Add support for archives

Now that we have an policy option since 4192a2da (thanks to @dkg), it would be nice to add support for archives in MAT2.

• Implement archive support
  • zip
  • tar.gz
  • tar.bz2
  • tar.xz
  • tar (!39 (merged))
    • Handle path traversal issues ( 95169906 )
    • Handle symlink attacks ( 95169906 )
    • All the shits mentioned here
• Add tests
  • Add valid tests (those could be created at runtime by the testsuite)
  • Add corrupted archives
  • Add valid zip containing invalid files
  • Add archives containing read-only files
• Add documentation

assigned to @smagnin

Just need to write a test for the zip, maybe tomorrow it should be good.

@jvoisin Do we want to have the metadata list of the archive only or the metadata list of the archive and all the files inside ? (When we are doing a mat2 -s foobar.zip)

Since we are cleaning the metadata of all files inside the archive, I think we should print them too.

We want to have the metadata from files inside the archive too. You might want to change their display in the CLI, to convey which ones are from the archive or the files themselves.

ok that's what I was thinking :)

Another question : How do we want to handle errors ? If we have an unsupported type files inside the archives, currently with my code it will fail. Are we just throwing an error and stop or throwing a warning, continue the cleaning and maybe name the output : foobar.partialcleaned.zip ?

There is the policy thingy, introduced by @dkg in !6 (merged) that you should use :)

changed the description

changed milestone to %0.5.0 - Slug

marked the checklist item zip as completed

marked the checklist item Implement archive support as completed

marked the checklist item Add valid tests (those could be created at runtime by the testsuite) as completed

marked the checklist item Add valid tests (those could be created at runtime by the testsuite) as incomplete

marked the checklist item Add valid zip containing invalid files as completed

marked the checklist item Add valid tests (those could be created at runtime by the testsuite) as completed

Zip support implemented in 3a070b0a

changed milestone to %0.7.0 - Roaster

About tar files, what filetypesdo we want to keep/discard? Hardlinks? Symlinks? Block-devices?

changed the description

changed milestone to %0.8.0 - Lobster

assigned to @jvoisin and unassigned @smagnin