Commit Graph

1762 Commits

Author SHA1 Message Date
John Cupitt 5d19bcf955 Merge branch '8.8' 2019-08-30 11:03:24 +01:00
John Cupitt 6ea76f9632 improve data_length handling in jpg load
libjpeg uses unsigned ints for data length, so we must use size_t
everywhere.
2019-08-30 11:00:25 +01:00
John Cupitt 349e76bb85 Merge branch '8.8' 2019-08-29 15:30:20 +01:00
John Cupitt 3161de3b52 verify bands/format for coded vips images
WHen loading a vips image with Coding set, make sure that Bands and
BandFmt are correct.
2019-08-29 15:28:55 +01:00
John Cupitt 9a5dca1ef8 Merge branch '8.8' 2019-08-28 10:05:12 +01:00
John Cupitt 46212e92b1 fix default sharpen
sharpen with sigma 0.5 was doing nothing, since the new int precision
rules meant that we generated a point-point convolution.

This patch increases the int precision for sharpen, so we now work with
sigma down to 0.5.

Also: restore input colourspace. Previously, the output image was always
LabS (the computation space for sharpen). Now, it transforms back to the
input space. This is more in line with how other operators work.

Thanks 2h4dl for pointing this out.

See https://github.com/libvips/pyvips/issues/123
2019-08-28 09:16:40 +01:00
John Cupitt 2f6cc1fad9 Merge branch '8.8' 2019-08-27 14:49:42 +01:00
John Cupitt d61a85f6c9 prevent over-pre-shrink in thumbnail
We could pre-shrink so much that an axis went to 0.

See https://github.com/lovell/sharp/issues/1782#issuecomment-525249430
2019-08-27 14:47:36 +01:00
John Cupitt d2ab993931 Merge branch '8.8' 2019-08-27 13:06:27 +01:00
John Cupitt 5cce83a294 check image bounds for GIF load
It seems giflib does no checking of image dimensions at all, not even
height == 0.
2019-08-27 13:04:28 +01:00
John Cupitt 23d0d5603d Merge branch '8.8' 2019-08-24 17:15:45 +01:00
John Cupitt 4691260540 better feof() handling in gif load
Relaxing the read error rules made looping possible .. make sure we
always stop explicitly on eof.
2019-08-24 17:14:10 +01:00
John Cupitt 77de39320a Merge branch '8.8' 2019-08-24 11:21:48 +01:00
John Cupitt 038409093f clip coding and interpretation on image read
They could be out of bounds enums. This used not to matter, but we use
them more now, so out of bounds values can cause coredumps.
2019-08-24 11:20:45 +01:00
John Cupitt 1b57c6d066 Merge branch '8.8' 2019-08-23 15:43:12 +01:00
John Cupitt 4b677fb16a make GIF parse less strict
Some GIFs don't follow the standard very closely. Let them through.

See https://github.com/libvips/libvips/issues/1404
2019-08-23 15:40:48 +01:00
John Cupitt 19a80e4efb Merge branch '8.8' 2019-08-21 17:37:16 +01:00
John Cupitt b5e8e99746 fix a read-one-byte-beyond issue in jpeg load
libvips could harmlessly read beyond the end of a string with a crafted jpg
file
2019-08-21 17:17:54 +01:00
John Cupitt ed2054dbbc revise arithmetic with const
there's a problem with out of bounds values, for example:

	vips relational_const k2.jpg x.v equal 1000

actually finds pixels == 255, since 1000 is saturated converted to 255
before the test starts.

This patch reworks arithmetic against const values to fix this.
2019-08-21 10:35:48 +01:00
John Cupitt 2d0c21279a add tests 2019-08-19 16:27:44 +01:00
John Cupitt d80ce4bf15 fix a problem with shrinkv tail processing
Tail processing in shrinkv had an implicit assumption of round-down, but of
course we round to nearest. Thanks angelmixu.

see https://github.com/libvips/libvips/issues/1396
2019-08-13 11:21:01 +01:00
John Cupitt cdee0feb93 Merge branch '8.8' 2019-08-09 10:12:55 +01:00
John Cupitt aafc2c7948 add changelog note for profiles.c fix 2019-08-09 10:11:45 +01:00
John Cupitt 988f9594f4 Merge branch '8.8' 2019-08-08 16:11:02 +01:00
John Cupitt ff58c67e33 add locks for pdfium load
We used to lock within documents, ie. we did not allow two threads to
work on the same file.

