Not decrypting streams in encrypted pdf about pdf HOT 6 CLOSED

Hello,

the decryption happens here: https://github.com/pdf-rs/pdf/blob/master/pdf/src/file.rs#L75

from pdf.

and it is set here: https://github.com/pdf-rs/pdf/blob/master/pdf/src/file.rs#L243

from pdf.

Ok, as far as I understand the following is happening:

In _parse_with_lexer_ctx objects are found and decrypted. For examples for Strings here.

But for streams, the data is not really read, only the range withing the file where the stream is located is strored (that happens in parse_stream_object which is called here and defined here.

Then, in Storage::decode the data at the range stored for the stream is taken from the file and decoded in decode. But it is not decrypted there. Decryption at that point is also not easy, because the Context with the key is not available (as far as I can see).

Is my analysis correct?

If so, I wonder how to fix that.

• Option A: Instead of only storing the range for the stream, store the decrypted stream.
• Option B: Make the key available at the point where the stream is decoded (in FlateDecode for example).
• Option C: I am totally wrong (not unlikely).

@s3bk What do you think, how should this be done? I am willing to try to implement this but I would like your opinion so I am not running in the wrong direction.

from pdf.

From my understanding there is only one key per file, which is why decrypting happens in storage.
Why is it not decrypted?

https://github.com/pdf-rs/pdf/blob/master/pdf/src/object/stream.rs#L84 should call
https://github.com/pdf-rs/pdf/blob/master/pdf/src/file.rs#L129 which calls
https://github.com/pdf-rs/pdf/blob/master/pdf/src/file.rs#L70
and decrypts it.

from pdf.

Mmmh, should decrypt work "in place"? Maybe https://github.com/pdf-rs/pdf/blob/master/pdf/src/file.rs#L75 should be

            data = Vec::from(t!(decoder.decrypt(id, &mut data)));

from pdf.

ugh. yes.

from pdf.