.py → libmat2/bubblewrap.py

@@ -11,24 +11,26 @@ import os
 import shutil
 import subprocess
 import tempfile
-from typing import List, Optional
+import functools
+from typing import Optional, List
 
 
 __all__ = ['PIPE', 'run', 'CalledProcessError']
 PIPE = subprocess.PIPE
 CalledProcessError = subprocess.CalledProcessError
 
+# pylint: disable=subprocess-run-check
 
+
+@functools.lru_cache(maxsize=None)
 def _get_bwrap_path() -> str:
-    bwrap_path = '/usr/bin/bwrap'
-    if os.path.isfile(bwrap_path):
-        if os.access(bwrap_path, os.X_OK):
-            return bwrap_path
+    which_path = shutil.which('bwrap')
+    if which_path:
+        return which_path
 
     raise RuntimeError("Unable to find bwrap")  # pragma: no cover
 
 
-# pylint: disable=bad-whitespace
 def _get_bwrap_args(tempdir: str,
                     input_filename: str,
                     output_filename: Optional[str] = None) -> List[str]:
@@ -37,7 +39,7 @@ def _get_bwrap_args(tempdir: str,
 
     # XXX: use --ro-bind-try once all supported platforms
     # have a bubblewrap recent enough to support it.
-    ro_bind_dirs = ['/usr', '/lib', '/lib64', '/bin', '/sbin', cwd]
+    ro_bind_dirs = ['/usr', '/lib', '/lib64', '/bin', '/sbin', '/etc/alternatives', cwd]
     for bind_dir in ro_bind_dirs:
         if os.path.isdir(bind_dir):  # pragma: no cover
             ro_bind_args.extend(['--ro-bind', bind_dir, bind_dir])
@@ -49,9 +51,16 @@ def _get_bwrap_args(tempdir: str,
 
     args = ro_bind_args + \
         ['--dev', '/dev',
+         '--proc', '/proc',
          '--chdir', cwd,
-         '--unshare-all',
+         '--unshare-user-try',
+         '--unshare-ipc',
+         '--unshare-pid',
+         '--unshare-net',
+         '--unshare-uts',
+         '--unshare-cgroup-try',
          '--new-session',
+         '--cap-drop', 'all',
          # XXX: enable --die-with-parent once all supported platforms have
          # a bubblewrap recent enough to support it.
          # '--die-with-parent',
@@ -69,7 +78,6 @@ def _get_bwrap_args(tempdir: str,
     return args
 
 
-# pylint: disable=bad-whitespace
 def run(args: List[str],
         input_filename: str,
         output_filename: Optional[str] = None,

libmat2/epub.py

 import logging
 import re
 import uuid
+import zipfile
 import xml.etree.ElementTree as ET  # type: ignore
+from typing import Any, Dict
 
 from . import archive, office
 
+
 class EPUBParser(archive.ZipParser):
     mimetypes = {'application/epub+zip', }
     metadata_namespace = '{http://purl.org/dc/elements/1.1/}'
@@ -15,11 +18,27 @@ class EPUBParser(archive.ZipParser):
             'META-INF/container.xml',
             'mimetype',
             'OEBPS/content.opf',
+            'content.opf',
+            'hmh.opf',
+            'OPS/.+.xml'
             }))
+        self.files_to_omit = set(map(re.compile, {  # type: ignore
+            'iTunesMetadata.plist',
+            'META-INF/calibre_bookmarks.txt',
+            'OEBPS/package.opf',
+             }))
         self.uniqid = uuid.uuid4()
 
-    def _specific_get_meta(self, full_path, file_path):
-        if file_path != 'OEBPS/content.opf':
+    def is_archive_valid(self):
+        super().is_archive_valid()
+        with zipfile.ZipFile(self.filename) as zin:
+            for item in self._get_all_members(zin):
+                member_name = self._get_member_name(item)
+                if member_name.endswith('META-INF/encryption.xml'):
+                    raise ValueError('the file contains encrypted fonts')
+
+    def _specific_get_meta(self, full_path, file_path) -> Dict[str, Any]:
+        if not file_path.endswith('.opf'):
             return {}
 
         with open(full_path, encoding='utf-8') as f:
@@ -30,14 +49,31 @@ class EPUBParser(archive.ZipParser):
             except (TypeError, UnicodeDecodeError):
                 return {file_path: 'harmful content', }
 
