Forcing check and fix commands to brute-force -i folder

[watthefak]

Hello,

This is a really niche edge-case which I admit is caused mostly by myself. However, I am reaching out in an attempt to maybe get help and answers.
My current goal is to have snapraid check -i /mnt/rec sift through the whole folder block-by-block in search for any matches against the missing blocks.

The setup I have is 1 parity P with 3 drives A B C all ext4. All data was erasure encoded duplicacy chunks. Drive B got essentially rm -rf through a series of unfortunate accidents. Drive A and C had 1% of files deleted since last sync. Snapraid fix recovered 99% of files. Unrecoverable files are those which were both among the 1% deleted and also were in B.
I did an undelete operation on A and C. That recovered a whole mess of chunks both intact and mangled. Those are all in /mnt/rec

When running snapraid check -f unrecoverable_chunkfile -i /mnt/rec it finishes in seconds so my assumption is that snapraid does not actually read the whole /mnt/rec folder only checks if any file has exactly the same size and gives up.
I would like it to try and read the whole folder and calculate if any combination of data in /mnt/rec with data on A and C give a match to recover an extra block otherwise not recovered by snapraid.

[amadvance]

'-i' does read all files under the specified directory, but it doesn't brute-force every possible offset. It simply reads each file sequentially in SnapRAID-sized blocks and hashes those blocks. So the operation is quite efficient, especially if '/mnt/rec' is not very large or the data is already in the filesystem cache. The fact that it takes only a few seconds doesn't mean it is only checking files with a matching size.