I am still new to RB, and we are looking for something similar as well. We
have an SCR in our defect tracking system that aggregates all work for
change. Developers create a series of changesets and then tag each of those
with the SCR. When it's time to do a code review, we'd like to have one code
review for the entire SCR. So if I'm not mistaken, we are looking for the
same thing you are ("review request being for a series of changesets").

Here are my thoughts (FWIW) on this:

There needs to be a way add multiple changeset ids or revision numbers or
whatever on the command line. Currently we handle it through a manual

       use "hg log" to a list of changesets
       make a list of the changeset revision ids for the SCR, say (88, 89
and 93)
       starting with the oldest revision (88), create the review using "hg
postreview 88" (the mercurial-postreview allows this, don't know if the
others do)
       say the new review is id 33
       add the remaining revisions one by one from oldest to newest:
                 postreview -e 33  89
                 postreview -e 33  93

       when you go into RB GUI, the three changesets show up as three
revisions 1, 2, and 3 where 1->revision88, 2->revision89 and 3->revision93.

   - manual steps are error prone
   - if you screw up and add a revision you shouldn't have, not possible to
remove that from the review
   - if you screw up and forget to add a revision, not possible to insert
that into the existing review
   - the multiple revisions in the code review are confusing to the
   - During a review, it's not possible to tell if a problem in the code in
revision x is actually fixed later on in revision y.

So, in short, the extensions I'd like to see are:

  - the manual part of the review request automated. Perhaps it's as simple
as allowing multiple changesets to be put on the command line. Or it may
have to get more  complicated (e.g. list the revisions for the developer
through some gui and allow them to select the ones they want from that list)

  - have a "rollup" mechanism. As the additional revisions are added, the
changes are rolled up into one aggregate diff. (I assume this would be very
complicated code and so might not be doable without significant risk)

I'm not sure if there are other options available...


On Tue, Oct 26, 2010 at 7:25 PM, Andrew Schwartz <>wrote:

> I think that what I'm talking about is the same thing that Eduardo called
> "Remodelling diff versioning".  Does that seem right to you?
> On Tue, Oct 26, 2010 at 7:17 PM, Andrew Schwartz <>wrote:
>> Christian,
>> Sorry for letting this thread slide for so long.
>> I'm talking about a review request being for a series of changesets.  In
>> our case, we often will split a single feature change into small, easily
>> reviewed changesets.  It would be nice to be able to put all of these
>> changesets into a single review request.
>> Does that make sense?

Want to help the Review Board project? Donate today at
Happy user? Let us know at
To unsubscribe from this group, send email to
For more options, visit this group at

Reply via email to