[PATCH] issue 1364

'Benoit Boissinot' bboissin at gmail.com
Mon Oct 27 10:48:07 CDT 2008


On Mon, Oct 27, 2008 at 06:34:52PM +0300, Andrei Vermel wrote:
>  
> >> - this is one potential proposed way to fix the problem - I do not 
> >> like the extend of changes required - it is nice to have one location 
> >> where stat calls are routed through though
> >
> > I guess another possibility is to implement a buggy statfiles() for PYTHON
> < 2.5.
> > How hard would it be ?
> Easy. Just allow for DST as if mtime were localtime (as on FAT).
> But honest fix of os.stat has an added benefit of sorting out the annoying
> dirstate rebuild twice a year after DST change. 

On the other hand, this only affects windows users with python2.4.
Is there a reason for Windows user not to update Python ? (for Linux it
might be harder because of the stable distributions that can ship old
software for a long time).

Personnaly I prefer a small hack that mimick the behaviour from python,
even if it's buggy. Is the dirstate rebuild really slow ?

regards,

Benoit

-- 
:wq


More information about the Mercurial-devel mailing list