[PATCH 0 of 2] patchbomb: ask for confirmation displaying summary of series

[Christian Ebert]

* Matt Mackall on Friday, January 29, 2010 at 13:44:31 -0600
> On Fri, 2010-01-29 at 19:12 +0000, Christian Ebert wrote:
>> * Matt Mackall on Friday, January 29, 2010 at 12:27:01 -0600
>>> On Fri, 2010-01-29 at 16:55 +0000, Christian Ebert wrote:
>>>> 1) Never prompt for cc, treat it like bcc
>>> 
>>> I don't really like this change. If I've used patchbomb for years
>>> expecting it to give me a cc: prompt when I give no args, I'll be
>>> bitten. I don't much like the interactive nature of patchbomb but my
>>> earlier suggestion (skip cc: prompt when from: is given)
>> 
>> You mean: when to: is given?
> 
> Right.

Ok. Others were pointing out though that the ui is bad if it
needs a table like this one:

opts     hgrc     prompt   result
                  new/old

to                NO/cc    to (opts)         *
to       to       NO/cc    to (opts)         *
to       cc       NO       to (opts), cc (hgrc)
to       to, cc   NO       to (opts), cc (hgrc)

to, cc            NO       to (opts), cc (opts)
to, cc   to       NO       to (opts), cc (opts)
to, cc   cc       NO       to (opts), cc (opts)
to, cc   to, cc   NO       to (opts), cc (opts)

cc                to       to (prompt), cc (opts)
cc       to       NO       to (hgrc), cc (opts)
cc       cc       to       to (prompt), cc (opts)
cc       to, cc   NO       to (hgrc), cc (opts)

                  to, cc   to (prompt), cc (prompt)
         to       cc       to (hgrc), cc (prompt)
         cc       to       to (prompt), cc (hgrc)
         to, cc   NO       to (hgrc), cc (hgrc)

But I'll go with the above then.

>>> is really about as far as I think we should bend things.
>>> 
>>>> 2) Before sending display Subject (with -v/--verbose only),
>>>>  From, To, Cc, Bcc and ask for user's confirmation
>>> 
>>> And I'm not sure about this one. It's not a fix for the shortcomings of
>>> (1) because it involves a new option.

Only to avoid misunderstandings: With --verbose I meant the
already existing global option.

>> I was taking out the --verbose switch anyway.
>> Confirmation with display of headers: yes or no?
> 
> Without the --verbose switch, definitely not.

Ok.

> That's going to interrupt people's well-established work flows,
> especially anyone who's gone as far as scripting their
> patchbombing. I'd take a dry-run option though.

patchbomb already has -n/--test, so I'll try to make this depend
on --verbose, right?

c
-- 
\black\trash movie   _S A M E_  _T I M E_  _S A M E_  _P L A C E_
                                  New York, in the summer of 2001

              --->> http://www.blacktrash.org/underdogma/stsp.php