[turbofan] Check word32 -> float64 conversion with input type/output truncation.

[turbofan] Check word32 -> float64 conversion with input type/output truncation.

This makes sure we are using the right signedness - either the input
specifies signedness or the use must declare it does not care.

This requires some hole-punching in for-in so that the
representation selector understands the smi-ness of index.
Moreover, phi needs to pass the truncation along (maybe
slightly scary).

Committed: https://crrev.com/e5ea13e2309ba6a9bf7bcd22b1062987c70bb8de
Cr-Commit-Position: refs/heads/master@{#32081}

[Jarin, 2015-11-18 10:57:02]

Description was changed from

==========
[turbofan] Check word32 -> float64 conversion with input type/output truncation.

This makes sure we are using the right signedness - either the input
specifies signedness or the use must declare it does not care.
==========

to

==========
[turbofan] Check word32 -> float64 conversion with input type/output truncation.

This makes sure we are using the right signedness - either the input
specifies signedness or the use must declare it does not care.

This requires some hole-punching in for-in so that the 
representation selector understands the smi-ness of index.
Moreover, phi needs to pass the truncation along (maybe
slightly scary).
==========

[Jarin, 2015-11-18 10:57:02]

jarin@chromium.org changed reviewers:
+ bmeurer@chromium.org

[Jarin, 2015-11-18 10:57:31]

Could you take a look, please?

[Benedikt Meurer, 2015-11-18 11:54:31]

As discussed offline previously, I'm scared by the use of Guards to make this
"work", but I don't have a better idea yet. Can you add TODOs to the Guards that
we might want to find better ways?

Otherwise LGTM.

[Jarin, 2015-11-18 12:10:11]

On 2015/11/18 11:54:31, Benedikt Meurer wrote:
> As discussed offline previously, I'm scared by the use of Guards to make this
> "work", but I don't have a better idea yet. Can you add TODOs to the Guards
that
> we might want to find better ways?
> 
> Otherwise LGTM.

Actually, it turns out I do not need the guards - in the absence of good types,
representation selector will insert a truncation rather than a change. 

So the guards are gone now.

[Jarin, 2015-11-18 12:17:31]

The CQ bit was checked by jarin@chromium.org

[Jarin, 2015-11-18 12:17:31]

The patchset sent to the CQ was uploaded after l-g-t-m from bmeurer@chromium.org
Link to the patchset: https://codereview.chromium.org/1455123002/#ps20001
(title: "Get rid of the guards")

[commit-bot: I haz the power, 2015-11-18 12:17:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1455123002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1455123002/20001

[commit-bot: I haz the power, 2015-11-18 13:49:31]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-11-18 13:50:03]

Patchset 2 (id:??) landed as
https://crrev.com/e5ea13e2309ba6a9bf7bcd22b1062987c70bb8de
Cr-Commit-Position: refs/heads/master@{#32081}