Obsolete Development

Pierre-Yves David pierre-yves.david at logilab.fr
Tue Aug 7 12:34:03 CDT 2012


On Tue, Aug 07, 2012 at 12:25:03PM -0500, Sean Farley wrote:
> > Thanks for your interrest in evolve!
> 
> Thanks to you and Patrick (and unnamed others, I'm sure) for answering
> and providing the patches.
> 
> > Both document and code are hosted on bitbucket:
> >
> >     https://bitbucket.org/marmoute/hgview-mutable
> >
> > Contribution can be submitted by pull request from either bitcket or obsolete
> > enabled repo[1].
> 
> Ok, will do. I have a lot of feedback for documentation, mostly, and
> some for code but I'm still new to the codebase and it's a little much
> for me right now. I will try to at least provide some tracebacks.

Beware that I'm currently doing a major cleanup for the obsolete and evolve
code base (I may even merge them both). You may want to hold yourself a few day
before working in the code.

This does not apply to documention.

> > I believe that discussion can happen on the mercurial list for now. We'll move
> > to a dedicated list if the amount get out of control. As a lot of code are
> > aimed for code, UI and terminolgy discussion deserves to happen on this list
> > anyway.
> 
> First of all, using evolve is amazingly powerful and is a breath of
> fresh air coming from using mq. As has been mentioned by others
> before, I don't quite know what to do about the abandoned changesets.
> So far, they get in the way of
> 
> 1) histedit (aborts because of orphaned nodes)

I plan to work on histedit obsolescence aware soon™

> 2) local numbering is out-of-whack (these have been left out of the
> documentation I guess for this reason?)

Yeah, I do not think this will change. People should not rely on local
numbering too much. Moreover this local numbering make sense when you use

    hg log --hidden


> 
> If I haven't pushed to anywhere, I usually run:
> 
> $ hg strip -r 'hidden()'
> 
> and that seems to work. So far, this has worked in my simple test
> cases even with pushing to other repos but is there a danger to
> stripping these changesets?

There is not danger exect the fact you "loose" your backup and ability to
consulte rewritting history.

Note: I'm interrested in people keeping their extinct changeset for performance
      impact.

-- 
Pierre-Yves David

http://www.logilab.fr/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20120807/034acbe6/attachment.pgp>


More information about the Mercurial-devel mailing list