On Tue, Nov 4, 2008 at 5:37 PM, Dominique Pelle wrote: > > 2008/11/4 Matt Wozniski: >> >> ... My WAG is that they're seeing a >> font rendering regression (Pango? Freetype?), since I can reproduce >> this problem in Debian Unstable with GTK2-GNOME gvim (105 seconds for >> Dominique's benchmark) but not with the X11-Motif gvim (14 seconds for >> Dominique's benchmark) or xterm/screen (13 seconds). In GTK2 gvim, I >> used "DejaVu Sans Mono 7", in the others I use >> "-misc-fixed-medium-r-semicondensed--13-120-*-*-c-60-iso10646-1". >> >> Yep - just tested, and sure enough... switching my xterm font to >> DejaVu size 8 brings the Dominique's benchmark up to 94 seconds from >> 13. > > Maybe the font explains it, but I tried the test again with several > fonts and gvim always performed the test in 12s or 13s seconds > on my laptop. I tried it by adding the line: > > set gfn=Courier\ 10
Slow for me with Courier 10 and Courier New 10 as well. Seems to be a recent regression. I wonder which package, though... Kazuo Teramoto said that he could reproduce it on Archlinux, and since Arch and Debian Sid are two of the biggest rolling-release distros, it makes me think that this is an upstream regression in one of the libraries, and not a Debian/Ubuntu specific regression. If someone with Gentoo could do an emerge world, it would probably help to confirm or refute that theory... ~Matt --~--~---------~--~----~------------~-------~--~----~ You received this message from the "vim_use" maillist. For more information, visit http://www.vim.org/maillist.php -~----------~----~----~----~------~----~------~--~---
