Re: [PATCH] parse_date(): fix parsing 03/10/2006

[Andrew Morton <akpm@osdl.org>]

Junio C Hamano <junkio@cox.net> wrote:
>
> Andrew Morton <akpm@osdl.org> writes:
> 
> > But there was a second problem.  Once the parsing had misbehaved, Len
> > managed to create a commit which was six months in the future:
> >
> > commit 8313524a0d466f451a62709aaedf988d8257b21c
> > Author: Bob Moore <robert.moore@intel.com>
> > Date:   Tue Oct 3 00:00:00 2006 -0400
> >
> >     ACPI: ACPICA 20060310
> >
> > Will your fix prevent that from happening?  If not, perhaps some basic
> > sanity checking might be appropriate.
> 
> You _might_ get an e-mail to fix kernel problems from yourself
> in the future, in which case you would want to commit with
> future author date, like this ;-).
> 
> People would often deal with dates in the past (way in the past
> when talking about importing foreign SCM history), but probably
> it would never make sense to do dates way into the future.  I'll
> think about it.
> 

Well it doesn't have to be fatal, of course.  Some "do you really want to
do this [y/n]?" prompt, with a command option to override it.  Or simply
print a big warning.

Whatever.