036bd0bb4b
convsep calls convasep when it can
2016-07-09 19:20:33 +01:00
f294ec5d9b
oops
2016-07-09 19:10:17 +01:00
4c1e6dd023
turn off chroma subsample for Q > 90
...
see https://github.com/jcupitt/libvips/issues/482
2016-07-09 18:49:22 +01:00
e30bdadfb8
notes
2016-07-09 18:45:03 +01:00
c43757d8da
add restrict on unpack loop
2016-07-09 18:24:15 +01:00
ce0bddaf4f
update notes
2016-07-09 18:22:37 +01:00
9d7667cbff
oop found a problem
2016-07-09 18:01:46 +01:00
a9d1c654e2
fixes
2016-07-09 14:15:28 +01:00
c0bffd53ad
use PingImage() etc
...
much faster now we have a proper header for libMagick
2016-07-09 13:12:31 +01:00
43395d4e7b
tests pass
2016-07-09 12:19:56 +01:00
5fae93bc61
faster and simpler magickload
...
but tests are failing now
2016-07-08 22:38:14 +01:00
a5d9cf204c
seems to work, needs more testing
2016-07-08 17:56:30 +01:00
de981cd9ec
docs tweak
2016-07-08 12:18:07 +01:00
40f8a8af85
add autoconf stuff
...
todo:
- copypaste magick2vips into magick7load
- move to magick7 API
- test with magick6 and GM too, of course
see https://github.com/jcupitt/libvips/issues/480
2016-07-08 10:45:29 +01:00
0f2d1a6b0b
improve docs
2016-07-08 09:15:56 +01:00
bfdafdb733
intize input mask
2016-07-07 14:45:36 +01:00
f873a51b09
sort-of working
2016-07-06 14:56:28 +01:00
e0cfcc5f47
convasep done
...
though not tried compiling yet
2016-07-06 10:02:53 +01:00
0aa453ee0a
start on convasep
2016-07-05 15:52:49 +01:00
3130b9b118
add note on Peter's CM work
2016-07-05 08:48:08 +01:00
65067fd58b
add some more auto init
...
vips_image_new_from_file() was missing a check_init
2016-07-04 14:51:00 +01:00
b81ac67a98
tag tiff alpha as UNASSALPHA
...
we were using ASSOCALPHA, but that's only for pre-multiplied alpha
channels
2016-07-04 14:49:04 +01:00
5e58187902
improve docs
2016-07-04 10:18:39 +01:00
85cbebbac4
remove old im_conv() code
2016-07-01 16:18:48 +01:00
9442c225ba
kinder, gentler underflow check
2016-07-01 15:30:07 +01:00
d2aeec56b5
convi done
2016-07-01 15:11:17 +01:00
df81a6fe24
almost there
2016-06-30 14:59:20 +01:00
e45c5a587e
int vector path done
...
seems to sort-of work
2016-06-29 17:49:02 +01:00
0c2259d34b
improve similarity docs
...
rotation direction was wrong, add type comments
https://github.com/jcupitt/libvips/issues/475
2016-06-28 17:07:14 +01:00
bd7bad1916
almost done vector path
2016-06-28 11:07:20 +01:00
5ab0001ec6
Add configure check for zlib with inflateInit2
...
and replace old unused and poorly named FIND_ZIP. The new code prefers
pkgconfig and only falls back to manual detection if needed.
This also prioritizes detected zlib flags and includes to be preferred
over the ones that are added by pkgconfig for libpng, tifflib etc. which
would otherwise have caused the system default zlib to be used.
2016-06-27 03:03:45 +02:00
14953c19e7
Add support for .svgz files to svgload
2016-06-27 01:54:53 +02:00
5e660addc8
convi working
...
need to add vector path next
2016-06-24 16:03:15 +01:00
1a5aa2125e
add convi
2016-06-24 12:48:08 +01:00
e60e90b0da
im_conv_f rewrite done
2016-06-23 20:42:03 +01:00
93a55310d3
compiles
2016-06-23 15:00:26 +01:00
65105a9442
upsize with something other than nearest
...
vips_resize() uses to just use nearest when upsizing, following standard
practice in repro. This is often unfortunate for image processing, where
small nearest upsizes will produce obvious aliasing.
It now picks a VipsInterpolate which corresponds (roughly) to the
selected VipsKernel and uses that with affine for any upsizing.
2016-06-22 10:33:08 +01:00
6f52f14fc5
fix comment
2016-06-19 09:37:26 +01:00
3c7ce5f6fe
better invalid test
2016-06-18 19:42:43 +01:00
11b136b2df
pyvips8 can create new metadata
...
previously it tried to lookup the type of the field in set(), now if
OK if there's no field there already
2016-06-17 11:27:13 +01:00
d56cd4dcb2
note py problem
2016-06-17 09:36:04 +01:00
f3326c8126
better vips_resize() for cubic/linear/nearest
...
vips_resize() used to do most of a downsize with vips_shrink() and the
final 200 - 300% with vips_reduce(). This was correct for lanczos2/3,
but not right for linear/cubic, which need more shrink and less
reduce to avoid aliasing.
This patch makes vips_resize() leave the final 100 - 200% to
vips_reduce() for linear/cubic, and leave everything to reduce for
nearest.
2016-06-16 09:46:02 +01:00
a5bef08d4a
better reducev multiplication
...
more accurate, no slower
add more tests too
2016-06-15 13:56:19 +01:00
58be330714
oop problems found in reducev
2016-06-14 14:56:56 +01:00
eb7768ff61
Merge branch 'master' into add-assignment-overloads
2016-06-11 20:15:42 +01:00
83b1c33441
add tests for C++ += etc
...
all seems to work!
2016-06-11 20:14:48 +01:00
28efdf1695
working!
2016-06-11 17:36:06 +01:00
8667dd0d18
VImage::ifthenelse() fix
...
the prototype was missing =0 on options for scalar constant forms
2016-06-10 17:34:41 +01:00
93951fd04a
start adding
2016-06-10 14:57:46 +01:00
6fa665d6de
notes
2016-06-10 14:45:33 +01:00
John Cupitt
Felix Bünemann