Date: Thu, 3 Oct 1996 10:41:27 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: rb@gid.co.uk (Bob Bishop) Cc: terry@lambert.org, jmacd@cs.berkeley.edu, freebsd-hackers@FreeBSD.org Subject: Re: XDrawString Message-ID: <199610031741.KAA06338@phaeton.artisoft.com> In-Reply-To: <v01510107ae791ce00fd1@[194.32.164.2]> from "Bob Bishop" at Oct 3, 96 08:38:49 am
next in thread | previous in thread | raw e-mail | index | archive | help
> >> >My results showed that XDrawString on a 2048 character string, > >> >where only 50 or so are on-screen takes much longer than the same > >> >call with only the visible length supplied. > >>[etc] > > > >I personally expect the Xlib to imply a cliprect on the data it sends > >to the X server.[etc] > > If *someone* isn't clipping, then things is surely broke. Trouble is, when > there is clipping the algorithm seems to go exponential if one end of the > draw is in the window and the other is far away. I guess it depends on whether you think that the most likely event will be: 1) Tiny bit of what is drawn is off screen 2) Most of what is drawn is off screen Case (1) would make you want to take the hit of drawing a small amount of non-visible objects that get clipped by the server as a trade for putting a clip region comparison in the inner loop. Case (2) would make you want to take the hit of putting a clip region comparison in the inner loop as a trade for draw a large amount of non-visible objects that get clipped by the server. I would think it would depend on whether you expected the programmer using the system to be careful (case 1) or sloppy (case 2). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610031741.KAA06338>