Re: bypassing countdown cycle for trivial patches?

[David Kastrup]

Urs Liska <address@hidden> writes:

> Am 03.07.2014 12:31, schrieb Mark Polesky:
>> Do I have permission to push trivial patches straight to staging?
>> Sometimes I want to fix typos and such without bothering with
>> an issue number and a countdown cycle.
>> Here's an example:
>> http://codereview.appspot.com/104500043/
>>
>> I don't want to make waves by violating protocol, but would this
>> be alright?
>>
>> - Mark
>
> I just had this issue and was strongly suggested to wait at least for
> the "passes ..." message.

Which requires an issue number, as created by git cl.

> After that it seems to be considered OK to push trivial patches to staging.

Yup.

A Rietveld review is sort of pointless without a corresponding issue.  I
have, on occasion, pushed trivial patches without review to staging, but
you better make very, very sure that the patch does what you intended
and nothing else.  Also things like changes in snippets/new tend to
create diffs in unrelated makelsr runs.  I lean towards running makelsr
(in a separate commit) in order to avoid this.  It also avoids pushing
bad material unnoticed by Patchy tests that will blow up only once
makelsr is run for an unrelated change.

-- 
David Kastrup