New issue
Advanced search Search tips

Issue 662834 link

Starred by 1 user

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug

Blocking:
issue 662423



Sign in to add a comment

Difference between fullcode and crankshaft_opt: float->int array copy

Project Member Reported by machenb...@chromium.org, Nov 7 2016

Issue description

# Minimized program:
var a = new Float64Array(1);
function copy() { a[0] = undefined; }
copy();
print(new Int32Array(a.buffer)[1]);


# Compared nocrankshaft with noturbo_opt

# Flags of nocrankshaft:
--abort_on_stack_overflow --expose-gc --allow-natives-syntax --invoke-weak-callbacks --omit-quit  --random-seed -1285689746 --nocrankshaft
# Flags of noturbo_opt:
--abort_on_stack_overflow --expose-gc --allow-natives-syntax --invoke-weak-callbacks --omit-quit  --random-seed -1285689746 --always-opt --turbo-filter=~

Difference:
- 2146959360
+ -524288

### Start of configuration nocrankshaft:
2146959360

### End of configuration nocrankshaft

### Start of configuration noturbo_opt:
-524288

### End of configuration noturbo_opt

 
Status: WontFix (was: Untriaged)
Okay, this one we won't fix because it's not actually a spec violation.
Repurposing the float64array memory as int32 memory just shows that we've got different data in there. And the reason for different data that truncation occurs in one of the copy() calls and not the other. That's not an observable difference (until you repurpose the memory).

It's a bummer because we have to figure out a way to mark this as a known/won'tfix failure.
Labels: -Pri-1 Pri-2
Owner: machenb...@chromium.org
Status: Assigned (was: WontFix)
I put it on my pile of things to suppress.
Labels: -Restrict-View-Google

Sign in to add a comment