Rockbox mail archiveSubject: Re: FS#8961 - Anti-Aliased Fonts
Re: FS#8961 - Anti-Aliased Fonts
From: Frank Gevaerts <frank_at_gevaerts.be>
Date: Fri, 25 Feb 2011 17:25:15 +0100
On Fri, Feb 25, 2011 at 05:04:06PM +0100, Thomas Martitz wrote:
> I'm using FS#8961 and I really like it. So I would like to commit it. I
> reviewed the patch and I think it's good enough.
I haven't looked at the actual code, but I've done some measurements.
According to test_gfx, lcd_putsxy() is 1.3% slower on ipod video for
mono fonts. Presumably this goes down when using larger fonts than
sysfont (due to the slowdown being entirely caused by the one single
if() that decides between mono and AA rendering). Note also that this
is *pure* lcd_putsxy() performance, which isn't likely to be an actual
bottleneck for anything.
The binsize impact is on the order of one kilobyte. This should be
qualified by adding that the impact is only there on colour targets,
which all have at least 8MB of RAM, except for the c200v2. It can be
argued that the c200v2 doesn't have a good enough screen to benefit from
AA anyway, so it can be disabled there if needed, but that might of
course break theme compatibility with the v1.
I think both of those costs are small enough to be considered insignificant.
This means that practically speaking the patch has no impact for people
who do not want the added feature (there's not even a new setting), and
if it's technically OK I see no reason to delay committing.
-- "Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." - Brian W. KernighanReceived on 2011-02-25