However pdfium is not threadsafe in any way, and this is not supported,
see:

	https://groups.google.com/forum/#!msg/pdfium/kyIdh_J4csg/K1LvfPiHDwAJ

This patch adds locks around pdfium calls.

see:

	https://github.com/libvips/libvips/issues/1380
	https://github.com/libvips/libvips/issues/1275
2019-08-08 15:54:16 +01:00
John Cupitt 7250b169d4 Merge branch '8.8' 2019-08-07 16:19:23 +01:00
John Cupitt f36927e78d fix build against graphicsmagick
- add tests for more features missing in GM
- simplify the IM/GM configure tests
- revise the test suite to pass with GM and IM
2019-08-07 16:17:27 +01:00
John Cupitt f0d1a5d82b note PNG comment fix 2019-08-02 10:11:16 +01:00
John Cupitt 332e97272a Merge branch '8.8' 2019-07-29 10:07:29 +01:00
John Cupitt 5e77ab948b fix loop with malformed ppm
skip_line in ppm parse was not testing for EOF, so it could get stuck
with malformed files

thanks Kyle-Kyle

see https://github.com/libvips/libvips/issues/1377
2019-07-29 10:05:55 +01:00
John Cupitt 6a75776272 Merge branch 'loader-minimise-experiment' 2019-07-28 17:23:01 +01:00
John Cupitt 0e5447e537 final cleanup 2019-07-28 17:15:54 +01:00
John Cupitt 264f41df1c Merge branch '8.8' 2019-07-24 16:07:39 +01:00
John Cupitt 3510e7abcf fetch HEIC metadata from the main image
since thumbs don't have it, see https://github.com/libvips/libvips/issues/1373
2019-07-24 16:06:49 +01:00
John Cupitt 7f3174c3e5 Merge branch 'master' into loader-minimise-experiment 2019-07-23 11:45:24 +01:00
John Cupitt 03bf5a05fa note fuzz in ChangeLog 2019-07-21 18:02:57 +01:00
John Cupitt 2c654060f9 experiment with a different early-close strategy
We close loaders early in order to save file handles, and on Windows to
make sure that files can be deleted as soon as possible.

Currently loaders do this by watching the Y coordinate of requests and
freeing the fd when the final line of the file is fetched. This is messy
and does not always work, since there are cases when the final line is
not fetched.

Instead, this patch gets loaders to listen for "minimise" on their
output and close on that. This signal is emitted on all upstream images
whenever a threadpool finishes a scan of an image and is usually used to
trim caches after computation.

See https://github.com/libvips/libvips/issues/1370
2019-07-20 16:31:30 +01:00
John Cupitt 64a6a27326 Merge branch '8.8' 2019-07-19 11:59:38 +01:00
John Cupitt 8cf6a9f9ca don't attempt to save large XMP to JPG
jpeg_write_marker() with some libjpeg versions will throw a fatal
error with large chunks.

To write >64kb XMP it you need to parse the whole XMP object,
pull out the most important fields, code just them into the main
XMP block, then write any remaining XMP objects into a set of
extended XMP markers.

http://wwwimages.adobe.com/content/dam/Adobe/en/devnet/xmp/pdfs/XMPSpecificationPart3.pdf

Instead, just warn and skip large XMP blocks.

see https://github.com/libvips/libvips/issues/1372
2019-07-19 11:37:36 +01:00
John Cupitt 044f8b74f6 Merge branch 'master' into add-tiff-webp-zstd 2019-07-11 13:50:47 +01:00
John Cupitt c3e7504835 Merge branch '8.8' 2019-07-10 17:44:39 +01:00
John Cupitt 5e2d66d14b better early shutdown behaviour in shrinkv
read the tail of the input to force early shutdown in seq readers

does reducev need something similar?

see https://github.com/kleisauke/net-vips/issues/12
2019-07-09 16:58:30 +01:00
John Cupitt 52ee3b083f add webp and zstd support to tiffsave
needs some tests still
2019-07-08 10:39:44 +01:00
John Cupitt c7f98ad0a5 Merge branch '8.8' 2019-07-07 13:12:51 +01:00
John Cupitt e4cff876a8 fix make dist
we'd renamed a file
2019-07-07 11:20:28 +01:00
John Cupitt f45f0ecef4 limit resize if shrink would drop axis <1px
vips_resize() will break the aspect ratio and limit resize on an axis if
it would result in an image of less than 1px on that axis

see https://github.com/lovell/sharp/issues/1782#issuecomment-508921306
2019-07-07 10:19:33 +01:00
John Cupitt 3871369d3e Merge branch 'master' into deftomat-master 2019-07-06 14:50:40 +01:00
John Cupitt 777c360774 safer gifload delay array init
- remove assumptions about n_pages
- minor reformatting
2019-07-06 11:34:20 +01:00
John Cupitt 80f247cfb2 Merge branch '8.8' 2019-07-03 15:33:43 +01:00
John Cupitt 8cb2b613b7 note new heifsave param in changelog
plus tiny reformatting
2019-07-03 15:04:58 +01:00