[PATCH 2] Speeding up kdebug - using print member
Roger Larsson
roger.larsson at norran.net
Fri Oct 17 19:08:39 CEST 2003
On Friday 17 October 2003 17.28, David Faure wrote:
> On Friday 17 October 2003 17:25, Roger Larsson wrote:
> > On Friday 17 October 2003 11.33, David Faure wrote:
> > > On Friday 17 October 2003 08:52, Roger Larsson wrote:
> > > > No comments on this patch? (Updated to newest CVS)
> > > >
> > > > Really three parts that are in one patch.
> > > > 1) All pipe operators should check 'print'
> > > > 2) We should determine if printout is wanted up front, not when most
> > > > of the work is done.
> > >
> > > Looks fine to me.
> > > The next step would be to compare kcachegrind results, to see if it's
> > > indeed faster ;)
> >
> > Tried to clock this - there are to few printouts to make a difference...
>
> calltree/kcachegrind gives very precise results, not clock related,
> but instructions count related.
Yea, but to retest it I would need to recompile it... Hmm... Maybe if I do not
change the header...
Problem is that I will go on a two week vacation tomorrow and that I have a
worse problem. My mailbox fills up in a day with the trojan Swen (or
something resembling it). So I will loose all mails from day one :-(
And I do not like to have my workstation running all the time... I plan to
add fetchmail retrieval on my router/firewall.
/RogerL
--
Roger Larsson
Skellefteå
Sweden
More information about the Kde-optimize
mailing list