JS speed - Konq vs. Mozilla

[Koos Vriezen]

On Wed, 4 Sep 2002, David Faure wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Wednesday 04 September 2002 19:46, Koos Vriezen wrote:
> > Hmmm, now that all free slots are at the end of the block, the 'firstfree'
> > value of a block is always the same as 'filled'. So I stripped that out.
>
> Heh - indeed ;)
>
> > There is also no need for a 'memset' call in the CollectorBlock
> > constructor and setting '*r = 0L' in the collector anymore (because we
> > can iterate from 0 to filled). This also eliminates the need for
> > 'if (imp)' in the collector.
>
> Excellent stuff!

If no other comments, I'll commit it tomorrow.

>
> > I think this will be my final patch for the Collector for now, because
> > there is only 5-7% speed to gain here (compared to a allocator that just
> > returns 'malloc(size)').
> > Real speed gain should be found elsewhere. A gprof output for the tested
> > script, posted earlier in this thread, shows that KJS::Value::~Value() is
> > called 299335 and KJS::Value::imp() 356709 times!
>
> Shouldn't those be inlined, simply?

Would it not break BC (value.h is installed)? Btw, any ideas why
'putValue' is the only function where 'const Value' is not passed by
reference?

Koos