]>
git.imager.perl.org - imager.git/log
Tony Cook [Sat, 26 May 2012 02:33:49 +0000 (12:33 +1000)]
adjust the test error messages to match the new module load errors
Tony Cook [Sat, 26 May 2012 02:32:53 +0000 (12:32 +1000)]
[rt #77063] improve rotate(degrees => ...) results.
Tony Cook [Thu, 24 May 2012 09:09:41 +0000 (19:09 +1000)]
[rt #77173] improve error reporting on an API version mismatch
Tony Cook [Fri, 11 May 2012 12:28:50 +0000 (22:28 +1000)]
elimintate IIM_new and IIM_DESTROY from Imager's internals
They only matter for XS
Tony Cook [Fri, 11 May 2012 11:52:32 +0000 (21:52 +1000)]
yes 0.90 has been released
Tony Cook [Mon, 30 Apr 2012 09:02:56 +0000 (19:02 +1000)]
0.90 release
Tony Cook [Mon, 30 Apr 2012 09:00:04 +0000 (19:00 +1000)]
update PNG Changes/.pm for release
Tony Cook [Sun, 29 Apr 2012 04:54:18 +0000 (14:54 +1000)]
Merge branch 'png'
Tony Cook [Sun, 29 Apr 2012 04:53:35 +0000 (14:53 +1000)]
PNG re-work: prepare for merge
Tony Cook [Sun, 29 Apr 2012 03:11:32 +0000 (13:11 +1000)]
PNG re-work: finish up the tag documentation
Tony Cook [Mon, 23 Apr 2012 05:18:08 +0000 (15:18 +1000)]
PNG re-work: start documenting new tags
Tony Cook [Fri, 20 Apr 2012 10:15:25 +0000 (20:15 +1000)]
PNG re-work: saving imager tags as PNG chunks
Tony Cook [Fri, 20 Apr 2012 10:14:51 +0000 (20:14 +1000)]
PNG re-work: fix line length calculations for pngdump.pl
Tony Cook [Mon, 16 Apr 2012 09:05:53 +0000 (19:05 +1000)]
PNG re-work: process the bKGD chunk into an i_background tag
Tony Cook [Sun, 15 Apr 2012 04:50:52 +0000 (14:50 +1000)]
PNG re-work: test tags from PNG reads and re-work a little
libpng synthesizes gAMA and cHRM chunks on input when sRGB is present,
so ignore them if sRGB is present.
No longer generate a png_text tag when the keyword is recognized.
Added png_time
Tony Cook [Sun, 15 Apr 2012 04:48:07 +0000 (14:48 +1000)]
PNG re-work: ignore some build and test detritus for the PNG manifest
Tony Cook [Sun, 15 Apr 2012 04:47:14 +0000 (14:47 +1000)]
PNG re-work: update pngdump.pl to dump IDAT data
Tony Cook [Sun, 15 Apr 2012 04:46:49 +0000 (14:46 +1000)]
PNG re-work: avoid the comment in qw() warning
Tony Cook [Sun, 15 Apr 2012 03:26:13 +0000 (13:26 +1000)]
PNG re-work: update Changes for the most recent changes
Tony Cook [Sun, 15 Apr 2012 03:20:19 +0000 (13:20 +1000)]
PNG-rework: forward png flush output through to i_io_flush()
Tony Cook [Sat, 14 Apr 2012 02:52:22 +0000 (12:52 +1000)]
PNG re-work: support writing 16-bit/sample images
Tony Cook [Thu, 12 Apr 2012 12:55:45 +0000 (22:55 +1000)]
PNG re-work: handle writing bilevel files
Tony Cook [Thu, 12 Apr 2012 09:55:56 +0000 (19:55 +1000)]
PNG re-work: paletted file writes
Tony Cook [Tue, 10 Apr 2012 08:32:59 +0000 (18:32 +1000)]
PNG re-work: split out write handling
Tony Cook [Mon, 9 Apr 2012 11:03:42 +0000 (21:03 +1000)]
PNG re-work: collect more text tags
Tony Cook [Mon, 9 Apr 2012 11:03:22 +0000 (21:03 +1000)]
PNG re-work: remove unused variables
Tony Cook [Mon, 9 Apr 2012 11:02:39 +0000 (21:02 +1000)]
PNG re-work: test results of new and old read handlers
Tony Cook [Mon, 9 Apr 2012 04:59:49 +0000 (14:59 +1000)]
PNG re-work: expand the pngdump.pl tool a bit
Tony Cook [Mon, 9 Apr 2012 02:58:43 +0000 (12:58 +1000)]
PNG re-work: add more test images and fix the PNG/ MANIFEST
Tony Cook [Mon, 9 Apr 2012 02:49:11 +0000 (12:49 +1000)]
PNG re-work: change note for the previous + some TODO
Tony Cook [Fri, 6 Apr 2012 13:18:45 +0000 (23:18 +1000)]
PNG re-work: read 16-bit PNG as 16-bit
Tony Cook [Fri, 6 Apr 2012 07:30:39 +0000 (17:30 +1000)]
PNG re-work: read 1 bit grey images as the common internal bilevel format
which is really paletted
Tony Cook [Fri, 6 Apr 2012 05:13:20 +0000 (15:13 +1000)]
PNG re-work: implement paletted read
Tony Cook [Mon, 2 Apr 2012 11:53:16 +0000 (21:53 +1000)]
PNG re-work: remove obsolete comment and the associated unused code
This appears to have been copied from libpng's example.c
Tony Cook [Mon, 2 Apr 2012 11:49:37 +0000 (21:49 +1000)]
PNG re-work: remove direct image access
the PNG driver has long directly accessed Imager idata data
directly, which is invalid.
