[Gnu-arch-users] Re: tla1.2 on cygwin

[Greek0]

On 11 Mar 2004 15:22:32 -0500
Stefan Monnier <address@hidden> wrote:

> 
> And if we really want to detect corruption, how about MD5 instead of
> inodes?

Inode signatures are fast, MD5 is slow. You have to read the whole file
to determinate the md5 sum, and that's exactly what you want to prevent
using inode sigs.
I think on FAT ctime & filenames should be enough to detect corruption.
I don't know how tla should detect corruption on NFS though. Actually I
don't think there's a right way to it. Either you don't do it at all,
which is bad (since you work without a safety net), or you do it with
md5 (or something similar), which is bad (since it kills performance).

Greek0