Open Stack

On Mon, Sep 25 2017, Doug Hellmann wrote:

>> I was trying to ignore the thread in the hopes it would die out quick.
>> But torches and pitchforks all came out from the far corners, so I'm
>> going to push back on that a bit.
>> 
>> I'm not super clear why there is always so much outrage about these
>> patches. They are fixing real things. When I encounter them, I just
>> approve them to get them merged quickly and not backing up the review
>> queue, using more CI later if they need rebasing. They are fixing real
>> things. Maybe there is a CI cost, but the faster they are merged the
>> less likely someone else is to propose it in the future, which keeps
>> down the CI cost. And if we have a culture of just fixing typos later,
>> then we spend less CI time on patches the first time around with 2 or 3
>> iterations catching typos.
>> 
>> I think the concern is the ascribed motive for why people are putting
>> these up. That's fine to feel that people are stat padding (and that too
>> many things are driven off metrics). But, honestly, that's only
>> important if we make it important. Contributor stats are always going to
>> be pretty much junk stats. They are counting things to be the same which
>> are wildly variable in meaning (number of patches, number of Lines of
>> Code).
>> 
>> My personal view is just merge things that fix things that are wrong,
>> don't care why people are doing it. If it gets someone a discounted
>> ticket somewhere, so be it. It's really not any skin off our back in the
>> process.
>> 
>> If people are deeply concerned about CI resources, step one is to get
>> some better accounting into the existing system to see where resources
>> are currently spent, and how we could ensure that time is fairly spread
>> around to ensure maximum productivity by all developers.

Well said, agreed.

What I hate about this thread is that it's not the first one to pop up,
and it does not help to get the real problem getting acknowledge, which
IMHO is, that sometimes people spams Gerrit with small *wrong* patches.

But as you said, if they are real fixes, just blame yourself for merging
mistakes or typos in the first place.

> I'm less concerned with the motivation of someone submitting the
> patches than I am with their effect. Just like the situation we had
> with the bug squash days a year or so ago, if we had a poorly timed
> set of these trivial patches coming in at our feature freeze deadline,
> it would be extremely disruptive. So to me the fact that we're
> seeing them in large batches means we have people who are not fully
> engaged with the community and don't understand the impact they're
> having. My goal is to reach out and try to improve that engagement,
> and try to help them become more fully constructive contributors.

I think it's a good idea but I'm also pretty sure somebody already said
that (and probably did that) last time we had this discussion. While it
might or not (from my experience) improve this batch of contributors,
the next ones are already waiting to spam next cycle. ;)

-- 
Julien Danjou
;; Free Software hacker
;; https://julien.danjou.info
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 800 bytes
Desc: not available
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20170925/58f99622/attachment.sig>