traperto/magic-bytes-validator

File validator that checks by magic bytes and MIME types

magic-bytes-validator's Issues

Set Default FileTypes

Idea for v2

This is not a real "issue" but rather an idea that I had. Currently, we cannot handle special quirks (like offsets, variable bytes inside a sequence or final byte sequences) really well. This is based on our architecture, which handles every FileType the same. To mitigate those problems, we could refactor our matching logic:

An IFileType has a bool Matches(byte[] bytes) method, which returns whether a given (complete) file stream matches a file type. An abstract AbstractStartSequencesFileType : IFileType could implement the Matches method to validate its magic byte sequences (given by its inheritors) just like we match the file types today.
File types with special magic bytes logic could be implemented clean and encapsulated in the file type only. A drawback would be, that we need to load the complete file stream before matching. But this is necessary to be completely sure that the file type matches.

This would obviously be a major breaking change. Therefore we would be forced to make it a v2. Depending on how many other people depend on this package, we should support the v1 alongside v2, at least for some time.

What do you think about this?

Create Default FileType Classes

e.g. PNG

Magic Bytes offset on Streams

Thanks for this library, I've found it very useful in a project I'm working on. However I've found that when evaluating magic bytes with the StreamFileTypeProvider, the offset is never considered, meaning that the mp4 format will never be discovered (or others that might be supported in the future that have offset bytes)

I'm sending a pull request with this fix and an optimization when ordering the sizes to find the maximum bytes size for the buffer,

Thanks for your hard work!

Missing LICENSE file in Repository

You are missing a LICENSE file in your repository. Please add it using the guide in the following link. https://docs.github.com/en/communities/setting-up-your-project-for-healthy-contributions/adding-a-license-to-a-repository
It is not enough to simply mention that the license is an MIT license, otherwise everyone using this repo is legally liable and this could have major monetary consequences for your users.

Add functionality to detect mimetypes from a data-stream

Sometimes you don't have a extension nor a mimetype and the only way is to detect is by inspecting the data and look for magic bytes.

Since the library also knows the magic bytes it would be great if it also offers a detection mechanism.