Update audio code to use size_t more correctly,

Update audio code to use size_t more correctly,
webrtc/common_audio/signal_processing/ portion.

This is a piece of https://codereview.webrtc.org/1230503003 , split out to a
separate change to make reviewing easier.

BUG=chromium:81439
TEST=none

[Peter Kasting, 2015-07-10 16:47:08]

pkasting@chromium.org changed reviewers:
+ andrew@webrtc.org

[Peter Kasting, 2015-07-17 00:08:37]

pkasting@chromium.org changed reviewers:
+ jan.skoglund@webrtc.org

[Peter Kasting, 2015-07-17 00:09:01]

Adding jan.skoglund as alternate OWNER; can one of you reviewers take a look at
this please?

[jks, 2015-07-22 14:23:35]

LGTM

[Andrew MacDonald, 2015-07-22 18:53:08]

On 2015/07/17 00:09:01, Peter Kasting wrote:
> Adding jan.skoglund as alternate OWNER; can one of you reviewers take a look
at
> this please?

Sorry, was on vacation and haven't gotten to this yet.

[Peter Kasting, 2015-07-22 19:42:07]

On 2015/07/22 at 18:53:08, andrew wrote:
> On 2015/07/17 00:09:01, Peter Kasting wrote:
> > Adding jan.skoglund as alternate OWNER; can one of you reviewers take a look
at
> > this please?
> 
> Sorry, was on vacation and haven't gotten to this yet.

That's OK, Jan approved it.

[Andrew MacDonald, 2015-07-22 19:54:59]

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
File webrtc/common_audio/signal_processing/auto_correlation.c (right):

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
webrtc/common_audio/signal_processing/auto_correlation.c:23: if (order >
in_vector_length) {
assert this instead and return a size_t?

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
File webrtc/common_audio/signal_processing/min_max_operations.c (right):

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
webrtc/common_audio/signal_processing/min_max_operations.c:155: if (vector ==
NULL || length == 0) {
How would you feel about removing the error checks and returning a size_t
instead?

We can let it crash if vector == NULL, and assert(length > 0). Same for the rest
of these index functions.

For consistency, I think we can remove the checks in all the non-index functions
in this file as well.

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
File webrtc/common_audio/signal_processing/min_max_operations_mips.c (right):

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
webrtc/common_audio/signal_processing/min_max_operations_mips.c:27: if (vector
== NULL || length == 0) {
Remove the checks in this file as well, if that's what we end up with in the C
version.

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
File webrtc/common_audio/signal_processing/min_max_operations_neon.c (right):

https://codereview.webrtc.org/1228823002/diff/20001/webrtc/common_audio/signa...
webrtc/common_audio/signal_processing/min_max_operations_neon.c:20: if (vector
== NULL || length == 0) {
Again here.

[Peter Kasting, 2015-07-22 20:49:14]

I think your suggestions are reasonable, and I will implement them, but I'd like
to do so as a followup to the large size_t conversion happening here.  In the
current change I'm trying to avoid behavioral changes, as the patch is so large
and thus dangerous already that intentional behavioral changes increase the
likelihood of needing to roll it back.

I'll send you a separate CL to implement these and land it once the larger
change here is able to land.

[Peter Kasting, 2015-07-22 23:30:43]

On 2015/07/22 at 20:49:14, Peter Kasting wrote:
> I'll send you a separate CL to implement these and land it once the larger
change here is able to land.

You can now look at the CL to do this, plus convert all the connected bits to
size_t, at https://codereview.webrtc.org/1244383003 .

[Andrew MacDonald, 2015-07-22 23:43:14]

lgtm

On 2015/07/22 23:30:43, Peter Kasting wrote:
> On 2015/07/22 at 20:49:14, Peter Kasting wrote:
> > I'll send you a separate CL to implement these and land it once the larger
> change here is able to land.
> 
> You can now look at the CL to do this, plus convert all the connected bits to
> size_t, at https://codereview.webrtc.org/1244383003 .

Thanks.