how to *really* revert?

Matt Mackall mpm at selenic.com
Tue Aug 9 12:33:25 CDT 2005


On Tue, Aug 09, 2005 at 06:14:39PM +0200, Thomas Arendsen Hein wrote:
> * Andrew Thompson <andrewkt at aktzero.com> [20050809 17:55]:
> > Mark Williamson wrote:
> > >>yeh, hg status returns:
> > >>
> > >>dargo:~/xen-unstable.hg> hg status
> > >>M linux-2.6-xen-sparse/arch/xen/kernel/reboot.c
> > >>M tools/python/xen/lowlevel/xs/xs.c
> > >>M tools/python/xen/xm/main.py
> > >>M xen/arch/x86/mm.c
> > >>M xen/include/asm-x86/page.h
> > >
> > >Just a thought: from my memory, the changes detection code compares 
> > >timestamps before it compares file contents.  Perhaps if you've touched 
> > >those files at some point, they've got the wrong timestamps and that's 
> > >throwing this out?
> > 
> > That should not be the case. I believe the logic is check 
> > file-modified-time, if different look further(diff?), else continue on.
> 
> Yes, but before this, compare file size with dirstate system, if it
> is different, the file is changed in any case.
> 
> And the bug is, that the saved file size in dirstate has problems
> when merging in certain cases, see older mails here or just run the
> test suite.

No, this is the case where it's right but confusing.

-- 
Mathematics is the supreme nostalgia of our time.


More information about the Mercurial mailing list