Got any early evidence, personal evidence, or citations? I haven’t done a big research recently.
I don’t think it’s used enough to say definitively, but the 2.0.9.107 report which continues here is awaiting further isolation by @Taomyn because error is likely not fixable without an actual case which could be given to the SharpCompress project as an issue or pull request for an LZMA fix:
We got it narrowed from three dblock files to one, but even one dblock can contain sensitive data. Same is true for one block, but with reduced chance. Maybe copy the database for a future check about which files use the bad block. As more time passes, the check will become less trustworthy because source file versions will be deleted by retention and connection to the blocks will worsen.