[PATCH 2 of 2] qseries: add --edit option for editing .hg/patches/series

[Augie Fackler]

On Fri, Apr 30, 2010 at 7:16 PM, Mads Kiilerich <mads at kiilerich.com> wrote:
>>
>> On Thursday, 29 April 2010 at 21:54, Dirkjan Ochtman wrote:
>>
>>>
>>> On Thu, Apr 29, 2010 at 13:31, Martin Geisler<mg at aragost.com>  wrote:
>>>
>>>>
>>>> Verify that the file can still can be parsed and that the user hasn't
>>>> reordered patches that are currently applied. Something like that.
>>>>
>>>
>>> That sounds nice, but I'd be happy to just take this for now and have
>>> that added later.
>>>
>>>
>>>>
>>>> I don't know what would be a good command line interface for this.
>>>>
>>>
>>> I think it's not very trivial, and I think just invoking the editor on
>>> series is kind of nice and simple.
>>>
>>
>> I don't think I agree. It's pretty easy to screw up your patch series
>> by misediting the series file. Adding this flag encourages monkeying
>> with the file.
>>
>>
>
> But ... sometimes the series file has to be monkeyed anyway. The proposed
> qpush --move might be usable in some situations, but it is not an efficient
> and sufficient UI for all tasks.
>
>> I think the editor ought to launch on a _copy_ of the series file, and
>> verify after editing that the series file contains the same set of
>> patches that it did before, and that no applied patches have changed
>> order, at the very least.
>>

Chiming in a bit late, but I'd put in a request for not verifying that
applied patches didn't change order. I frequently edit the series
order of applied patches, then qpop -a/qpush -a to reorder things.

>
> That would be easy to add, but sometimes other changes than reordering are
> done, so I'm not sure it would be nice to work with. Examples could be to
> disable patches by prefixing them with # (which often is simpler than using
> guards), or to mass-remove or -add guards.
>
> And what should it do after an invalid edit? Just drop the changes? Prompt
> the user and open the editor again?
>
> /Mads
> _______________________________________________
> Mercurial-devel mailing list
> Mercurial-devel at selenic.com
> http://selenic.com/mailman/listinfo/mercurial-devel
>