-    def _specific_cleanup(self, full_path: str):
-        if full_path.endswith('OEBPS/content.opf'):
+    def _specific_cleanup(self, full_path: str) -> bool:
+        if full_path.endswith('hmh.opf') or full_path.endswith('content.opf'):
             return self.__handle_contentopf(full_path)
         elif full_path.endswith('OEBPS/toc.ncx'):
             return self.__handle_tocncx(full_path)
+        elif re.search('/OPS/[^/]+.xml$', full_path):
+            return self.__handle_ops_xml(full_path)
+        return True
+
+    def __handle_ops_xml(self, full_path: str) -> bool:
+        try:
+            tree, namespace = office._parse_xml(full_path)
+        except ET.ParseError:  # pragma: nocover
+            logging.error("Unable to parse %s in %s.", full_path, self.filename)
+            return False
+
+        for item in tree.iterfind('.//', namespace):  # pragma: nocover
+            if item.tag.strip().lower().endswith('head'):
+                item.clear()
+                break
+        tree.write(full_path, xml_declaration=True, encoding='utf-8',
+                   short_empty_elements=False)
         return True
 
-    def __handle_tocncx(self, full_path: str):
+    def __handle_tocncx(self, full_path: str) -> bool:
         try:
             tree, namespace = office._parse_xml(full_path)
         except ET.ParseError:  # pragma: nocover
@@ -53,7 +89,7 @@ class EPUBParser(archive.ZipParser):
                    short_empty_elements=False)
         return True
 
-    def __handle_contentopf(self, full_path: str):
+    def __handle_contentopf(self, full_path: str) -> bool:
         try:
             tree, namespace = office._parse_xml(full_path)
         except ET.ParseError:
@@ -71,7 +107,7 @@ class EPUBParser(archive.ZipParser):
                 item.append(uniqid)
 
                 # items without mandatory content
-                for name in {'language', 'title'}:
+                for name in ['language', 'title']:
                     uniqid = ET.Element(self.metadata_namespace + name)
                     item.append(uniqid)
                 break  # there is only a single <metadata> block

libmat2/exiftool.py

@@ -2,13 +2,12 @@ import functools
 import json
 import logging
 import os
-from typing import Dict, Union, Set
+import shutil
+import subprocess
+from typing import Union, Set, Dict
 
 from . import abstract
-from . import subprocess
-
-# Make pyflakes happy
-assert Set
+from . import bubblewrap
 
 
 class ExiftoolParser(abstract.AbstractParser):
@@ -16,12 +15,21 @@ class ExiftoolParser(abstract.AbstractParser):
     from a import file, hence why several parsers are re-using its `get_meta`
     method.
     """
-    meta_allowlist = set()  # type: Set[str]
+    meta_allowlist: Set[str] = set()
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
-        out = subprocess.run([_get_exiftool_path(), '-json', self.filename],
-                             input_filename=self.filename,
-                             check=True, stdout=subprocess.PIPE).stdout
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
+        try:
+            if self.sandbox:
+                out = bubblewrap.run([_get_exiftool_path(), '-json',
+                                      self.filename],
+                                     input_filename=self.filename,
+                                     check=True, stdout=subprocess.PIPE).stdout
+            else:
+                out = subprocess.run([_get_exiftool_path(), '-json',
+                                      self.filename],
+                                     check=True, stdout=subprocess.PIPE).stdout
+        except subprocess.CalledProcessError:  # pragma: no cover
+            raise ValueError
         meta = json.loads(out.decode('utf-8'))[0]
         for key in self.meta_allowlist:
             meta.pop(key, None)
@@ -29,8 +37,7 @@ class ExiftoolParser(abstract.AbstractParser):
 
     def _lightweight_cleanup(self) -> bool:
         if os.path.exists(self.output_filename):
-            try:
-                # exiftool can't force output to existing files
+            try:  # exiftool can't force output to existing files
                 os.remove(self.output_filename)
             except OSError as e:  # pragma: no cover
                 logging.error("The output file %s is already existing and \
@@ -49,24 +56,25 @@ class ExiftoolParser(abstract.AbstractParser):
                '-o', self.output_filename,
                self.filename]
         try:
-            subprocess.run(cmd, check=True,
-                           input_filename=self.filename,
-                           output_filename=self.output_filename)
+            if self.sandbox:
+                bubblewrap.run(cmd, check=True,
+                               input_filename=self.filename,
+                               output_filename=self.output_filename)
+            else:
+                subprocess.run(cmd, check=True)
         except subprocess.CalledProcessError as e:  # pragma: no cover
             logging.error("Something went wrong during the processing of %s: %s", self.filename, e)
             return False
         return True
 
-@functools.lru_cache()
+@functools.lru_cache(maxsize=None)
 def _get_exiftool_path() -> str:  # pragma: no cover
-    possible_pathes = {
-        '/usr/bin/exiftool',              # debian/fedora
-        '/usr/bin/vendor_perl/exiftool',  # archlinux
-    }
+    which_path = shutil.which('exiftool')
+    if which_path:
+        return which_path
 
