e081ae2d99
experiment with app14 decode
2019-04-15 10:50:42 +01:00
37613c196d
fix jpeg xmp read/write
...
it was not removing/reattaching the magic XMP URL
see https://github.com/libvips/libvips/issues/1229
2019-02-11 18:16:20 +00:00
cff4af34ca
add vips_image_set_blob_copy()
...
and use it where we can
2018-11-23 16:27:39 +00:00
0b4cef06b7
close early for tiffload
2018-08-16 16:00:14 +01:00
b836749b75
close jpeg read early
...
The current behaviour (close input handles on unref) works for languages
like C / C++ / Python / Rust / etc. where things get unreffed automatically
when they go out of scope.
On languages like Ruby / C# / node / etc. where things are unreffed on GC,
files can stay open for a long time after you've finished with them. This
interacts in an unfortunate way with the Windows default of refusing to
remove open files.
This change closes file handles as soon as the scan of the input file
finishes, and therefore produces something closer to expected behaviour
for GCd languages on Windows.
see https://github.com/kleisauke/net-vips/issues/12
2018-08-16 15:47:48 +01:00
4f22e8d1dc
add jpeg-chroma-subsample to jpegload
...
The jpeg loader now sets the field jpeg-chroma-subsample to record
chroma subsample.
See https://github.com/jcupitt/libvips/issues/954
2018-04-25 16:04:20 +01:00
5b55dd8ed4
Merge branch '8.6'
2018-04-11 09:38:59 +01:00
24b146790f
oop reorder crop to come after cache
...
or we'll write beyond the buffer end
2018-04-10 15:39:51 +01:00
d1dd41a21f
strict round down on jpeg shrink
...
libjpeg rounds up on shrink-on-load. In some cases this can leave a dark
line along the right and bottom edge, since it only contains (for
example) 1/4 of a pixel of data.
This change adds a crop after jpeg load so that only complete pixels are
output.
See https://github.com/lovell/sharp/issues/1185
2018-04-10 15:18:18 +01:00
15abbbe986
set "interlaced" for jpg and png
...
set "interlaced" for interlaced png and jpg images
2018-04-09 17:14:38 +01:00
5176b4a17e
better header sniffing for small files
...
vips__get_bytes() used to fail if the file was too small for the
allocated buffer, which was a problem for svg, since files can be extremely
small.
This change makes vips__get_bytes() return the number of bytes read, so
the is_a testers can work on files smaller than the max header size.
2018-03-22 12:08:39 +00:00
6f2861df5b
move META_SEQ support into foreign load
...
simplifies loaders a bit
2018-01-01 11:14:27 +00:00
83e36880da
more compat for older IPCT code
...
add IPTC data under the name "ipct-data" as well
2017-12-31 10:38:35 +00:00
77287a938f
fix misspelling of IPTC as IPCT
...
the letters have been accidentally swapped for years
add a compat macro so older code still works
2017-12-31 10:23:27 +00:00
9c8790b40a
setting EXIF data blocks sets other tags too
...
calling vips_image_set() to set the EXIF data block VIPS_META_EXIF_NAME
will automaticaly set other tags, like orientation etc.
2017-11-27 15:15:21 +00:00
2b70348d23
Merge branch '8.5'
2017-08-29 10:33:45 +01:00
e2c83fe4bd
fix jpegload fail
...
we were only failing on libjpeg errors, but libjpeg treats very serious
things thaht we want to be able to catch (like truncated files) as warnings
... we must therefore error out if fail is set and libjpeg issues a warning
see https://github.com/lovell/sharp/issues/793
2017-08-29 09:49:38 +01:00
10003f3f3c
revise libjpeg new from buffer
...
in line with latest libjpeg recommendations
this fixes a segv with corrupt input found by libFuzzer and asan
2017-08-25 12:51:06 +01:00
0ef01e9ca5
fix jpegload from empty buffer
...
could try to read a couple of non-existent bytes before failing ... now we
inject an EOI marker
2017-08-24 17:28:07 +01:00
942b0446a3
revise fail handling, again
...
clarify policy on file read errors and warnings:
* if the file format library reports a warning, we log it in vips, but
that's all
* if the file format library reports an error, we log it, but try to
continue
* if the file format library reports an error and fail is set, we log it
and fail
all loaders now implement this
2017-05-12 17:22:49 +01:00
362c7efdfc
argh reapply seq removal
2017-03-06 11:55:38 +00:00
dd5a108854
fix up new seq mode stuff
...
seems to work
2017-03-05 21:59:53 +00:00
96b1aba1a7
Revert "experiment with removal of seq stall"
...
