486a08d8a6
relive debug stuff
2016-05-23 16:49:47 +02:00
5678f93257
Add buffer-write support for Radiance
2016-05-23 16:37:28 +02:00
14d7a97afc
change write_new to make it independent from file I/O
2016-05-20 11:28:33 +02:00
4c599ea272
fix typo
2016-05-20 11:27:41 +02:00
23579d4928
delete obsolete line
2016-05-19 22:37:33 +02:00
10a3782590
Add buffer write support for HDR Radiance
2016-05-19 22:32:17 +02:00
fa79fcde35
more overcomp reporting
2016-05-19 14:40:16 +01:00
c84f642d09
removed intltool again
...
and ignire more standard m4/ macros
2016-05-18 16:24:06 +01:00
bf70a4d0bf
oops droppped aclocal somehow
2016-05-18 14:27:57 +01:00
9c9e97bd80
try to fix travis
2016-05-18 14:22:20 +01:00
6e94df7214
set version before import pyvips8
...
stops some warnings
2016-05-18 13:41:41 +01:00
10cfbc553f
get rid of m4-safe
...
and don't copy the .m4 files recommended by glib-gettextize either
2016-05-18 13:41:31 +01:00
95b9c2aeb6
argh
2016-05-18 10:54:40 +01:00
cc349198a1
oops
2016-05-18 10:49:28 +01:00
91aa288a44
break up acinclude.m4
...
it's now in a more maintainable set of .m4 files in m4-safe
2016-05-18 10:41:05 +01:00
1add582047
Merge pull request #448 from felixbuenemann/fix-python-detection
...
Fix Python and GI path detection problems
2016-05-18 09:45:11 +01:00
2b71756352
Merge branch '8.3'
2016-05-18 09:30:22 +01:00
7ec63c4451
more vips image read fixing
...
thanks AFL!
2016-05-18 09:00:30 +01:00
4285118880
more robust vips file read
2016-05-18 08:58:05 +01:00
c5dfc1f840
Fix VIPS_LIBDIR being "NONE" without --prefix
...
If the --prefix switch is not given to configure, the prefix variable is
NONE, so we need to the the default value from ac_default_prefix first.
2016-05-18 04:28:59 +02:00
dcbd8397a5
Warn if the gi module could not be loaded
2016-05-18 04:28:58 +02:00
e24334beb2
Disable vips7 SWIG bindings for python3
...
because they cause a compile error (see issue #334 ).
2016-05-18 04:28:58 +02:00
6956bcf8a4
Fix several problems with gi path warnings
...
* Do not warn when the prefix appears to be a homebrew cellar.
* Fix configure errors when python is not detected
* Fix configure errors when pygobject is installed in prefix but
not present in the python `sys.path`.
* Use the actual overrides path, instead of the gi module path.
2016-05-18 04:28:58 +02:00
f0a5ac5bbc
Vendor python.m4 to fix wrong multiarch exec paths
...
The python.m4 that ships with automake hardcodes the library path for
platfrom dependent files to `lib`, which is wrong for multiarch systems.
Instead we import the `m4/python.m4` from pygobject, which does not have
this problem and also ensures we have the same behavior as pygobject
which we are depending on for pyvips8.
Due to this change some minor changes to how the `bootstrap.sh` script
clean up the `m4` folder where needed.
The python search order has been modified to prefer the default python
over python2 over python3.
2016-05-18 04:28:58 +02:00
6a445ea9c8
more vips image read fixing
...
thanks AFL!
2016-05-17 21:58:52 +01:00
af34f8f3ef
more robust vips file read
2016-05-17 19:43:47 +01:00
92a6ff19a1
add some more webp tests
2016-05-17 14:08:46 +01:00
77c9616952
oh argh last change
2016-05-17 13:56:16 +01:00
9057328208
reword config messages
2016-05-17 13:55:37 +01:00
bbac46817c
Merge branch 'master' of github.com:jcupitt/libvips
2016-05-17 13:44:58 +01:00
8b592585d7
try to py install help
...
at the end of configure, vips now tries to suggest typelib and overrides
install hints
2016-05-17 10:52:11 +01:00
1654ac7e5b
fix a comment
2016-05-16 14:32:52 +01:00
54d8e4808f
restore bandfmt after unpremultiply in thumbnail
...
vips_unpremultiply() always outputs float. If you save the output of
vipsthumbnail to a format that supports float files, like tiff, it won't
ever get cast back to the source format.
Example:
$ vipsheader Opera-icon-high-res.png
Opera-icon-high-res.png: 3056x3325 uchar, 4 bands, srgb, pngload
$ vipsthumbnail Opera-icon-high-res.png -o x.tif
$ vipsheader x.tif
x.tif: 117x128 float, 4 bands, scrgb, tiffload
This change makes it note and restore BandFmt aropund pre/unpremultiply.
See https://github.com/jcupitt/libvips/issues/447
2016-05-16 13:18:59 +01:00
db11d3b973
allow nested [] in CLI args
...
this now works:
vips copy x y.dz[suffix=.jpg[Q=90]]
2016-05-16 11:09:47 +01:00
70abf4efde
merge pdfload range to master
2016-05-15 10:02:46 +01:00
ffc5279d2f
Merge branch 'master' into loadpdf-range
2016-05-15 10:01:27 +01:00
7f02e843dc
add ack for dzsave tile skip
...
see https://github.com/jcupitt/libvips/issues/352
2016-05-15 09:53:37 +01:00
46f2776d22
fix up changelog order
2016-05-15 09:50:40 +01:00
72b67d0ec2
Merge branch 'master' into no-blank-tiles
2016-05-15 09:48:21 +01:00
2cfe4842d3
notes
2016-05-15 09:48:02 +01:00
a613841a49
oop various fixes
...
seems to work now
2016-05-13 09:16:34 +01:00
24a51098b1
compiler warning
2016-05-12 14:54:28 +01:00
f18fbe1fa2
add @n option to pdfload
...
so you can load many pages at once as a tall strip
2016-05-12 14:13:50 +01:00
9fba3b5577
notes on TODO
2016-05-12 08:55:19 +01:00
10fcbb6129
Merge pull request #443 from felixbuenemann/remove-jpegsave-quant_table-debug-print
...
Remove jpegsave default quant_table debug print
2016-05-09 18:44:31 +01:00
cb97d3d786
Remove jpegsave default quant_table debug print
...
This was only ut in for debugging which quant_table is used in the
JCP_FASTEST profile and should've never been committed.
2016-05-09 18:37:00 +02:00
9db05c30c2
tiny README fixes
2016-05-09 14:06:44 +01:00
b206ff18cb
Merge pull request #442 from felixbuenemann/test-system-webp-on-travis
...
Test system webp on Travis CI and fix alpha_q default value
2016-05-09 13:07:53 +01:00
3b7242f743
Fix webpsave alpha_q default value
2016-05-09 13:08:11 +02:00
4640aab388
Test against system libwebp on Travis CI
...
* Dropped custom ppa used only for libwebp backport
* Dropped lossless test from `make check` (not supported in 0.1.3)
2016-05-09 12:29:23 +02:00
Henri Chain
John Cupitt
Felix Bünemann