AudioEncoderOpusTest.PacketLossRateOptimized: Fix bug and make prettier

AudioEncoderOpusTest.PacketLossRateOptimized: Fix bug and make prettier

Fix bug 4981, which caused the second half (decreasing loss rates) to
not test anything. In the process, the test is changed slightly to
make it less dependent on the exact rounding behavior of doubles (by
not testing exactly at the the points where the effective loss rate
goes through a step---just very very close). A bunch of symbolic
constants are also replaced with easy-to-read literal numbers.

BUG=4981
Committed: https://crrev.com/942a699f14317927f9069a908e74c28338c5c5a8
Cr-Commit-Position: refs/heads/master@{#9908}

[kwiberg-webrtc, 2015-09-09 11:09:54]

kwiberg@webrtc.org changed reviewers:
+ minyue@webrtc.org

[minyue-webrtc, 2015-09-09 11:40:29]

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
File webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc
(left):

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc:117: for
(double loss = from; loss <= to;
all right, this was silly. Nice pick!

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
File webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc
(right):

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc:100: //
If a < b, returns a vector with the n evenly-spaced numbers min(a,b) + eps,
do you add/minus eps in this function?

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc:115:
return points;
does't returning std::vector involves a copy ctor? which might be less
efficient.

[kwiberg-webrtc, 2015-09-09 11:51:42]

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
File webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc
(right):

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc:100: //
If a < b, returns a vector with the n evenly-spaced numbers min(a,b) + eps,
On 2015/09/09 11:40:29, minyue-webrtc wrote:
> do you add/minus eps in this function?

I used to, but then I changed it. Patch set #2 changes this comment to match.

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc:115:
return points;
On 2015/09/09 11:40:29, minyue-webrtc wrote:
> does't returning std::vector involves a copy ctor? which might be less
> efficient.

Yes, in principle. But (1) this is a test, so we care much more about
readability than performance; (2) the lists only contain 10 doubles each, so
copying them isn't that expensive anyway; and (3) the compiler will use NVRO and
eliminate the copy (see e.g.
http://en.cppreference.com/w/cpp/language/copy_elision).

[minyue-webrtc, 2015-09-09 12:13:39]

lgtm

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
File webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc
(right):

https://codereview.webrtc.org/1316673010/diff/1/webrtc/modules/audio_coding/c...
webrtc/modules/audio_coding/codecs/opus/audio_encoder_opus_unittest.cc:115:
return points;
On 2015/09/09 11:51:42, kwiberg-webrtc wrote:
> On 2015/09/09 11:40:29, minyue-webrtc wrote:
> > does't returning std::vector involves a copy ctor? which might be less
> > efficient.
> 
> Yes, in principle. But (1) this is a test, so we care much more about
> readability than performance; (2) the lists only contain 10 doubles each, so
> copying them isn't that expensive anyway; and (3) the compiler will use NVRO
and
> eliminate the copy (see e.g.
> http://en.cppreference.com/w/cpp/language/copy_elision).

Acknowledged.

[kwiberg-webrtc, 2015-09-09 12:15:40]

The CQ bit was checked by kwiberg@webrtc.org

[commit-bot: I haz the power, 2015-09-09 12:15:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1316673010/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1316673010/20001

[commit-bot: I haz the power, 2015-09-09 13:43:00]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-09-09 13:43:12]

Patchset 2 (id:??) landed as
https://crrev.com/942a699f14317927f9069a908e74c28338c5c5a8
Cr-Commit-Position: refs/heads/master@{#9908}