[PATCH] mq: include involved files in qrefresh -e constructed message (issue3647)

Pierre-Yves David pierre-yves.david at ens-lyon.org
Fri Dec 7 18:37:48 CST 2012


On 7 déc. 2012, at 22:13, Davide Bolcioni wrote:

> Kevin Bullock [kbullock+mercurial at ringworld.org] wrote
> 
>> On Dec 3, 2012, at 6:45 PM, Davide Bolcioni wrote:
>> 
>>> # HG changeset patch
>>> # User Davide Bolcioni <dbolcioni at fb.com> # Date 1354053904 28800 # 
>>> Node ID 85615e499caecbaf758363fd90ad52b42e118a1c
>>> # Parent  fb14a5dcdc62987512820531fe60719d650491b6
>>> mq: include involved files in qrefresh -e constructed message
>>> (issue3647)
> 
> [cut]
> 
>>> +        if not msg:
>>> +            msg.append(_('(replace this line with the message)'))
>> 
>> We don't fill in a placeholder message for commit, why should qrefresh 
>> -e be any different?
> 
> The MQ extension adds more stuff to juggle in your mind as part of the context during the workflow, so making this look different from a commit seems advisable to me in order to remind the user that the two visually similar operations actually are somewhat different. I am open to suggestions about the wording, e.g. "... with the patch refresh reason".

My point is:

Is there any strong semantical different between

$ hg qrefresh
$ hg commit --amend

I do not think so and the commit --amend seems right to me currently.

You have yet to convinces me of the contrary.

Adding more details that an amend/refresh/histedit is in progress may be fine, but we should not alter the core logic.


-- 
Pierre-Yves


More information about the Mercurial-devel mailing list