mercurial mq <was> Re: RFR 8010325 : Remove hash32() method and hash32 int field from java.lang.String

[Paul Sandoz]

On Jun 14, 2013, at 10:36 PM, Martin Buchholz <martinrb at google.com> wrote:

> On Thu, Jun 13, 2013 at 12:46 PM, Brent Christian <
> brent.christian at oracle.com> wrote:
> 
>> On 6/12/13 7:55 PM, David Holmes wrote:
>> 
>>> Something of an aside but ...
>>> 
>>> On 13/06/2013 3:45 AM, Martin Buchholz wrote:
>>> 
>>>> Hi Brent,
>>>> 
>>>> Thanks for doing this.
>>>> 
>>>> Your webrev does not include mercurial changeset information, which I
>>>> think
>>>> is supported by recent webrevs.
>>>> 
>>> 
>>> Given the changeset has to be created after the review is complete
>>> most/many people will not have a changeset prepared at review time.
>>> 
>> 
>> That's it exactly.  If at all possible, I don't commit until the code has
>> completed code review.  I tell webrev to do its thing based on modified
>> files, rather than outgoing changesets.
> 
> 
> Hmmm.... I've been using mq for so long it's hard for me to imagine working
> without it.
> 
> It allows others to review the mercurial changeset metadata, which is also
> the best summary to decide whether to review further.


+1 

mq is the best way i have found to keep multiple patches in flight either in the same queue or using multiple queues, and avoid those annoying merge commits.

I wish there were better tooling support in IDEs for it.

Paul.