[Adium-devl] Version Control

[Colin Barrett]

On Feb 6, 2008, at 9:21 PM, Peter Hosey wrote:

> Evan, as Colin noted, works on both Adium and Pidgin. So, if he  
> commits+pushes something to Pidgin's hypothetical hg repository, he  
> would use his @pidgin.im address; when he pulls it into Adium's  
> repo, it will keep that address, and we'll know that he committed it  
> as a Pidgin developer. When he commits+pushes directly to Adium,  
> he'd use his @adiumx.com address, and we'll know that he committed  
> it as an Adium developer.

How is this useful? Evan committed it. That's all we really need to  
know, right? What if he commits a patch in Adium's libpurple  
repository which then gets pulled into the main pidgin one? Which  
email address should he use then?

> More generally, a project-identifying email address clearly says at  
> a glance that the committer is or is not affiliated with our project  
> or Pidgin. We don't even need to look at the username: if the domain  
> is either adiumx.com or pidgin.im, it's an official patch;  
> otherwise, it's a user-submitted patch.

How is this useful? Either the patch is in the main repository or it  
isn't. What about this situation:

James Hacker takes an interest in the Adium project. He starts  
submitting patches, initially under j.random at hacker.com. Eventually,  
he becomes an official adium developer, and his @adiumx.com email is james at adiumx.com 
.

He's still the same person. The purpose of having an email address in  
the commit name, as you correctly argue, is to uniquely identify  
someone. James is still James, wether or not he's an Adium developer  
or not.

Or what about the case where someone "retires" from adium development,  
but still submits a few patches now and again. What email address  
should *they* use? It's not as simple a situation as you present, and  
honestly it's much simpler to have the email address have no semantic  
value other than "an email address people can use to contact you."

-Colin