I may change the internal format of the data in the future for simpler
MMX/SSE implementation.
Tony Cook [Mon, 2 Apr 2012 11:32:03 +0000 (21:32 +1000)]
PNG re-work: simple structural dump script for PNG files
Tony Cook [Sat, 12 Nov 2011 05:38:41 +0000 (16:38 +1100)]
WIP PNG re-work
Tony Cook [Mon, 7 Nov 2011 09:11:18 +0000 (20:11 +1100)]
PNG re-work: save png_bits and png_interlace tags
Tony Cook [Wed, 2 Nov 2011 11:58:53 +0000 (22:58 +1100)]
PNG re-work: add more image type TODO tests
Tony Cook [Tue, 14 Jun 2011 13:49:51 +0000 (23:49 +1000)]
PNG re-work: simple tests for reading various images types
including TODOs for paletted checks
Tony Cook [Tue, 14 Jun 2011 13:34:47 +0000 (23:34 +1000)]
PNG re-work: test size limit handling of very wide or tall PNG images
Tony Cook [Tue, 14 Jun 2011 13:34:11 +0000 (23:34 +1000)]
PNG re-work: provide perl level access to the libpng version number
Tony Cook [Sat, 11 Jun 2011 03:55:07 +0000 (13:55 +1000)]
PNG re-work: improve error reporting for PNG read/write
Tony Cook [Sun, 15 Apr 2012 04:47:40 +0000 (14:47 +1000)]
ignore some build detritus for the manifest
Tony Cook [Mon, 16 Apr 2012 09:04:59 +0000 (19:04 +1000)]
i_get_file_background[f]?() now return int
Tony Cook [Sat, 28 Apr 2012 02:48:49 +0000 (12:48 +1000)]
fix a typo
Tony Cook [Sat, 28 Apr 2012 02:45:13 +0000 (12:45 +1000)]
[rt #74043] handle static linking of libpng with libz
Tony Cook [Fri, 27 Apr 2012 12:01:43 +0000 (22:01 +1000)]
Tony Cook [Fri, 27 Apr 2012 11:56:10 +0000 (21:56 +1000)]
[rt #75869] make sure EU::MM can find libraries found via pkg-config
Tony Cook [Fri, 27 Apr 2012 10:18:35 +0000 (20:18 +1000)]
more verbose probe output
Tony Cook [Tue, 14 Jun 2011 13:32:59 +0000 (23:32 +1000)]
PNG re-work: handle libpng 1.5 correctly
Tony Cook [Wed, 25 Apr 2012 02:18:10 +0000 (12:18 +1000)]
[rt #76782] improve documentation and error reporting for missing I/O callbacks
Tony Cook [Tue, 24 Apr 2012 09:50:29 +0000 (19:50 +1000)]
change notes for the BMP updates
Tony Cook [Tue, 24 Apr 2012 09:23:21 +0000 (19:23 +1000)]
odd run lengths for BI_RLE4 compressed BMP files were handled incorrectly
Tony Cook [Tue, 24 Apr 2012 07:12:46 +0000 (17:12 +1000)]
handle odd-width BI_RLE4 width overflow
Some BMP BI_RLE4 writers compress scanlines as if they were an extra
pixel in width, eg. GIMP.
Imager would fail to read these images, I now allow one pixel of
overflow.
Tony Cook [Mon, 23 Apr 2012 12:18:36 +0000 (22:18 +1000)]
report basic information to the log for one scanline overflow
Tony Cook [Mon, 23 Apr 2012 12:17:17 +0000 (22:17 +1000)]
unshipped regression tests for BMP files
Tony Cook [Mon, 23 Apr 2012 12:16:01 +0000 (22:16 +1000)]
start of a BMP file structure dumper
Tony Cook [Mon, 23 Apr 2012 12:05:32 +0000 (22:05 +1000)]
[rt #76736] fix BI_BITFIELDS handling
I didn't have any test images when I wrote the BI_BITFIELDS code,
unfortunately it was completely broken.
