obliterate functionality?

cowwoc cowwoc at bbs.darktech.org
Tue Mar 18 17:27:41 CDT 2008


Matt Mackall wrote:
> On Tue, 2008-03-18 at 14:35 -0400, cowwoc wrote:
>> 1) Parent node may obliterate on its own code, when you pull changes 
>> from it you're forced to accept their obliterate (since they own that code).
> 
> So, for instance, if you were writing a Mercurial feature in your branch
> I thought was completely stupid (let's call it "burninate") but other
> people liked, I could pull your branch, obliterate your burninate work,
> and then suddenly everyone who pulls from me (including you) would have
> their burninate work permanently erased from their local history? That
> sounds very interesting!

	I think you misunderstood what I meant when I talked about a 
hierarchical architecture. The point is, you wouldn't pull code from me. 
Sun Microsystem's OpenJDK repository wouldn't pull code from individual 
contributors, rather those contributors would push code up to it.

	Those contributors could obliterate code in their local repository but 
no one would see that obliteration unless they set the contributor as 
their parent (as opposed to the official repository) so that's fine. 
Some contributors (say Sun employees) would be able to push obliteration 
up to the parent but that's okay because they have permissions to do so.

Gili


More information about the Mercurial mailing list