optimizing initial import

Matt Mackall mpm at selenic.com
Mon Jun 13 02:25:48 UTC 2005


On Sun, Jun 12, 2005 at 05:27:20PM -0400, Kevin Smith wrote:
> Arun Sharma wrote:
> >hg addremove
> >hg commit
> >
> >pays the penalty of scanning the filesystem twice, even though the user is 
> >unlikely to have made any changes to the filesystem in between.
> >
> >So why not add an option to either of these two commands to do 
> >addremove+commit?
> >
> >It's arguable that this is not a frequent operation that's worth 
> >optimizing for,
> >but might help win new users.
> 
> Interesting idea. I know darcs had some serious performance issues with 
> the initial import because it wasn't treated as a special case in the 
> code. Because darcs is already slow, importing a new large tree could 
> literally take hours. Hopefully mercurial is already better than that :-)
> 
> Forgetting the initial import case, I think I would like an option on 
> commit that would fail if addremove would have done something. I don't 
> think I want the commit to add new files automatically, but I would like 
> to be notified before I have committed without some necessary files.

This can be added to $HGEDITOR pretty easily, but I'll think about
adding an actual option..

-- 
Mathematics is the supreme nostalgia of our time.



More information about the Mercurial mailing list