-    for possible_path in possible_pathes:
-        if os.path.isfile(possible_path):
-            if os.access(possible_path, os.X_OK):
-                return possible_path
+    # Exiftool on Arch Linux has a weird path
+    if os.access('/usr/bin/vendor_perl/exiftool', os.X_OK):
+        return '/usr/bin/vendor_perl/exiftool'
 
     raise RuntimeError("Unable to find exiftool")

libmat2/harmless.py

 import shutil
-from typing import Dict, Union
+from typing import Union, Dict
 from . import abstract
 
 
 class HarmlessParser(abstract.AbstractParser):
     """ This is the parser for filetypes that can not contain metadata. """
-    mimetypes = {'text/plain', 'image/x-ms-bmp'}
+    mimetypes = {'text/plain', 'image/x-ms-bmp', 'image/bmp'}
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
         return dict()
 
     def remove_all(self) -> bool:

libmat2/images.py

-import imghdr
 import os
-from typing import Set, Dict, Union
+import re
+from typing import Union, Any, Dict
 
 import cairo
 
@@ -9,10 +9,7 @@ gi.require_version('GdkPixbuf', '2.0')
 gi.require_version('Rsvg', '2.0')
 from gi.repository import GdkPixbuf, GLib, Rsvg
 
-from . import exiftool
-
-# Make pyflakes happy
-assert Set
+from . import exiftool, abstract
 
 class SVGParser(exiftool.ExiftoolParser):
     mimetypes = {'image/svg+xml', }
@@ -24,25 +21,40 @@ class SVGParser(exiftool.ExiftoolParser):
                       }
 
     def remove_all(self) -> bool:
-        svg = Rsvg.Handle.new_from_file(self.filename)
-        dimensions = svg.get_dimensions()
-        surface = cairo.SVGSurface(self.output_filename,
-                                   dimensions.height,
-                                   dimensions.width)
+        try:
+            svg = Rsvg.Handle.new_from_file(self.filename)
+        except GLib.GError:
+            raise ValueError
+
+        try:
+            _, _, _, _, has_viewbox, viewbox = svg.get_intrinsic_dimensions()
+            if has_viewbox is False:
+                raise ValueError
+            _, width, height = svg.get_intrinsic_size_in_pixels()
+        except AttributeError:
+            dimensions = svg.get_dimensions()
+            height, width = dimensions.height, dimensions.width
+
+        surface = cairo.SVGSurface(self.output_filename, height, width)
         context = cairo.Context(surface)
-        svg.render_cairo(context)
+        try:
+            svg.render_document(context, viewbox)
+        except AttributeError:
+            svg.render_cairo(context)
+
         surface.finish()
         return True
 
-    def get_meta(self) -> Dict[str, Union[str, dict]]:
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
         meta = super().get_meta()
 
         # The namespace is mandatory, but only the …/2000/svg is valid.
         ns = 'http://www.w3.org/2000/svg'
-        if meta.get('Xmlns', ns) == ns:
+        if meta.get('Xmlns') == ns:
             meta.pop('Xmlns')
         return meta
 
+
 class PNGParser(exiftool.ExiftoolParser):
     mimetypes = {'image/png', }
     meta_allowlist = {'SourceFile', 'ExifToolVersion', 'FileName',
@@ -56,12 +68,10 @@ class PNGParser(exiftool.ExiftoolParser):
     def __init__(self, filename):
         super().__init__(filename)
 
-        if imghdr.what(filename) != 'png':
-            raise ValueError
-
         try:  # better fail here than later
             cairo.ImageSurface.create_from_png(self.filename)
-        except MemoryError:  # pragma: no cover
+        except:  # pragma: no cover
+            # Cairo is returning some weird exceptions :/
             raise ValueError
 
     def remove_all(self) -> bool:
@@ -95,7 +105,6 @@ class GdkPixbufAbstractParser(exiftool.ExiftoolParser):
 
     def __init__(self, filename):
         super().__init__(filename)
-        # we can't use imghdr here because of https://bugs.python.org/issue28591
         try:
             GdkPixbuf.Pixbuf.new_from_file(self.filename)
         except GLib.GError:
@@ -107,9 +116,16 @@ class GdkPixbufAbstractParser(exiftool.ExiftoolParser):
 
