[Webkit-unassigned] [Bug 94712] High memory usage spike in AssemblerBuffer
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu Aug 23 14:54:00 PDT 2012
https://bugs.webkit.org/show_bug.cgi?id=94712
--- Comment #9 from Benjamin Poulain <benjamin at webkit.org> 2012-08-23 14:53:58 PST ---
> Not ready for review yet. But thanks for reviewing.
Yep, I noticed there was no r? only after making a few comments :)
> > > Source/WTF/wtf/SegmentedVector.h:-44
> > > - ~SegmentedVectorIterator() { }
> >
> > I think that was here so that some compiler do not generate a real function for the empty destructor.
>
> I can recover it easily. But isn't that a compiler bug? as generating a default destructor is so fundamental in C++. If we do think this could happen, shouldn't there be a coding rule saying we have to explicitly provide a destructor for every class in WebKit? Actually there are tons of classes in WebKit that don't have dtor specified.
Yeah, I may be wrong. It is just a vague memory from profiling.
> > > Source/WTF/wtf/SegmentedVector.h:112
> > > + size_t indexInSegment() const { return m_index; }
> > > +
> > > + void proceedToNextSegment()
> > > + {
> > > + if (++m_segment >= m_vector.m_segments.size()) {
> > > + // Points to the "end" symbol
> > > + m_segment = 0;
> > > + m_index = SegmentSize;
> > > + } else
> > > + m_index = 0;
> > > + }
> >
> > This is weird. Isn't this the wrong abstraction?
> Why?
This is exposing what SegmentedVector abstract, defeating the purpose of having an adt.
> > Why not give the "copy" capability to SegmentedVector?
>
> Like SegmentedVector::copyTo(T* dest, size_t pos, size_t len)? yeah, probably better.
Yep, that sounds good.
--
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
More information about the webkit-unassigned
mailing list