RTC_[D]CHECK_op: Remove superfluous casts

RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.

NOPRESUBMIT=true
BUG=webrtc:6645
Committed: https://crrev.com/352444fcace861c0b0e2b1654187f5244e1be1f4
Cr-Commit-Position: refs/heads/master@{#15281}

[kwiberg-webrtc, 2016-11-26 09:49:23]

The CQ bit was checked by kwiberg@webrtc.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-26 09:49:26]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2016-11-26 09:56:16]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-26 09:56:17]

Dry run: Try jobs failed on following builders:
  win_clang_dbg on master.tryserver.webrtc (JOB_FAILED,
http://build.chromium.org/p/tryserver.webrtc/builders/win_clang_dbg/builds/8641)

[kwiberg-webrtc, 2016-11-26 22:11:40]

The CQ bit was checked by kwiberg@webrtc.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-26 22:11:48]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2016-11-26 22:23:18]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-26 22:23:19]

Dry run: Try jobs failed on following builders:
  win_x64_dbg on master.tryserver.webrtc (JOB_FAILED,
http://build.chromium.org/p/tryserver.webrtc/builders/win_x64_dbg/builds/3552)

[kwiberg-webrtc, 2016-11-26 22:59:08]

Patchset #1 (id:1) has been deleted

[kwiberg-webrtc, 2016-11-26 22:59:16]

Patchset #1 (id:20001) has been deleted

[kwiberg-webrtc, 2016-11-26 23:01:05]

Description was changed from

==========
RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

NOPRESUBMIT=true
BUG=webrtc:6645
==========

to

==========
RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.

NOPRESUBMIT=true
BUG=webrtc:6645
==========

[kwiberg-webrtc, 2016-11-26 23:01:05]

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

[kwiberg-webrtc, 2016-11-26 23:01:44]

This should be the final CL for this bug!

[ossu, 2016-11-28 14:14:34]

I've only one question, mainly because I'm interested. I don't think it's reason
enough to block this CL.

lgtm

https://codereview.webrtc.org/2534683002/diff/40001/webrtc/base/safe_compare.h
File webrtc/base/safe_compare.h (right):

https://codereview.webrtc.org/2534683002/diff/40001/webrtc/base/safe_compare....
webrtc/base/safe_compare.h:164: static char Test(...);
I tried to find an answer to this, and clearly this isn't the case on any of our
current platforms, but could char and int be the same type? I guess they could
be the same size on some machine that only reads words, but would that also mean
they're the same type? If so, the check below would break down.

[kwiberg-webrtc, 2016-11-28 23:20:37]

https://codereview.webrtc.org/2534683002/diff/40001/webrtc/base/safe_compare.h
File webrtc/base/safe_compare.h (right):

https://codereview.webrtc.org/2534683002/diff/40001/webrtc/base/safe_compare....
webrtc/base/safe_compare.h:164: static char Test(...);
On 2016/11/28 14:14:34, ossu wrote:
> I tried to find an answer to this, and clearly this isn't the case on any of
our
> current platforms, but could char and int be the same type? I guess they could
> be the same size on some machine that only reads words, but would that also
mean
> they're the same type? If so, the check below would break down.

Yes, they can be the same size, but I think they'd still be separate types. You
can run into problems on real systems today because long long and long, although
they're both 64 bits, aren't the same type. Example where the compiler is able
to tell them apart, even though they're both 64 bits:
https://godbolt.org/g/rVMNcI

[kwiberg-webrtc, 2016-11-28 23:27:46]

The CQ bit was checked by kwiberg@webrtc.org

[commit-bot: I haz the power, 2016-11-28 23:27:52]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2016-11-28 23:58:54]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1480375666574640,
"parent_rev": "ab62b56fe6a05d08920e611fd804571cefe27a43", "commit_rev":
"a39ccbc271bb7ac35da80c0e43b4d59edfd75c19"}

[commit-bot: I haz the power, 2016-11-28 23:58:57]

Description was changed from

==========
RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.

NOPRESUBMIT=true
BUG=webrtc:6645
==========

to

==========
RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.

NOPRESUBMIT=true
BUG=webrtc:6645
==========

[commit-bot: I haz the power, 2016-11-28 23:58:58]

Committed patchset #1 (id:40001)

[commit-bot: I haz the power, 2016-11-28 23:59:08]

Description was changed from

==========
RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.

NOPRESUBMIT=true
BUG=webrtc:6645
==========

to

==========
RTC_[D]CHECK_op: Remove superfluous casts

There's no longer any need to make the two arguments have the same
signedness, so we can remove a bunch of superfluous (and sometimes
dangerous) casts.

It turned out I also had to fix the safe_cmp functions to properly handle
enums that are implicitly convertible to integers.

NOPRESUBMIT=true
BUG=webrtc:6645

Committed: https://crrev.com/352444fcace861c0b0e2b1654187f5244e1be1f4
Cr-Commit-Position: refs/heads/master@{#15281}
==========

[commit-bot: I haz the power, 2016-11-28 23:59:09]

Patchset 1 (id:??) landed as
https://crrev.com/352444fcace861c0b0e2b1654187f5244e1be1f4
Cr-Commit-Position: refs/heads/master@{#15281}

[ossu, 2016-11-29 10:28:23]

https://codereview.webrtc.org/2534683002/diff/40001/webrtc/base/safe_compare.h
File webrtc/base/safe_compare.h (right):

https://codereview.webrtc.org/2534683002/diff/40001/webrtc/base/safe_compare....
webrtc/base/safe_compare.h:164: static char Test(...);
On 2016/11/28 23:20:37, kwiberg-webrtc wrote:
> On 2016/11/28 14:14:34, ossu wrote:
> > I tried to find an answer to this, and clearly this isn't the case on any of
> our
> > current platforms, but could char and int be the same type? I guess they
could
> > be the same size on some machine that only reads words, but would that also
> mean
> > they're the same type? If so, the check below would break down.
> 
> Yes, they can be the same size, but I think they'd still be separate types.
You
> can run into problems on real systems today because long long and long,
although
> they're both 64 bits, aren't the same type. Example where the compiler is able
> to tell them apart, even though they're both 64 bits:
> https://godbolt.org/g/rVMNcI

Cool beans! Then I'm satisfied. :)