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

[Stefan Monnier]

>> Actually no: I meant ctime.
>> mtime can be tweaked with touch, so you can't rely on it if you want to
>> be safe.  But admittedly, CVS relies on exclusively on mtime (not even the
>> size) and problems related to that have been extremely rare.
> hardlinking updates ctime.  ctime cannot be used.

Ah... good point.

>> That can happen while keeping the inode constant as well.
> not without deliberate intervention a la touch.  We're not aiming to
> prevent people shooting their own foot off.  We're trying to provide
> sights so that they know when they've pointed the gun at said foot.

Then why in the world check the device number?

>> > All of the "changes" and "file diffs" will produce faulty output if the
>> > basis for comparison is corrupt.
>> Sure.  Corruption can and does happen without changing any inode number,
>> mtime, or size.
> Oh? (Not disk random bits changing though - that disk sector checksums
> are for).

Checksums aren't always sufficient, but more seriously, OS bugs or crashes
or mishandling that end up wrongly overwriting a sector are definitely
part of the real world, tho very rare.

I still haven't seen a definition of which kinds of corruption we're trying
to defend ourselves against.  BTW many/most programs are perfectly happy
with just an mtime check.


        Stefan