Refactor byte_io to avoid potentially undefined behavior

Make sure ByteReader and ByteWriter classes (and their specializations) don't perform operations that have implementation-specific or undefined behavior.

Pitfalls:

* Left shift of signed integer has undefined behavior
* Right-shift of signed integer has platform-specific behavior is value is negative
* Cast from unsigned to signed has undefined behavior if value is negative

BUG=webrtc:4824
Committed: https://crrev.com/be9b7b6881e5b0e0b54e7d2fb79c5af5f68c015b
Cr-Commit-Position: refs/heads/master@{#9854}

[sprang_webrtc, 2015-07-08 11:01:28]

sprang@webrtc.org changed reviewers:
+ stefan@webrtc.org

[stefan-webrtc, 2015-07-08 12:01:50]

https://codereview.webrtc.org/1226993003/diff/1/webrtc/modules/rtp_rtcp/sourc...
File webrtc/modules/rtp_rtcp/source/byte_io.h (right):

https://codereview.webrtc.org/1226993003/diff/1/webrtc/modules/rtp_rtcp/sourc...
webrtc/modules/rtp_rtcp/source/byte_io.h:92: return ~used_bits_mask | val;
I think this deserves a comment. :)

[sprang_webrtc, 2015-07-08 12:57:03]

Addressed comment: Added comment.

https://codereview.webrtc.org/1226993003/diff/1/webrtc/modules/rtp_rtcp/sourc...
File webrtc/modules/rtp_rtcp/source/byte_io.h (right):

https://codereview.webrtc.org/1226993003/diff/1/webrtc/modules/rtp_rtcp/sourc...
webrtc/modules/rtp_rtcp/source/byte_io.h:92: return ~used_bits_mask | val;
On 2015/07/08 12:01:50, stefan-webrtc (holmer) wrote:
> I think this deserves a comment. :)

Done.

[stefan-webrtc, 2015-07-08 13:29:10]

lgtm

[sprang_webrtc, 2015-07-10 15:23:03]

I did further refactoring of this in order to remove any and all instances where
undefined behavior could potentially occur. Even though it doesn't seem to
happen on any of the platforms covered by the try bots. This is probably (by
far) the most complex no-op I have every written :D

[pbos-webrtc, 2015-07-27 14:11:15]

pbos@webrtc.org changed reviewers:
+ pbos@webrtc.org

[pbos-webrtc, 2015-07-27 14:31:19]

As discussed offline make a simpler check for two's complement and assume that
if one type is two's complement then all are.

[stefan-webrtc, 2015-09-01 12:24:51]

Is this CL pending review or action? Seems like it would be good to get it in to
resolve the bug.

[sprang_webrtc, 2015-09-02 12:49:30]

Please review.

[pbos-webrtc, 2015-09-03 15:18:08]

lgtm

https://codereview.webrtc.org/1226993003/diff/270001/webrtc/modules/rtp_rtcp/...
File webrtc/modules/rtp_rtcp/source/byte_io.h (right):

https://codereview.webrtc.org/1226993003/diff/270001/webrtc/modules/rtp_rtcp/...
webrtc/modules/rtp_rtcp/source/byte_io.h:247: // TODO(sprang): Check if these
actually help or if generic cases will be
Do you want to add these before you know that it's useful?

[sprang_webrtc, 2015-09-04 07:24:06]

https://codereview.webrtc.org/1226993003/diff/270001/webrtc/modules/rtp_rtcp/...
File webrtc/modules/rtp_rtcp/source/byte_io.h (right):

https://codereview.webrtc.org/1226993003/diff/270001/webrtc/modules/rtp_rtcp/...
webrtc/modules/rtp_rtcp/source/byte_io.h:247: // TODO(sprang): Check if these
actually help or if generic cases will be
On 2015/09/03 15:18:08, pbos-webrtc wrote:
> Do you want to add these before you know that it's useful?

With this comment I meant all the specializations below.

[sprang_webrtc, 2015-09-04 07:24:14]

The CQ bit was checked by sprang@webrtc.org

[sprang_webrtc, 2015-09-04 07:24:15]

The patchset sent to the CQ was uploaded after l-g-t-m from stefan@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1226993003/#ps270001 (title:
"Use numeric limits to get signed min-val")

[commit-bot: I haz the power, 2015-09-04 07:24:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1226993003/270001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1226993003/270001

[commit-bot: I haz the power, 2015-09-04 08:07:01]

Committed patchset #15 (id:270001)

[commit-bot: I haz the power, 2015-09-04 08:07:12]

Patchset 15 (id:??) landed as
https://crrev.com/be9b7b6881e5b0e0b54e7d2fb79c5af5f68c015b
Cr-Commit-Position: refs/heads/master@{#9854}