Re: [PATCH] Add a 1-second sleep to git-cvsexportcommit test

From: Marius Storm-Olsen <marius@trolltech.com>
Date: Jul 25, 2007, 07:43 UTC
Message-ID: <46A6FF23.8050704@trolltech.com>
In-Reply-To: <200707250835.21181.andyparkins@gmail.com>

Andy Parkins said the following on 25.07.2007 09:35:

Show 10 quoted lines

> On Tuesday 2007 July 24, Linus Torvalds wrote:
> 
>> So if you sleep for one second, the filesystem times will update by one
>> second, but if you try to *synchronize* to exactly one second, it's not at
>> all certain that the *filesystem* clock will be synchronized to the same
>> second! Time skew is simply a fact of life.
> 
> I think it's even worse; if memory serves one of the Windows file systems 
> (spit) only stores times to a two-second resolution.  So half the time, 
> waiting for one second won't change the time stamp _at all_.

"File time stamps on FAT drives are rounded to the nearest two seconds (even number) when the file is written to the drive. The file time stamps on NTFS drives are rounded to the nearest 100 nanoseconds when the file is written to the drive."

http://support.microsoft.com/kb/127830

-- 
.marius