On Tue, Dec 13, 2005 at 06:08:52PM +0100, Jean-Marc Lasgouttes wrote: > >>>>> "Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes: > > Martin> Ah yes of course. Is there a way to suppress this? What I see > Martin> is that always another call to redoParagraph seems to follow, > Martin> from BufferView::Pimpl::metrics. My feeling is that one of > Martin> them is superfluous. > > It is done in setCursor, maybe Alfredo knows why?
First off, I commented out the redoParagraph statement in setTargetX, and LyX started crashing on backspace. It seems to be needed... > Martin> No, that amazes me... because even in this case, I have > Martin> reduced the number of calls to redoParagraph by 20% at least, > Martin> even if you play the devil's advocate. Yet, Bennett sees no > Martin> more than half of that in speedup. I reduced the area drawn / > Martin> number of drawings by 50% at least, no matter how you count > Martin> it, and that seems to show nowhere. So where else is LyX > Martin> eating time if not in rebreaking and drawing? > > Remember that the total is always 100% whatever you do :) The change > you may see is that some part becomes less important than it was. Doesn't really apply here... my savings are based on operation counts from my instrumentation output (lyxerr stmts) and Bennett's from his stopwatch :-) > Martin> The width querying is mostly from redoParagraph too, I > Martin> suspect. > > It seems that the rows we build at this point should remember the > position of all text chunks in each row. This would lower a lot the > number of width computations. I seem to remember Andre having a patch. Did it do this? - Martin
pgpX0z3poy9B0.pgp
Description: PGP signature
