Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase >> mail # dev >> suggestion for smoother code review process

Copy link to this message
Re: suggestion for smoother code review process
Step 1 for Phabricator is to reach parity with the current Review Board
utilities.  Step 2 is to improve formatting and minimize redundancy.  I
agree with Jonathan's comments: if you see something got added to RB/Phab,
interact with the dialog there instead of trying to use JIRA directly.

On 10/20/11 3:22 PM, "Ted Yu" <[EMAIL PROTECTED]> wrote:

>I looked at John Sichi's comment, obviously issued from phabricator, for
>HBASE-4532 @ 18/Oct/11 23:41
>I don't see much difference from feedback from review board - AFFECTED
>were included.
>One thing I do like the postback from review board is the nice layout
>viewable in Yahoo email but not gmail (strangely).
>On Thu, Oct 20, 2011 at 3:11 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
>> I think the easiest improvement is to strip the list of files from
>> reviewboard feedback.
>> I would wait for a while to see if any volunteer comes up for the above
>> task :-)
>> I am not sure about phabricator which requires an account.
>> I remember seeing phabricator feedback in JIRA. The format is different.
>> Cheers
>> On Thu, Oct 20, 2011 at 3:05 PM, Todd Lipcon <[EMAIL PROTECTED]> wrote:
>>> Hey Ted,
>>> I agree the formatting of the reviewboard comments back onto JIRA
>>> could be improved. I wrote the original script that does it - it's
>>> some nasty procmail and python.
>>> It sounds like the FB folks are working on getting phabricator up -
>>> maybe it will have better JIRA integration?
>>> Let me know if you have some time to spend on improving the
>>> python/procmail setup with RB. I can connect you with the right infra
>>> people to make the change.
>>> -Todd
>>> On Thu, Oct 20, 2011 at 3:03 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
>>> > Hi,
>>> > We have been using review board for a while to conduct code review.
>>> > One aspect I don't like the integration is that every round of review
>>> would
>>> > result in the summary and list of files (both of which could be
>>>long) to
>>> be
>>> > reposted to JIRA.
>>> > For a large project, such as HBASE-2856 or HBASE-3777, it is
>>> > (without exaggeration) for a developer who didn't closely follow the
>>> > development to understand what was going on.
>>> >
>>> > I want to share what I have been doing recently (by not commenting on
>>> review
>>> > board, if possible):
>>> > I would quote the snippet of code in the patch and make my comment
>>> >
>>> > I think the person asking for review can post the url for review
>>> > request on the JIRA. By not filling Bugs field, we don't incur extra
>>> > housekeeping that I mentioned earlier.
>>> > If the Groups and People fields are filled properly, there is no
>>>risk of
>>> > losing review request. In the worst case, one sentence on the JIRA
>>> > remind related people to look at the patch again.
>>> >
>>> > Note the above is just personally advice. Please don't interpret it
>>> rule
>>> > or anything like that.
>>> >
>>> > Cheers
>>> >
>>> --
>>> Todd Lipcon
>>> Software Engineer, Cloudera