constexpr NaN,+Inf,-Inf

constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

I did try to write

    static constexpr float float_nan() { ... }

and completely failed.  constexpr seems a bit too restrictive in C++11 to make 
it work, but Clang kept telling me, you'll be able to do this with C++14.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002

Committed: https://skia.googlesource.com/skia/+/3ff2cc81a58b3fe5287980cf618147c84a6c81f6

[mtklein_C, 2016-08-10 13:32:07]

Description was changed from

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

BUG=skia:
==========

to

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002
==========

[mtklein_C, 2016-08-10 13:32:42]

mtklein@chromium.org changed reviewers:
+ reed@google.com

[mtklein_C, 2016-08-10 13:32:43]

The CQ bit was checked by mtklein@chromium.org

[commit-bot: I haz the power, 2016-08-10 13:32:57]

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

[mtklein_C, 2016-08-10 13:35:08]

Description was changed from

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002
==========

to

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

I did try to write

    static constexpr float float_nan() { ... }

and completely failed.  constexpr seems a bit too restrictive in C++11 to make
it work, but Clang kept telling me, you'll be able to do this with C++14.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002
==========

[mtklein_C, 2016-08-10 13:35:19]

Description was changed from

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

I did try to write

    static constexpr float float_nan() { ... }

and completely failed.  constexpr seems a bit too restrictive in C++11 to make
it work, but Clang kept telling me, you'll be able to do this with C++14.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002
==========

to

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

I did try to write

    static constexpr float float_nan() { ... }

and completely failed.  constexpr seems a bit too restrictive in C++11 to make 
it work, but Clang kept telling me, you'll be able to do this with C++14.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002
==========

[reed1, 2016-08-10 14:29:06]

reed@google.com changed reviewers:
+ caryclark@google.com

[reed1, 2016-08-10 14:29:06]

cary, do you remember why we deliberately put nan and inf in globals? Was it to
defeat some compiler inline for a test or something?

[caryclark, 2016-08-10 14:52:38]

On 2016/08/10 14:29:06, reed1 wrote:
> cary, do you remember why we deliberately put nan and inf in globals? Was it
to
> defeat some compiler inline for a test or something?

I don't remember. Defeating an inline seems likely though.
That code was last touched by Brian and Derek according to git blame.

[reed1, 2016-08-10 15:21:41]

reed@google.com changed reviewers:
+ bsalomon@google.com, djsollen@google.com

[reed1, 2016-08-10 15:21:41]

dreek, brian -- anyone object to NOT using a global to access infinity/nan? For
some reason in the past, we made these values globals, but cary and I can't
remember why.

[reed1, 2016-08-10 15:29:24]

after chat, brian can't remember either, so maybe whatever the issue was is now
gone (since your trybots are passing). lgtm

[commit-bot: I haz the power, 2016-08-10 15:31:45]

Description was changed from

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

I did try to write

    static constexpr float float_nan() { ... }

and completely failed.  constexpr seems a bit too restrictive in C++11 to make 
it work, but Clang kept telling me, you'll be able to do this with C++14.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002
==========

to

==========
constexpr NaN,+Inf,-Inf

Reading extern values meant these couldn't be compile-time constants.

math.h has INFINITY, which is macro that is supposed to expand to float +inf.
On MSVC it seems it's natively a double, so we cast just to make sure.

There's nan(const char*) in math.h for NaN too, but I don't trust that
to be compile-time evaluated.  So instead, we keep reinterpreting a bit pattern.

I did try to write

    static constexpr float float_nan() { ... }

and completely failed.  constexpr seems a bit too restrictive in C++11 to make 
it work, but Clang kept telling me, you'll be able to do this with C++14.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2233853002

Committed:
https://skia.googlesource.com/skia/+/3ff2cc81a58b3fe5287980cf618147c84a6c81f6
==========

[commit-bot: I haz the power, 2016-08-10 15:31:47]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/3ff2cc81a58b3fe5287980cf618147c84a6c81f6