Re: [PATCH 24/54] md/raid1: Improve another size determination in setup_conf()

From: Jes Sorensen
Date: Mon Oct 10 2016 - 09:11:10 EST

Dan Carpenter <dan.carpenter@xxxxxxxxxx> writes:
> On Thu, Oct 06, 2016 at 11:29:20AM +0200, Richard Weinberger wrote:
>> On Thu, Oct 6, 2016 at 11:22 AM, SF Markus Elfring
>> <elfring@xxxxxxxxxxxxxxxxxxxxx> wrote:
>> > From: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx>
>> > Date: Tue, 4 Oct 2016 21:46:18 +0200
>> >
>> > Replace the specification of a data structure by a pointer dereference
>> > as the parameter for the operator "sizeof" to make the corresponding size
>> > determination a bit safer.
>> Isn't this pure matter of taste?
>> Some developers prefer sizeof(*ptr) because it is easier to type, other
>> developers prefer sizeof(struct foo) because you can determine the type
>> at first sight and makes review more easy.
> sizeof(*ptr) is more future proof and normally more obvious and easier
> to review. That said, I've tried to tell Markus to only send bugfix
> patches because these are a waste of time and regularly introduce bugs.

This is totally a matter of taste. I for one find it way easier to
review something which says 'sizeof(struct ....)' because it stands out
more. I am curious what you mean by it being more future proof - if the
code says 'struct foo' in the sizeof argument, what is the problem?

The one area where there is a higher risk is if the type is changed, but
that is outweighed by the fact the spelled out version is easier to