Hi Sven -

Code snippet:

 (while (< y height)

             (gimp-rect-select inImage 0 0 width y
                               REPLACE FALSE 0) ; select some pixels
             (gimp-edit-copy maskchannel)       ; copy the selection
             (gimp-selection-translate inImage 0 y) ; shift down
             (gimp-floating-sel-anchor
              (car
               (gimp-edit-paste maskchannel FALSE))) ; and paste back

             (set! y (+ y y))           ; move down by as much as we did
             )

The snippet above no longer has the anomaly at y = 256; instead it's around y = 150.  

The snippet below has the anomaly at y = 2 ^ n:


 (while (< y height)

             (gimp-rect-select inImage 0 0 width y
                               REPLACE FALSE 0) ; select some pixels
             (gimp-edit-copy maskchannel)       ; copy the selection
             (gimp-selection-translate inImage 0 y) ; shift down
             (gimp-floating-sel-anchor
              (car
               (gimp-edit-paste maskchannel TRUE))) ; and paste back

             (set! y (+ y y))           ; move down by as much as we did
             )

The location of the anomaly depends on the second argument to gimp-edit-paste, which I 
don't really understand anyway.

The problem does not arise when the image's y dimension is a power of 2.  (Or maybe 
it's some multiple of the step size:  the initial step I've been using has been 4).

To make life more interesting, I can't find it when the initial step size is 5.

I'll continue to research this and file a bug report when I can figure out what's 
really going on.

-- 
--Jeff

Jeff Trefftzs <[EMAIL PROTECTED]>
http://www.tcsn.net/trefftzs                    Home Page
http://gug.sunsite.dk/gallery.php?artist=68     Gimp Gallery
http://trefftzs.topcities.com/home.html         Photo galleries



_______________________________________________
Gimp-developer mailing list
[EMAIL PROTECTED]
http://lists.xcf.berkeley.edu/mailman/listinfo/gimp-developer

Reply via email to