This commit is contained in:
John Cupitt 2012-01-26 14:42:20 +00:00
parent 21f9a7f69b
commit 2b5963090f

48
TODO
View File

@ -1,3 +1,51 @@
- im_fwfft() does not set interpreation to fourier
- seems to do a lot of looping?
$ vips --vips-progress im_fwfft Gugg_coloured.jpg x.v
vips fwfft1:1: 2 threads, 972 x 16 tiles, groups of 64 scanlines
vips temp-6: 2 threads, 972 x 16 tiles, groups of 64 scanlines
vips temp-6: done in 0s
vips fwfft1:1: done in 0s
vips fft: 2 threads, 128 x 128 tiles, groups of 256 scanlines
vips fft: done in 0s
vips fwfft1:1: 2 threads, 972 x 16 tiles, groups of 64 scanlines
vips fwfft1:1: done in 0s
vips fft: 2 threads, 128 x 128 tiles, groups of 256 scanlines
vips fft: done in 0s
vips fwfft1:1: 2 threads, 972 x 16 tiles, groups of 64 scanlines
vips fwfft1:1: done in 0s
vips fft: 2 threads, 128 x 128 tiles, groups of 256 scanlines
vips fft: done in 0s
vips im__fftproc: 2 threads, 128 x 128 tiles, groups of 256 scanlines
vips im__fftproc: done in 0s
vips x.v: 2 threads, 128 x 128 tiles, groups of 256 scanlines
vips x.v: done in 0s
that's 8 separate loops over the image
- viewing an image tagged as fourier in nip2 is broken, which operation is
failing?
- im_invfftr() is broken
$ vips im_fwfft Gugg_coloured.jpg x.v
$ vips --vips-progress im_invfftr x.v x2.v
vips invfft1-1: 2 threads, 972 x 1 tiles, groups of 4 scanlines
vips invfft1-1: done in 0s
Segmentation fault
add a test to the suite
- we have VipsArrayObject and also vips_object_local_array() - we have VipsArrayObject and also vips_object_local_array()
can we make one use the other? can we make one use the other?