Feature request: Binary string diffing support

[gschlager]

Problem

When comparing binary strings (ASCII-8BIT encoding), super_diff shows unreadable output with escape sequences and control characters. This is common when testing file formats, network protocols, or any binary data.

Proposed solution

Add built-in support for binary string diffing that:

1. Detects binary strings - Strings with Encoding::ASCII_8BIT
2. Displays as hex dump - Shows offset, hex bytes, and printable ASCII:

   00000000: 7573 7461 7220 2000 0000 0000 0000 0000  ustar  .........                                                                                                                                         
   00000010: 0000 0000 0000 0000 0000 0000 0000 0000  ................                                                                                                                                         
   

3. Shows context around changes - Only displays a few lines around actual differences (similar to unified diff)
4. Inspects cleanly - Shows <binary string (512 bytes)> instead of garbled text

Working implementation

We have a working extension in mini_tarball that we'd be happy to contribute as a PR. It includes:

• Differs::BinaryString
• InspectionTreeBuilders::BinaryString
• OperationTreeBuilders::BinaryString
• OperationTrees::BinaryString
• OperationTreeFlatteners::BinaryString

The implementation follows super_diff's architecture and has been working well for us when testing tar archive generation.

Would you be interested in a PR for this feature?

[jas14]

Hi @gschlager , thanks for the suggestion!

I like this style of output, but this would be a big change from how binary strings are currently displayed. To avoid breaking people's expectations too much, I'd accept a PR to add this as an opt-in feature.

I took a quick look at your code and see you're doing something similar to a lesser-known feature of SuperDiff called diff elision. Would you be able to incorporate this into your PR? (Though I'm not sure if it will work with multi-line strings as currently written.)