Fix usage of many iterable functions on floating point typed lists.

Fix usage of many iterable functions on floating point typed lists.

BUG= http://dartbug.com/10823
R=lrn@google.com

Committed: https://code.google.com/p/dart/source/detail?r=23469

[Lasse Reichstein Nielsen, 2013-05-29 07:44:18]

lgtm

https://codereview.chromium.org/16135003/diff/1/runtime/lib/typed_data.dart
File runtime/lib/typed_data.dart (right):

https://codereview.chromium.org/16135003/diff/1/runtime/lib/typed_data.dart#n...
runtime/lib/typed_data.dart:305: dynamic combine(dynamic initialValue, element))
{
Inconsistent use of dynamic: it should be on element too, or not on
initialValue.
Using "dynamic" to represent one type and "omitted" (or, e.g., "var") to
represent another is too clever.

Either use dynamic for generic return type, or not, but do it consistently.

https://codereview.chromium.org/16135003/diff/1/runtime/lib/typed_data.dart#n...
runtime/lib/typed_data.dart:321: Iterable takeWhile(bool test(value)) {
value => element.
More below.

[floitsch, 2013-05-29 14:51:12]

PTAL.

https://codereview.chromium.org/16135003/diff/1/runtime/lib/typed_data.dart
File runtime/lib/typed_data.dart (right):

https://codereview.chromium.org/16135003/diff/1/runtime/lib/typed_data.dart#n...
runtime/lib/typed_data.dart:305: dynamic combine(dynamic initialValue, element))
{
On 2013/05/29 07:44:18, Lasse Reichstein Nielsen wrote:
> Inconsistent use of dynamic: it should be on element too, or not on
> initialValue.
> Using "dynamic" to represent one type and "omitted" (or, e.g., "var") to
> represent another is too clever.
> 
> Either use dynamic for generic return type, or not, but do it consistently.

Changed to `num`. That works for ints and doubles.

https://codereview.chromium.org/16135003/diff/1/runtime/lib/typed_data.dart#n...
runtime/lib/typed_data.dart:321: Iterable takeWhile(bool test(value)) {
On 2013/05/29 07:44:18, Lasse Reichstein Nielsen wrote:
> value => element.
> More below.

Done.

[Lasse Reichstein Nielsen, 2013-05-31 09:18:21]

lgtm

https://codereview.chromium.org/16135003/diff/5001/runtime/lib/typed_data.dart
File runtime/lib/typed_data.dart (right):

https://codereview.chromium.org/16135003/diff/5001/runtime/lib/typed_data.dar...
runtime/lib/typed_data.dart:285: void forEach(void f(num element)) {
'num' works for callbacks and return values, and we don't seem to use it for
elements to be inserted into the array, so this should be fine.

Ofcourse, it would be better if _TypedListBase was generic, and the type was T.
I assume that was avoided because the VM can't optimize generic types in
superclasses yet.

[kasperl, 2013-05-31 10:01:02]

https://codereview.chromium.org/16135003/diff/5001/tests/lib/typed_data/typed...
File tests/lib/typed_data/typed_list_iterable_test.dart (right):

https://codereview.chromium.org/16135003/diff/5001/tests/lib/typed_data/typed...
tests/lib/typed_data/typed_list_iterable_test.dart:4: //
VMOptions=--deoptimization_counter_threshold=1000
Intentional?

[floitsch, 2013-05-31 15:06:25]

https://codereview.chromium.org/16135003/diff/5001/runtime/lib/typed_data.dart
File runtime/lib/typed_data.dart (right):

https://codereview.chromium.org/16135003/diff/5001/runtime/lib/typed_data.dar...
runtime/lib/typed_data.dart:285: void forEach(void f(num element)) {
On 2013/05/31 09:18:21, Lasse Reichstein Nielsen wrote:
> 'num' works for callbacks and return values, and we don't seem to use it for
> elements to be inserted into the array, so this should be fine.
> 
> Ofcourse, it would be better if _TypedListBase was generic, and the type was
T.
> I assume that was avoided because the VM can't optimize generic types in
> superclasses yet.

Yes. I talked to the VM people. No for generic type. (also wouldn't compile
anymore).

https://codereview.chromium.org/16135003/diff/5001/tests/lib/typed_data/typed...
File tests/lib/typed_data/typed_list_iterable_test.dart (right):

https://codereview.chromium.org/16135003/diff/5001/tests/lib/typed_data/typed...
tests/lib/typed_data/typed_list_iterable_test.dart:4: //
VMOptions=--deoptimization_counter_threshold=1000
On 2013/05/31 10:01:02, kasperl wrote:
> Intentional?

good catch.

[floitsch, 2013-05-31 15:06:42]

Committed patchset #3 manually as r23469 (presubmit successful).