This reverts commit c6eb9ee0f2
2017-02-27 10:26:59 +00:00
e6b20cbd0f
Revert "remove stalling"
...
This reverts commit 959f412380
2017-02-27 10:26:32 +00:00
98f5421c10
Revert "remove readbehind"
...
This reverts commit cb83918a3f
2017-02-27 10:24:56 +00:00
cb83918a3f
remove readbehind
2017-02-21 17:14:33 +00:00
959f412380
remove stalling
2017-02-21 15:03:15 +00:00
c6eb9ee0f2
experiment with removal of seq stall
2017-02-21 09:44:21 +00:00
0231b512b4
fits isn't quite working still
2017-01-17 14:53:40 +00:00
c464503510
Merge branch 'master' into invalidate-on-load-error
2017-01-17 13:47:14 +00:00
5d61bd4982
Merge branch 'master' into add-webp-metadata
2017-01-04 16:13:39 +00:00
262d5a10f7
Merge branch 'master' into glib-logging
2017-01-04 10:13:06 +00:00
3a48ff9cc1
don't warn for missing exif res
...
don't warn for missing exif res (even though it should be there) since
we fall back to jfif
cs3 seems to avoid the exif res fields and just uses jfif
see https://github.com/lovell/sharp/issues/657
2017-01-04 09:22:32 +00:00
2be0b97dce
switch to g_warning()_/g_info()
...
we had vips_warn() and vips_info(), but they are a bit crappy ... switch
to g_warning() and g_info() instead
see https://github.com/jcupitt/libvips/issues/544
2017-01-03 15:52:27 +00:00
47222bc1f5
move exif handling out to exif.c
...
now used by webpload as well
2016-11-07 12:20:56 +00:00
c90b8be0b8
add tiffsave_buffer
...
add tests
also some cleanups to foreign include files
see https://github.com/jcupitt/libvips/issues/417
2016-10-15 12:29:14 +01:00
b43c171e79
set resunit as well when fixing exif res
2016-09-07 20:03:53 +01:00
af4282e0e2
Now assigning the jpeg resolution inside the exif header, if the exif header isnt set properly
2016-09-07 14:07:29 +02:00
4451f4d839
Don't use the exif resolution if not fully set
2016-09-07 12:56:58 +02:00
d8381c73da
more improvements to error handling during eval
...
we were dropping the error buffer in tilecache, thanks David
2016-07-28 09:46:10 +01:00
828b36dfe5
fix --fail option to jpegload
...
getting --fail on jpegload working means tilecache must terminate on tile
calc error
make openslideload not report tile calc errors, it might need a --fail
option too
see https://github.com/jcupitt/libvips/issues/474
2016-07-11 10:07:41 +01:00
916e5b5589
set filename on file read for more types
...
the new loaders were missing the line to set the output filename, so
`vipsheader fred.pdf` was not displaying the filename correctly
2016-06-01 09:14:01 +01:00
094470c9c1
done, just a few things left
...
- tif->jpg copy seems to loose tag
- add tests
2016-05-27 09:48:08 +01:00
1fbae62cf8
seems to work for jpg anyway
...
try some tests next
2016-05-05 20:07:17 +01:00
1f701a61ea
try signalling invalidate on load error
...
to knock load ops out of cache
2016-05-05 15:00:40 +01:00
01238a0fd5
more progressive jpg changes
...
- Remove the nocache stuff for prog jpg images. There was no saving in
practice, since operations downstream could be cached.
nocache is for objects that can change, not for memory saving
- Call _destroy on the decompress object much earlier, it frees a huge
amount of memoey for prog jpg images.
