feature request: remove some metadata from uploaded files

[akosmaroy]

it would be nice to have an automatism that removes some (maybe telling) metadata from uploaded files, or even from files within uploaded zip/tar/etc archives. I'm thinking about the following kind of metadata:

• EXIM data from JPEG and other images
• remove metadata from PDFs
• convert 'telling' formats to more raw formats, e.g. MS Office (Word) documents to RTF or other simpler documents that don't have capabilities to store metadata

this would make it less possible to trace uploaded file to the originator, in case he wasn't careful enough

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

[fpietrosanti]

Added in backlog of tickets.

This request related to two different requests:

• Metadata Removal: https://github.com/globaleaks/GlobaLeaks-0.1/issues/53
• File Sterilizer #270

The application design of file sterilizer + metadata removal should be done in a way that let create an independent application, very well sandboxed, interacting with globaleaks trough a simple REST interface.

[evilaliv3]

The recent ideas on this ticket are to:

• preserve the original
• provide to the user a sterilized copy of the file with no metadata.

this could be implemented on the backend by sending the file in plaintext (and so in regression with the upcoming browsercrypto design) or partially in the client supposing to be able to remove the metadata at runtime while making the user save the file.

[NSkelsey]

For text based content using an input box like a text area and saying copy and paste here sometimes works much better.

This is very common for online forms where you are asked to provide a CV or a resume.

If the content is what you are interested in then the formatting is only slightly important.