         _, extension = os.path.splitext(self.filename)
         pixbuf = GdkPixbuf.Pixbuf.new_from_file(self.filename)
+        pixbuf = GdkPixbuf.Pixbuf.apply_embedded_orientation(pixbuf)
         if extension.lower() == '.jpg':
             extension = '.jpeg'  # gdk is picky
-        pixbuf.savev(self.output_filename, type=extension[1:], option_keys=[], option_values=[])
+        elif extension.lower() == '.tif':
+            extension = '.tiff'  # gdk is picky
+        try:
+            pixbuf.savev(self.output_filename, type=extension[1:],
+                         option_keys=[], option_values=[])
+        except GLib.GError:  # pragma: no cover
+            return False
         return True
 
 
@@ -123,7 +139,7 @@ class JPGParser(GdkPixbufAbstractParser):
                       'MIMEType', 'ImageWidth', 'ImageSize', 'BitsPerSample',
                       'ColorComponents', 'EncodingProcess', 'JFIFVersion',
                       'ResolutionUnit', 'XResolution', 'YCbCrSubSampling',
-                      'YResolution', 'Megapixels', 'ImageHeight'}
+                      'YResolution', 'Megapixels', 'ImageHeight', 'Orientation'}
 
 
 class TiffParser(GdkPixbufAbstractParser):
@@ -137,4 +153,58 @@ class TiffParser(GdkPixbufAbstractParser):
                       'FileInodeChangeDate', 'FileModifyDate', 'FileName',
                       'FilePermissions', 'FileSize', 'FileType',
                       'FileTypeExtension', 'ImageHeight', 'ImageSize',
-                      'ImageWidth', 'MIMEType', 'Megapixels', 'SourceFile'}
+                      'ImageWidth', 'MIMEType', 'Megapixels', 'SourceFile', 'Orientation'}
+
+
+class PPMParser(abstract.AbstractParser):
+    mimetypes = {'image/x-portable-pixmap'}
+
+    def get_meta(self) -> Dict[str, Union[str, Dict]]:
+        meta: Dict[str, Union[str, Dict[Any, Any]]] = dict()
+        with open(self.filename) as f:
+            for idx, line in enumerate(f):
+                if line.lstrip().startswith('#'):
+                    meta[str(idx)] = line.lstrip().rstrip()
+        return meta
+
+    def remove_all(self) -> bool:
+        with open(self.filename) as fin:
+            with open(self.output_filename, 'w') as fout:
+                for line in fin:
+                    if not line.lstrip().startswith('#'):
+                        line = re.sub(r"\s+", "", line, flags=re.UNICODE)
+                        fout.write(line)
+        return True
+
+
+class HEICParser(exiftool.ExiftoolParser):
+    mimetypes = {'image/heic'}
+    meta_allowlist = {'SourceFile', 'ExifToolVersion', 'FileName', 'Directory',
+            'FileSize', 'FileModifyDate', 'FileAccessDate',
+            'FileInodeChangeDate', 'FilePermissions', 'FileType',
+            'FileTypeExtension', 'MIMEType', 'MajorBrand', 'MinorVersion',
+            'CompatibleBrands','HandlerType', 'PrimaryItemReference',
+            'HEVCConfigurationVersion', 'GeneralProfileSpace',
+            'GeneralTierFlag', 'GeneralProfileIDC',
+            'GenProfileCompatibilityFlags', 'ConstraintIndicatorFlags',
+            'GeneralLevelIDC', 'MinSpatialSegmentationIDC',
+            'ParallelismType','ChromaFormat', 'BitDepthLuma', 'BitDepthChroma',
+            'NumTemporalLayers', 'TemporalIDNested', 'ImageWidth',
+            'ImageHeight', 'ImageSpatialExtent', 'ImagePixelDepth',
+            'AverageFrameRate', 'ConstantFrameRate', 'MediaDataSize',
+            'MediaDataOffset','ImageSize', 'Megapixels'}
+
+    def remove_all(self) -> bool:
+        return self._lightweight_cleanup()
+
+class WEBPParser(GdkPixbufAbstractParser):
+    mimetypes = {'image/webp'}
+    meta_allowlist = {'SourceFile', 'ExifToolVersion', 'FileName',
+                      'Directory', 'FileSize', 'FileModifyDate',
+                      'FileAccessDate', "FileInodeChangeDate",
+                      'FilePermissions', 'FileType', 'FileTypeExtension',
+                      'MIMEType', 'ImageWidth', 'ImageSize', 'BitsPerSample',
+                      'ColorComponents', 'EncodingProcess', 'JFIFVersion',
+                      'ResolutionUnit', 'XResolution', 'YCbCrSubSampling',
+                      'YResolution', 'Megapixels', 'ImageHeight', 'Orientation',
+                      'HorizontalScale', 'VerticalScale', 'VP8Version'}