Remove unused parameters from test code.

Remove unused parameters from test code.

Newer versions of the Android build system have -Wunused-params enabled
by default. Drop the many unused argc/argv parameters from test C/C++
code and also (void) a parameter in a dummy malloc function. This avoids
having to come up with a convoluted way to disable the warning flag in
every test.

BUG=gyp:433
R=thakis@chromium.org

Committed: https://code.google.com/p/gyp/source/detail?r=1938

[Torne, 2014-06-05 12:39:02]

I had a go at running the gyp tests against the latest AOSP (which we'll need to
do in order to test multilib) and this was the first issue that came up. I can't
think of a nice way to disable the warning globally, and it was pretty easy to
search+replace the test code, so does this seem reasonable as a solution? :)

[Torne, 2014-06-05 12:47:23]

(not intending to land this until gypbot is well again and it can run try jobs)

[Nico, 2014-06-05 16:59:18]

Seems reasonable.

(Do warnings make tests fail?)

https://codereview.chromium.org/316163002/diff/1/test/actions/src/subdir1/pro...
File test/actions/src/subdir1/program.c (right):

https://codereview.chromium.org/316163002/diff/1/test/actions/src/subdir1/pro...
test/actions/src/subdir1/program.c:6: int main()
Isn't it technically required to be `int main(void)` in C files?

[Torne, 2014-06-05 17:01:03]

On 2014/06/05 16:59:18, Nico (away) wrote:
> Seems reasonable.
> 
> (Do warnings make tests fail?)

The tests expect nothing to be output on stderr during the build, it looks like.

>
https://codereview.chromium.org/316163002/diff/1/test/actions/src/subdir1/pro...
> File test/actions/src/subdir1/program.c (right):
> 
>
https://codereview.chromium.org/316163002/diff/1/test/actions/src/subdir1/pro...
> test/actions/src/subdir1/program.c:6: int main()
> Isn't it technically required to be `int main(void)` in C files?

I can never remember. I can change it if void is correct.

[Torne, 2014-06-06 09:20:51]

Updated to use int main(void) instead just in case that makes a difference
somewhere. :)

[Nico, 2014-06-12 20:14:10]

lgtm

[Torne, 2014-06-13 09:48:16]

Committed patchset #2 manually as r1938 (presubmit successful).