Someone just pointed something out which basically comes down to if you're using these two together: PRAGMA secure_delete=ON; PRAGMA journal_mode=PERSIST; then the first makes sure that evidence of deleted data should be missing from the main database file, but the second can leak such evidence in the journal file. For now I'm switching to TRUNCATE, but it did make me wonder. Should PERSIST devolve to TRUNCATE automatically in this combination? Or maybe PERSIST could overwrite the first journal_size_limit bytes of the file? [Obviously that could be optimized to be smaller in many cases.]
This isn't so much a feature request as wondering about what is correct and the least surprising outcome. -scott
