http://codereview.chromium.org/18089/diff/1/5
File src/jump-target-ia32.cc (right):

http://codereview.chromium.org/18089/diff/1/5#newcode106
Line 106: ASSERT(!cgen_->has_cc());
On 2009/01/15 12:21:54, William Hesse wrote:
> This looks like a duplicate line.

Oops.  Fixed.

http://codereview.chromium.org/18089/diff/1/7
File src/jump-target.h (right):

http://codereview.chromium.org/18089/diff/1/7#newcode38
Line 38: // TODO(): Update this comment.
On 2009/01/15 10:15:02, Kasper Lund wrote:
> You should probably just do this.

Done.

http://codereview.chromium.org/18089/diff/1/7#newcode102
Line 102: expected_frame_ = NULL;
On 2009/01/15 10:15:02, Kasper Lund wrote:
> Do we need to delete the expected frame here before setting it to
NULL? If not,
> it should probably be mentioned in a comment.

We don't want to delete it because it has normally been copied.  I think
the comment here before the function is pretty clear.

http://codereview.chromium.org/18089/diff/1/7#newcode127
Line 127: // become the current frame after the bind.
On 2009/01/15 12:21:54, William Hesse wrote:
> Are we allowing non-mergable expected frames, which might be changed
to be
> mergable or merged when bound, so that the frame after the binding
site won't be
> exactly the expected frame?

The frame after the binding site will always be the expected one.  I've
just changed the comment to be more specific.

http://codereview.chromium.org/18089/diff/1/7#newcode160
Line 160: void AddReachingFrame(VirtualFrame* frame) {
On 2009/01/15 10:15:02, Kasper Lund wrote:
> I don't know if this really should be inline. Maybe it would be better
to move
> it to the .cc file?

Done.

http://codereview.chromium.org/18089/diff/1/7#newcode184
Line 184: // jump target shadows other one, which is hidden as the
On 2009/01/15 12:21:54, William Hesse wrote:
> Not grammatical, confusing.

Added "the".

http://codereview.chromium.org/18089/diff/1/7#newcode199
Line 199: JumpTarget* other_target() const { return other_target_; }
On 2009/01/15 12:21:54, William Hesse wrote:
> Could this be called "active" target?  This is the JumpTarget passed
to the
> constructor, which always seems to be the active one, or the one that
is "seen".
>  The ShadowTarget represents the hidden target while it is masked by
another,
> and the target that was used to hide the original, after the hiding is
over,
> right?

No.  After shadowing they are both active.  I can't think of any better
name for this thing that is accurate for its whole life.  It is simply
the "other" target, the one that is not represented by the shadow
target.

http://codereview.chromium.org/18089/diff/1/4
File src/virtual-frame-ia32.cc (left):

http://codereview.chromium.org/18089/diff/1/4#oldcode321
Line 321: cgen_->SetFrame(this, &non_frame_registers);
On 2009/01/15 12:21:54, William Hesse wrote:
> Can you put an assert in, that the frame is its cgen's current frame?

Done.

http://codereview.chromium.org/18089

--~--~---------~--~----~------------~-------~--~----~
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
-~----------~----~----~----~------~----~------~--~---

Reply via email to