Delete unused TypeInfo class

Delete unused TypeInfo class

R=rossberg@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=18389

[Jakob Kummerow, 2013-12-13 16:21:47]

PTAL.

[rossberg, 2013-12-16 09:21:13]

LGTM sit comment

https://codereview.chromium.org/105313008/diff/1/src/property-details.h
File src/property-details.h (left):

https://codereview.chromium.org/105313008/diff/1/src/property-details.h#oldco...
src/property-details.h:116: // TODO(rossberg): this should die eventually.
Yes!

https://codereview.chromium.org/105313008/diff/1/src/type-info.h
File src/type-info.h (right):

https://codereview.chromium.org/105313008/diff/1/src/type-info.h#newcode45
src/type-info.h:45: static inline bool IsInt32Double(double value) {
We might consider moving this elsewhere, since it isn't really about "type
info". Maybe v8conversions.h?

[Sven Panne, 2013-12-16 09:50:40]

DBC

https://codereview.chromium.org/105313008/diff/1/src/type-info.h
File src/type-info.h (right):

https://codereview.chromium.org/105313008/diff/1/src/type-info.h#newcode45
src/type-info.h:45: static inline bool IsInt32Double(double value) {
On 2013/12/16 09:21:13, rossberg wrote:
> We might consider moving this elsewhere, since it isn't really about "type
> info". Maybe v8conversions.h?

... and perhaps refactor it to:

----------------------------------------
inline bool IsMinusZero(double value) {
  return DoubleRepresentation(value).bits == DoubleRepresentation(-0.0).bits;
}

static inline bool IsInt32Double(double value) {
  return !IsMinusZero(value) &&
      kMinInt <= value && value <= kMaxInt &&
      value == static_cast<int32_t>(value);
}
----------------------------------------

For added bonus points, use IsMinusZero at half a dozen similar places, add
operator=() to DoubleRepresentation, use IsInt32Double at several places,
perhaps add an IsInRange predicate. :-)

[Jakob Kummerow, 2013-12-17 16:38:48]

https://codereview.chromium.org/105313008/diff/1/src/type-info.h
File src/type-info.h (right):

https://codereview.chromium.org/105313008/diff/1/src/type-info.h#newcode45
src/type-info.h:45: static inline bool IsInt32Double(double value) {
On 2013/12/16 09:21:13, rossberg wrote:
> We might consider moving this elsewhere, since it isn't really about "type
> info". Maybe v8conversions.h?

Done.

https://codereview.chromium.org/105313008/diff/1/src/type-info.h#newcode45
src/type-info.h:45: static inline bool IsInt32Double(double value) {
On 2013/12/16 09:50:40, Sven Panne wrote:
> ... and perhaps refactor it to:
> 
> ----------------------------------------
> inline bool IsMinusZero(double value) {
>   return DoubleRepresentation(value).bits == DoubleRepresentation(-0.0).bits;
> }
> 
> static inline bool IsInt32Double(double value) {
>   return !IsMinusZero(value) &&
>       kMinInt <= value && value <= kMaxInt &&
>       value == static_cast<int32_t>(value);
> }
> ----------------------------------------

Done.

> For added bonus points, use IsMinusZero at half a dozen similar places, 

Done.

> add operator=() to DoubleRepresentation, 

Disagreed. DoubleRepresentation is a union; equality comparison semantics are
caller dependent, so having a default behavior by means of an operator=() would
just be confusing.

> use IsInt32Double at several places,

Done.

> perhaps add an IsInRange predicate. :-)

Unrelated.

[rossberg, 2013-12-17 16:51:17]

LEBTM

[Sven Panne, 2013-12-18 07:05:10]

https://codereview.chromium.org/105313008/diff/1/src/type-info.h
File src/type-info.h (right):

https://codereview.chromium.org/105313008/diff/1/src/type-info.h#newcode45
src/type-info.h:45: static inline bool IsInt32Double(double value) {
On 2013/12/17 16:38:48, Jakob wrote:
> > add operator=() to DoubleRepresentation, 
> 
> Disagreed. DoubleRepresentation is a union; equality comparison semantics are
> caller dependent, so having a default behavior by means of an operator=()
would
> just be confusing.

Nope, it would exactly be the right thing: The whole point of 
DoubleRepresentation is (surprise) a way to fiddle around with the
representation of doubles. So operator=() is extremely well-defined, namely
equality on the representation.

[Jakob Kummerow, 2013-12-18 10:41:42]

On 2013/12/18 07:05:10, Sven Panne wrote:
> https://codereview.chromium.org/105313008/diff/1/src/type-info.h
> File src/type-info.h (right):
> 
> https://codereview.chromium.org/105313008/diff/1/src/type-info.h#newcode45
> src/type-info.h:45: static inline bool IsInt32Double(double value) {
> On 2013/12/17 16:38:48, Jakob wrote:
> > > add operator=() to DoubleRepresentation, 
> > 
> > Disagreed. DoubleRepresentation is a union; equality comparison semantics
are
> > caller dependent, so having a default behavior by means of an operator=()
> would
> > just be confusing.
> 
> Nope, it would exactly be the right thing: The whole point of 
> DoubleRepresentation is (surprise) a way to fiddle around with the
> representation of doubles. So operator=() is extremely well-defined, namely
> equality on the representation.

Still don't agree, and also think the entire issue is utterly unrelated to the
point of this CL, but since this is the season for making other people happy...
Done.

[Jakob Kummerow, 2013-12-20 13:33:35]

Committed patchset #3 manually as r18389 (presubmit successful).