<div dir="ltr"><br><div class="gmail_extra"><div class="gmail_quote">On Sun, Jun 2, 2013 at 7:50 PM, Tom Fifield <span dir="ltr"><<a href="mailto:tom@openstack.org" target="_blank">tom@openstack.org</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 02/06/13 11:40, Lorin Hochstein wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
I've noticed that there have been several bugs in the grizzly branch<br>
that were fixed in the master but hadn't been backported to<br>
stable/grizzly. I'd like to propose the following automated approach to<br>
ensure backports happen more regularly:<br>
<br>
Any patch to the master branch must specify in the commit message<br>
whether the patch should be backported. For example:<br>
<br>
backport: stable/grizzly<br>
<br>
or<br>
<br>
backport: none<br>
<br>
or<br>
<br>
backport: stable/grizzly stable/folsom<br>
<br>
<br>
If this line is missing from the commit log, then a gating job will<br>
fail, and jenkins will link to the error message (e.g., "Missing<br>
'backport: ' line. Please specify "backport: stable/grizzly" if this<br>
should be backported to grizzly or "backport: none" if this shouldn't be<br>
backported).<br>
<br>
When the patch is merged into trunk, then jenkins automatically does a<br>
cherry-pick and merge proposal against the branches specified in the<br>
backport line.<br>
<br>
<br>
What do you folks think?<br>
</blockquote>
<br>
<br></div>
The automation sounds like an excellent idea. As it is, I'm a terrible offender of not backporting, so I would be very happy to have an easier way to do so.<br>
<br>
Regarding having a "default fail" scenario, I believe this should be a second step. The first step should be education of people so they put it in and think about what it means, rather than just putting in something to make the error go away - or worse, getting scared off and abandoning the patch :)<br>
<br>
Once we've got some basis of contributors doing things the right way, then it should be easy enough to enable "default fail" :)<br>
<br>
Thoughts?<br>
<br></blockquote><div><br></div><div style>Yeah, good idea. We can configure it as "non-voting", so jenkins will report if the string is missing, but it won't vote -1 if the check fails.</div><div style><br>
</div><div style>Lorin</div><div><br></div></div><div dir="ltr"></div>
</div></div>