From: Ramkumar Ramachandra <artagnon@gmail.com>
Date: Wed, 27 Jul 2011 18:03:30 GMT
Subject: Re: [PATCH 11/18] revert: Save data for continuing after conflict resolution
Message-ID: <CALkWK0nL2Lv8xSBQ0uBM-kuiK-aWDCqPmYUuyn5B7rPZMj5iAQ@mail.gmail.com>
In-Reply-To: <20110727155654.GB29924@elie>

Hi Jonathan,

Jonathan Nieder writes:
> Ramkumar Ramachandra wrote:
>> Junio complained about the warnings here [1], so I added MAYBE_UNUSED
>> to suppress them temporarily.  Would you really like to see
>> read_populate_todo and read_populate_opts being introduced MUCH later
>> in the '--continue' patch?
>>
>> [1]: http://thread.gmane.org/gmane.comp.version-control.git/176899
>
> Hmm.  Yes, I suppose I would.

Fixed.

> But as usual, what is relevant is not what I would like but what leads
> to clear, functional, and maintainable code.  It is from that point of
> view that the MAYBE_UNUSED is most problematic --- it suppresses a
> useful and true hint when compiling that the codebase implements some
> functionality that has no user.  For example, the warning would be
> triggered if the function gets renamed and added again and this old
> version is forgotten.

I actually think of it as a trade-off:
On one hand, I want to introduce functions to read and write
configuration files in the same patch.
On the other hand, I want to introduce functions when they're actually used.

-- Ram