[Nautilus-list] Re: AA rectangles in non-aa mode + speedups
- From: Darin Adler <darin bentspoon com>
- To: Alex Larsson <alexl redhat com>, Nautilus <nautilus-list lists eazel com>
- Subject: [Nautilus-list] Re: AA rectangles in non-aa mode + speedups
- Date: Sat, 26 Jan 2002 20:01:41 -0800
Overall, this looks great, and you have my OK to check in. But please read
these comments first.
I am surprised that you define, but don't use, RENDER_LIBS. I assume that
we're implicitly relying on gtk to drag in the Xrender and Xext if they are
present, but if so, then we shouldn't even define RENDER_LIBS. Also, you set
have_render to false, but then use have_xft. Because of that mistake, I bet
it won't work right on systems without Xrender.
You should kill that FIXME in start_rubberbanding. All you need to do is set
errno to 0 ahead of time, and then if it's non-zero afterward, make
fill_color be the default color.
I still don't understand exactly why the old behavior of re-raising the
rubberband was a problem. Isn't gnome_canvas_item_raise_to_top a no-op if
the item is already on top? If not, isn't the simplest fix to check for that
case in GnomeCanvas rather than adding the logic here? I agree for the case
where icons are raised, it's nice to raise them to the exact right level,
rather than raising them to the top and then putting the selection rectangle
back on top, but it doesn't seem like that would normally happen during
rubberbanding. What do you say?
-- Darin
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]