[PATCH 2 of 5] subrepo: dirtiness checks should iterate over subrepos

Edouard Gomez ed.gomez at free.fr
Mon Mar 15 03:22:26 CDT 2010


On Sun, 14 Mar 2010 22:29:57 +0100, Saint Germain wrote:
> Is it somehow related to my proposition:
> http://www.selenic.com/pipermail/mercurial-devel/2010-March/019313.html
> (the subject line is wrong, don't pay attention to it)
> 
> In order to guarantee that I am in a defined state, I wanted "hg update
> -C" to force updating for 'dirty' subrepo.
> 
> Is it also what you intend with your patch ?

My patch intends to ensure that for a layout looking like:
main/
main/sub1
main/sub1/sub2

If we modify sub2, and commit in main, we have to end up with the three
repos commited.

So i would say your patch and mine are not related. However they try to
both improve subrepo usability.

It seems essential that subrepos feel like atomic (my patch) and
sufficiently easy to use so that if someone wants a clean tree,
subrepos should be clean too (your patch).

subrepos are new, it's normal that they don't feel right for now.
But now I'm bound to them at work for multiple projects, I'll provide
patches for the more simple fixes and feedback for usability requiring
too much hacking for me to complete them.

-- 
Edouard Gomez



More information about the Mercurial-devel mailing list