Hi,

> Oh, wait, do you mean that when you commit a fix to your code, you
> would add a second parent to that commit which refers to the issue
> that's fixed? So instead of having a special marker (says "Fixes:
> issue#<foo>") in your commit message, you refer to the issue as a
> parent?
Yep.

> Hmm... that sounds interesting.  I wonder what that allows you to do
> that the "Fixes: issue#<foo>" doesn't, tho.  I can see that it makes
> sure your issues-history is always preserved in the code-history, but
> I'm not sure if that's a feature.
It's a question of what's your use-case or work-flow. Especially for
patch-sets, it often makes sense to have some sort of description
preserved for future reference. Kernel-devs, for example, post a
patch-set with a cover letter. For bug-reports, conserving at least
some of the discussion may also be important.

Anyway, we might as well introduce other ways of (weak) references
through trailers, for more flexibility.

> > For example bugs can be files on a certain commit  
> 
> I don't understand what you mean by that.
Neither do I.

> [ The main problem of BuGit (other than lack of manpower) is that it
>   needs to be rewritten from scratch since it's currently just an sh
>   script.  ]
Well, if you look at git-dit's history you'll discover that it was
developed as a collection of bash scripts as well.

> One more thing, in
> https://github.com/neithernut/git-dit/blob/master/doc/datamodel.md you
> say that message metadata is kept in the form of "git-tags".  Do you
> mean the same tags as those managed by "git tag"?  How does that work?
"git-tag" is a bit ambiguous. By "git-tag" we mean trailers. Maybe we
should be more explicit on that front.

Regards,
jule
_______________________________________________
dist-bugs mailing list
[email protected]
https://kitenet.net/cgi-bin/mailman/listinfo/dist-bugs

Reply via email to