2016-02-21 10:57:36 +00:00
2b46b7e889
many more const decls
...
ruby gobject-introspection is quite fussy about needing a lot of const
declarations ... these changes help vips_image_matrix_from_array()
appear in Ruby
2015-10-23 11:45:11 +01:00
01025328da
fix some cppcheck warnings
...
try:
$ cppcheck -f --enable=warning,performance,portability -j4 . 2> err.txt
there are still some left, see:
see https://github.com/jcupitt/libvips/issues/331
2015-09-28 22:01:15 +01:00
c2e8406ca0
check for unknown exif tags
...
exif_tag_get_name() can return NULL for unknown tags, check
2015-07-23 15:01:44 +01:00
fbe321eebf
change exif names argh
...
change exif names again: we were storing under @title, but that's both
subject to i18n, and unlookupable in libexif
we now use @name, which is not subject to i18n and can be searched for
... this will break most code which expects certain exif tag names
also, when we update exif, allow any tag, not just updates to existing
tags, see:
https://github.com/lovell/sharp/issues/189
2015-07-15 13:27:47 +01:00
fc5a4a917a
make file format readers close fds early
...
we were using two fds per open image, now we just use one
jpg, magick, tiff, png
2015-02-26 14:09:01 +00:00
6ff1d5320b
don't call jpeg_finish_decompress()
...
all it does is verify the trailing bytes in the file
see https://github.com/lovell/sharp/issues/151
2015-01-20 16:55:35 +00:00
ad86f45cd9
add autorot as an operator as well
2014-10-19 10:35:10 +01:00
0a51536162
seems to work
2014-10-17 13:17:44 +01:00
14d7ce1e91
first try, not very good
2014-10-16 19:16:45 +01:00
158237254d
fix jfif resunit read
...
see https://github.com/jcupitt/libvips/issues/170
the previous fix just did exif resolution "none", we need jfif "none" as
well
2014-09-09 16:46:16 +01:00
6269c852cc
support exif res unit "none"
...
well why would you want that, but it seems some images use it
2014-09-09 12:18:21 +01:00
e9ce212b62
make out of order read a fatal error
...
For png and jpg, out of order must always be a fatal error, since you
can't read from a specific y position.
This was a g_assert() before which meant that this test was missing from
production builds. It's now always a fatal error.
2014-09-02 09:37:08 +01:00
c85dd26627
oops, better filename set on foreign load
...
we were setting the filename in the wrong place
2014-06-10 13:46:05 +01:00
9dea549063
oops savers were not setting filename
...
when we added all the buffer loaders, we cul dno longer set filename in
VipsForeign, it's now the responsibility of each load class
2014-06-10 11:26:09 +01:00
d0f50febbb
added vips_foreign_load_buffer()
2014-04-22 16:27:43 +01:00
499b977043
back out the complex linear stuff
...
needs cooking for a while longer in a branch
2014-03-04 14:32:43 +00:00
4d85011a51
disable sequential skipahead
...
this was causing some problems on very-many-core systems
still need to update operation flags and disable seq mode for
insert/extract/etc.
2014-02-25 11:12:41 +00:00
0ef0328396
oops
2014-02-24 21:58:07 +00:00
d843521f77
jpeg load from buffer shouldn't modify buffer
...
fill_input_buffer() could write to the input buffer
see: https://github.com/jcupitt/libvips/pull/107
thanks Lovell
2014-02-24 17:10:07 +00:00
ae6ea3ba87
support many Radiance readers active at once
...
we had some globals for the Rad reader state, whcih was fine with the
old reader, but with the new sequential one you can have several active
at once
2014-01-23 15:06:37 +00:00
12454d3225
more jpeg read sanity checking
2014-01-23 14:19:55 +00:00
63591f6ac1
FATSTRIP for jpeg read
2013-12-16 16:22:44 +00:00
eb3ad6d64f
add lots of instrumentation
...
also svg output
2013-11-26 14:57:56 +00:00
12cf71a6a9
combin copy_fields and demand_hint
...
We had two API calls, vips_image_copy_fields() and vips_demand_hint().
They are now combined in the single vips_image_pipeline() call.
All operations are now slightly smaller and simpler.
2013-10-22 09:29:40 +01:00
d89007a8d4
catch jpeg startup errors
...
thanks mvgulik
2013-10-15 13:47:15 +01:00
6c0eced22c
build fixes for OS X
...
stupid libjpeg and their stupid bool type argh
2013-10-04 11:54:01 +01:00
440b800e0e
add a new unbuffered seq mode
...
things like
vips copy huge.png x.v
now run with only a small line cache
2013-09-29 11:04:11 +01:00
6167d4d97c
Fix three minor memleaks
2013-07-15 22:01:00 +01:00
02095763c9
null-terminate exif strings
...
libexif strings are not null-terminated, add a \0 before adding to the
vips image
thanks Mike
2013-07-06 11:51:34 +01:00
0c6da0c330
fix demand hints for shrink
...
