On Wed, Jul 29, 2009 at 3:22 PM, Casey McGinty<[email protected]> wrote: > If md5sum --check did not return an error with a truncated file, then the > file was likely broken to start with. I think the chances of a corrupted > file generating the same md5sum is close enough to 0 that you will never see > it in a life time.
I mis-spoke, sorry, what md5check does is what i'll call a quick check via metaflac --show-md5sum and then compares that to a previously stored ffp text file then it also calls flac -t which i'll call the extended check where i mis-spoke is it does not use md5sum --check (because that doesn't work on ffp files) > Also, you lost me on storing the md5sum value at the begging and the end? the quick check succeeds if the file has been truncated after the ffp.txt has been generated (when the file was still good) because the internal md5sum must be stored at the beginning of the file and is still retrieved ok by --show-md5sum so i was thinking that if it were stored at beg and end that would fix that but any other corruption could happen anywhere in middle and it would still be an issue - so using --test every time to make sure is the best way to go _______________________________________________ Flac mailing list [email protected] http://lists.xiph.org/mailman/listinfo/flac
