[Pulp-dev] Merging forward master -> 3.0-dev

[Brian Bouterse]

If we're not doing another merge forward, I think we should do the cleanup
soon.

The value of keeping that code around (to me) was this: you want to know
when a Pulp2 code change occurred and that code had already been ported to
Pulp3. That way you would know the Pulp2 changes had to also be ported to
Pulp3. The problem is: that value would only be known at "merge forward
time" via a merge conflicts. So if we're not merging forward again (for
core), then we can't realize that benefit again like we were. This is why I
think we're ready for cleanup soon.

Here is a cleanup task [0], in case others want to look/edit/comment on it.

[0]: https://pulp.plan.io/issues/2955

-Brian



On Thu, Aug 3, 2017 at 8:33 AM, Michael Hrivnak <mhrivnak at redhat.com> wrote:

>
>
> On Tue, Aug 1, 2017 at 5:33 PM, Brian Bouterse <bbouters at redhat.com>
> wrote:
>
>> Originally 3.0-dev was branched from master, and we've merged master back
>> into 3.0-dev twice, once at the end of 2016 and once in early 2017. I
>> haven't heard about any plans to merge master -> 3.0-dev again. I want to
>> check in if its time to declare that we aren't going to do that merge
>> forward again.
>>
>> I don't think there is much value in merging forward master -> 3.0-dev
>> again because the 3.0 code is so different than the 2.y code on master that
>> any diffs probably won't even make sense anymore. Also, the last merge
>> forward was pretty darn painful and it took a while.
>>
>> So with ^ in mind we could:
>>
>> 1) Decide that we don't need to merge forward master -> 3.0-dev again.
>>
>
> I agree. We don't likely have much to gain by trying another merge, and it
> would definitely be painful.
>
>>
>> 2) Clean up the 3.0-dev repo. Basically delete everything on 3.0-dev that
>> isn't part of the Pulp3 codebase. We can hammer out exactly what those
>> things are on an issue that isn't yet written.
>>
>>
> We were getting some value out of tracking what Pulp 2 code had been
> converted by removing it from the 3.0-dev branch as we went along. I'm not
> sure how well that's serving us though, and I'd be happy discontinuing that
> pattern. I figured at some point we would need a big purge, and doing that
> soon would be fine unless anyone finds the current pattern particularly
> helpful.
>
>
> --
>
> Michael Hrivnak
>
> Principal Software Engineer, RHCE
>
> Red Hat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-dev/attachments/20170803/8e4f9567/attachment.htm>