Karthik Nayak <karthik....@gmail.com> writes:

> On 06/08/2015 10:51 PM, Matthieu Moy wrote:

>> We could introduce ref-filter.h earlier, indeed. To me, the current
>> solution is good enough, but introducing ref-filter.h early and adding
>> function definition there in the same commit as you drop the "static"
>> keyword for them would clearly be an improvement.
>
> But that would break the flow, wouldn't it? I wanted ref-filter to be
> introduced together, hence right after ref-filter.h we move code to
> ref-filter.c

That's why I find the current solution good enough: it also has
advantages. But in the current series, when you say "make functions
public", you are not actually doing so since they're not exported in a
.h file.

Conversely, PATCH 07 does two things: move code from for-each-ref.c and
introduce new declarations. Had you introduced these declarations
earlier, this patch would have been pure code movement.

In both cases, you have intermediate states that are not fully
consistant: either you have public functions in the builtin/ directory
(which sometimes happen in Git's codebase, but we try to avoid it), or
you have non-static functions that are not declared in a .h.

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to