[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Time for a release?



Manoj Srivastava <srivasta@debian.org> writes:

>         Now, in this specific case, since there has been only one commit
>  to the bug####-rra branches each, this is not an issue -- I can just
>  look at the commit.  But if you ever add another change to any of these
>  branches, one would have to do
>   git diff bb27c26983818b9fd4ee8bcca705c0381c47010a bug172436-rra
>  to see what changed.
>
>         I also note that master seems to have the change present on the
>  branch bug367984-rra; but it is not obvious that it came from
>  bug367984-rra (well, the commit message subject is the same, but ..).
>  Perhaps this merge --squash is not a good idea; I think it is nicer to
>  see gitk --all display the merge even if we get to see _all_ the
>  history.  What do you think?

I would be very happy to just merge everything all the time.  I don't
really understand the Git dislike of that workflow; merging everything
captures all of the history and also lets us delete old branches (like
bug367984-rra) without Git complaining that the changes haven't been
merged.

If that sounds good, I'll merge from master into all the pending bug
branches and change the wiki to suggest a simple merge onto master when a
bug is finalized.

I suppose the primary objection is that it makes for a messy branch
history, but with Policy I expect to be deleting the branches once they're
merged anyway, so I don't think that applies.

>> Should we call this upcoming version 3.8.0.0?  I think it has enough
>> substantial changes to warrant the larger version bump.
>
>         Sounds good to me.

I'll make that change in the repository.

-- 
Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>


Reply to: