Bug 146867 – AI folding of IsObjectOrNull is broken for non-object types that may be null

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Filip Pizlo]

Patch forthcoming.

[Filip Pizlo]

Created attachment 256638 [details]
the patch

[WebKit Commit Bot]

Attachment 256638 [details] did not pass style-queue:


ERROR: Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:1044:  Should have only a single space after a punctuation in a comment.  [whitespace/comments] [5]
ERROR: Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:1053:  Should have only a single space after a punctuation in a comment.  [whitespace/comments] [5]
Total errors found: 2 in 4 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Ryosuke Niwa]

Comment on attachment 256638 [details]
the patch

View in context: https://bugs.webkit.org/attachment.cgi?id=256638&action=review

> Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:1056
> +            if (!(child.m_type & ((SpecObject - SpecFunction) | SpecOther))) {

Can we define a local variable immediately before this line to reduce the number of parentheses involved?
e.g.
SpeculatedType nonFunctionObject = ((SpecObject - SpecFunction) | SpecOther);

[Ryosuke Niwa]

or maybe an inline helper function somewhere?

[Filip Pizlo]

(In reply to comment #3)
> Comment on attachment 256638 [details]
> the patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=256638&action=review
> 
> > Source/JavaScriptCore/dfg/DFGAbstractInterpreterInlines.h:1056
> > +            if (!(child.m_type & ((SpecObject - SpecFunction) | SpecOther))) {
> 
> Can we define a local variable immediately before this line to reduce the
> number of parentheses involved?
> e.g.
> SpeculatedType nonFunctionObject = ((SpecObject - SpecFunction) | SpecOther);

Well, we could, but nonFunctionObject would be the wrong name for this.  It's really nonFunctionObjectOrUndefinedOrNull.  At that point, the variable name is more complicated than the set expression!

[Filip Pizlo]

(In reply to comment #4)
> or maybe an inline helper function somewhere?

https://bugs.webkit.org/show_bug.cgi?id=146870

[Filip Pizlo]

Landed in http://trac.webkit.org/changeset/186702