Bug#675790: "dch -r" refuses to proceed

[Luk Claes]

On 06/03/2012 10:20 PM, James McCoy wrote:
> 
> On Jun 3, 2012 4:12 PM, "Harald Dunkel" <harri at afaics.de
> <mailto:harri at afaics.de>> wrote:
>>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 06/03/12 21:13, James McCoy wrote:
>> >
>> > No, it's not a misunderstanding.  Just make vi write the file again
> (:wq).  We use the modification time to determine whether the changes
> are acceptable.
>>
>> Of course I see that you use the modification date to check if
>> the changelog file was modified, but there is no reason for this
>> check on "dch -r", since we know for sure that dch had changed it.
> 
> We know for sure that dch has modified the changelog when you just run
> dch, too. That doesn't mean that you've accepted the changes that were
> made.
> 
>> > Alternatively, you can one of the options suggested in the man page.
> Either "dch -r """ (i.e., run -r and supply an empty changelog message)
> or use --no-force-save-on-release.
>> >
>>
>> So you mean that all Debian developers using dch have to add empty
>> comments or long command line options, just to avoid some code
>> rework in dch?
> 
> No. They just run :wq to exit Vim (or whatever's appropriate for their
> editor) and everything works as expected.  If they don't want to do
> that, then they set the appropriate configuration options in ~/.devscripts.

No, it does not, that's the whole point of this bug report. Please try
$ dch -r
:wq!

a.k.a. without changing anything else.

> It's really not as hard as you're making it out to be.

So what's exactly wrong in the above way of using it and how is it fixed
in that case?

Cheers

Luk