Re: [PATCH] Ensure cwrapper compiles without warnings under -std=c99.

[Charles Wilson]

Eric Blake wrote:
> According to Charles Wilson on 5/5/2008 6:23 PM:
> | -# func_emit_wrapper arg
> | +# func_emit_wrapper_part1 arg
>
> Since you provide a default, I'd show that arg is optional, as in:
>
> # func_emit_wrapper_part1 [arg=no]

Ack.

> Is func_emit_wrapper_part1_arg1 even used?  Why not just delete it?

No, it is not (currently) used. However, that's just an artifact of 
where I happened to "split" the original emit function: it just so 
happens that the only place the original emit function used the 
*original* arg, ended up in the second half, after I split the emit 
function into two parts.

I figured symmetry was easier for future maintainance, than strict 
adherence to the rule/pattern/guideline concerning supplying only the 
minimum argument set to each (sub)function called.

But it doesn't matter that much to me: if you still think I should 
remove the superfluous argument and make the (sub) functions 
un-symmetric, I will.

> Isn't puts slightly more efficient than printf?  But it doesn't matter
> that much to me.

Not part of the scope of this patch: the old version used printf, the 
new version uses printf. I figured I'd get dinged for changing more than 
was strictly necessary to silence the observed warnings -- this was 
supposed to be a very simple, uncontroversial patch that could've made 
it in to 2.2.4...

The "out of the scope of this patch" comment also applies to /all/ of 
Ralf's comments on this patch, but I'll address that in a separate message.

Besides, we know that printf can deal with very large strings (because 
so far it is working with 4K+ ones, even if gcc -std=c99 *warns* about 
it. We don't know that puts() does. (It probably does, but we don't 
*know* that).

--
Chuck