Re: [Gmod-ajax] using the panel to do tiling

[Mitch S. <mit...@be...>]

I wrote:
> So far, I've been testing my changes by doing diffs of the tiles; I'm 
> pretty sure I've only committed changes that generate tiles that are 
> bit-exactly the same as before.  The tiles that I've generated with this 
> approach aren't the same bit-for-bit, but they do look the same (with 
> one exception: right ends of genes are now getting rendered correctly).  
> I think the difference is in the palette, so the tiles could still be 
> correct even if they're different.  So I'm not yet fully convinced that 
> it's rendering exactly correctly, but it does look right to me.
>   

I got the image diff program that the cairo project uses for its tests, 
which handles palette differences without a problem.  It also tells you 
how many pixels are different, which is a useful metric IMO.

I used it to compare the gbrowse-ajax-tiledimage version of tile 
generation with my experimental panel-based one, and after a few days of 
fiddling with the per-tile panel bounds, it's close.  It produces the 
same images pixel-for-pixel as TiledImage, with the exception of the 
2kbp and 1kbp zoom levels, where the gridlines are shifted by one 
pixel.  I spent some time trying to chase this down, without much 
success.  My best guess is that since 1kbp and 2kbp are right on a 
rounding boundary (1 and 0.5 bases per pixel, respectively) there's some 
kind of rounding difference.

At any rate, while the difference is annoying, I think it's close enough 
that I'm going to go ahead and commit this.  For the time being I plan 
to continue with this approach, unless something comes up that shows 
it's unworkable.

Mitch