"vips shrink" was broken due to wonky demand hints, thanks Jan
2013-04-08 13:08:24 +01:00
8cc76a6ba3
Update address for Free Software Foundation
...
Also update LGPLv2.1 license text from the copy currently on the GNU
website.
2013-03-07 00:40:19 -05:00
20ce700e23
jpeg tag read/write cleanups
2012-11-22 09:15:02 +00:00
c3ef2be4ab
preserve jpeg ipct data
...
now copies over ipct data from app13, thanks Gary
2012-11-21 21:33:50 +00:00
65b4056c84
jpeg read is less fussy about exif data
...
even read truncated exif
2012-11-21 18:34:20 +00:00
7025fe4081
more exif read improvements
...
some jpegs use a signed long to store the resolution unit (??), remove
the checks on tag format for res unit
2012-11-16 19:53:59 +00:00
74f545fc6f
keep exif rationals as a/b
...
we went via double before, which caused awful problems trying to
generate nice rationals again on save
keep everything as a/b as long as we can
2012-11-16 13:50:49 +00:00
fcc33020bd
better exif handling
...
- track ifd numbers for each item of exif
- be more careful about which ones we update
- don't update xres/yres in vips_shrink()
2012-11-16 10:03:31 +00:00
2048ec2356
jpeg read oops
2012-11-05 14:41:31 +00:00
8e0de67f65
sync
2012-10-30 17:16:55 +00:00
89065b391a
sequential needs a strip size param
...
strip tiff can only read in strip-sized chunks
2012-08-24 16:06:29 +01:00
39b05f5172
tuning, sort out demand hints
...
and disable extract_area seq mode
2012-08-24 12:46:10 +01:00
8653b23810
better handling of jpeg warnings
2012-08-23 21:44:14 +01:00
f5f0dda551
experiment with thread stalling
2012-08-21 16:18:25 +01:00
463058c149
note and use res unit on jpg load/save
...
previously, jpeg save always used pixels/inch and jpeg load converted to
vips pixels/mm
now on jpg load the image's res unit is recorded in
VIPS_META_RESOLUTION_UNIT, and on jpg save the res unit is set from
VIPS_META_RESOLUTION_UNIT (or defaults to inches).
you can now copy a cm-preferring tiff to a jpg and the unit is preserved
2012-08-07 14:10:19 +01:00
3694a2edc5
add sequential mode to vips7
...
the tiff and jpg readers let you give :seq as well
2012-06-24 09:44:50 +01:00
2f897315ad
fix a crash with corrupted jpg files
...
three was a double-free crash with a scrambled jpg file, thanks Grigoriy
2012-06-19 08:52:06 +01:00
8c6e3738dd
fix up ordering issues in seq mode
...
read order should be guaranteed now, hopefully
2012-02-20 15:03:29 +00:00
8af5522a96
sequential support for the jpg reader
2012-02-19 22:17:20 +00:00
bbadb8d681
jpeg2vips now keeps a load struct about
...
so we are ready to add lazy read
2012-02-19 10:54:41 +00:00
774e2fe495
read jfif resolution
...
read jfif, if available, before we read exif resolution
2012-01-09 10:34:47 +00:00
0983b50905
Merge remote-tracking branch 'origin/gobject-introspection' into goi-merge
...
Conflicts:
TODO
libvips/arithmetic/im_recomb.c
libvips/arithmetic/im_stats.c
libvips/deprecated/im_measure.c
libvips/include/vips/header.h
libvips/include/vips/image.h
libvips/include/vips/object.h
libvips/iofuncs/header.c
libvips/iofuncs/image.c
libvips/iofuncs/object.c
2011-12-24 17:36:55 +00:00
a62f5f8832
remove wio_output
2011-12-21 12:08:29 +00:00
05fbb6b3bc
fix up openslideload
...
there was a problem with the tile cache ... plus some small tidies
2011-12-12 11:34:28 +00:00
935c713e71
sort out jpegload
2011-11-29 12:11:33 +00:00
John Cupitt
Ulysse Manceron
Benjamin Gilbert