OT: Distributed bug tracking?

[Steve Borho]

On Wed, 2008-01-02 at 20:32 +0000, Paul Sargent wrote:
> Apologies for the off topic post, but it might swing around to being  
> sort-of on topic.
> 
> I've never found a bug tracker I liked, but that maybe because I don't  
> know about the options. The big thing for me is that with all of the  
> decentralised advantages we have now, the other major tool for doing  
> collaborative development work is normally accessed through a web  
> site. Often slow and in an entirely different environment to where I  
> do all my other development work (CLIs and text editors), it becomes  
> totally inaccessible when on the road.
> 
> It strikes me that if the bug information was part of the source tree  
> (with a web interface somewhere so that users / managers can still  
> access it) then the information would be exactly where the developer  
> needed it. It would also be easy to see which change sets resolved  
> particular bugs because the change-set would include the change to the  
> bug report.
> 
> It seems like a simple directory will one file per bug report would be  
> 80% of the way there. You'd then need some scripts to handle them in a  
> controlled way.
> 
> Does anybody know of a system like this out there?
> If not, would it make sense to implement this as a mercurial  
> extension, with an extension to hgweb too?
> 
> e.g.:
>    hg bug open            # opens a new bug in text editor
>    hg bug append <bug-id> # opens text editor to allow more  
> information to be added
>    hg bug resolve <state> # sets the bug state to closed, duplicate,  
> etc.
>    hg bug report <option> # report all bugs that match some predicate
> 
> I think you might need some special merge logic, so that two people  
> appending or resolving a bug clash in reasonable ways.

You've mostly described Bugs Everywhere.

http://www.panoramicfeedback.com/opensource/index.html

It's bzr/arch centric, but I sent them a patch many months ago to
support hg (rather clumsily, if I remember correctly).

It's not a lot of Python code, so could probably get pulled into a
Mercurial extension without too much trouble.

-- 
Steve Borho (steve at borho.org)
http://www.borho.org/~steve/steve.asc
Key fingerprint = 2D08 E7CF B624 624C DE1F  E2E4 B0C2 5292 F2C6 2C8C