Updated and tested (with some unshipped images for now.)
Tony Cook [Mon, 9 Apr 2012 11:02:58 +0000 (21:02 +1000)]
document the i_comment common tag
Tony Cook [Sat, 14 Apr 2012 04:06:07 +0000 (14:06 +1000)]
re-work sample down-conversion to rounding
Tony Cook [Sat, 14 Apr 2012 03:07:08 +0000 (13:07 +1000)]
change note for
360e674ef3d
Tony Cook [Sat, 14 Apr 2012 02:51:47 +0000 (12:51 +1000)]
fix the prototype for Imager::Test::is_imaged
Tony Cook [Tue, 20 Mar 2012 09:56:13 +0000 (20:56 +1100)]
fix the brokeness introduced in
5ee60746
as modified it only compared the first byte, change it to expand into
an array ref and use is_deeply() is compare.
Tony Cook [Sun, 18 Mar 2012 01:31:12 +0000 (12:31 +1100)]
0.89 release
Tony Cook [Sat, 10 Mar 2012 03:13:26 +0000 (14:13 +1100)]
[rt #70388] debug mymalloc() no longer builds a string, just stores the pointer
The debug malloc() code would use snprintf() (or even worse,
sprintf()) to build a string to track when the memory allocation was
performed for later tracking.
Now we just store the pointers, which should come from __FILE__ and
hence be static.
Tony Cook [Fri, 9 Mar 2012 08:15:32 +0000 (19:15 +1100)]
[rt #75560] don't fallback to using DynaLoader when XSLoader fails
This was masking errors when the loadable object failed to load for
some reason other than XSLoader being unavailable.
The alternative was to only eval the load of XSLoader and keep the
fallback, but XSLoader has been core perl for a while.
Tony Cook [Mon, 5 Mar 2012 10:34:15 +0000 (21:34 +1100)]
[rt #71495] method index links now point at the method documentation
Tony Cook [Mon, 5 Mar 2012 09:54:37 +0000 (20:54 +1100)]
[rt #75258] fix the utterly broken i_gpixf() for paletted images
This is implemented as a wrapper around the underlying i_gpix()
implementation and was handling the return value from i_gpix()
incorrectly.
Add tests and fix it.
Tony Cook [Sat, 25 Feb 2012 03:40:47 +0000 (14:40 +1100)]
compare unpacked data to avoid unintelligble cpan testers output
Tony Cook [Wed, 22 Feb 2012 05:08:28 +0000 (16:08 +1100)]
Imager 0.88 release
Tony Cook [Sat, 18 Feb 2012 08:53:13 +0000 (19:53 +1100)]
[rt #73697] re-work getpixel() and setpixel()
Tony Cook [Fri, 17 Feb 2012 12:27:31 +0000 (23:27 +1100)]
split Imager's typemap into internal, public and old perl bugfixes
Split as follows:
typemap.local - handle types specific to Imager.xs
typemap - types that Imager's API publishes
typemap.oldperl - fix broken typemap entries in old perl
typemap.oldperl only fixes issues I've run into with perl.
Tony Cook [Thu, 16 Feb 2012 13:00:28 +0000 (00:00 +1100)]
Merge branch 'psamp'
Tony Cook [Thu, 16 Feb 2012 12:58:07 +0000 (23:58 +1100)]
Changes note for the psamp/setsamples work
Tony Cook [Thu, 16 Feb 2012 07:41:26 +0000 (18:41 +1100)]
document setsamples() new capabilities
Tony Cook [Tue, 14 Feb 2012 11:52:56 +0000 (22:52 +1100)]
add offset and width parameters to setsamples() for new psamp/psampf
Tony Cook [Tue, 14 Feb 2012 09:31:43 +0000 (20:31 +1100)]
we don't use D::C's pod coverage, don't collect it
Tony Cook [Mon, 13 Feb 2012 09:57:59 +0000 (20:57 +1100)]
OO API
Tony Cook [Sun, 12 Feb 2012 00:59:59 +0000 (11:59 +1100)]
don't supply a default for a missing channel list
The underlying XS will supply a NULL chan_list which uses a faster
code path i_gsamp/i_psamp().
Tony Cook [Sat, 11 Feb 2012 13:33:33 +0000 (00:33 +1100)]
i_psamp()/i_psampf() coverage tests for paletted images
Tony Cook [Sat, 11 Feb 2012 01:52:45 +0000 (12:52 +1100)]
i_psamp()/i_psampf() coverage tests for masked images
Tony Cook [Fri, 10 Feb 2012 09:30:47 +0000 (20:30 +1100)]
i_psamp()/i_psampf() bad channel count tests
Tony Cook [Fri, 10 Feb 2012 09:30:14 +0000 (20:30 +1100)]
i_psamp()/i_psampf() coverage tests for double/sample images
Tony Cook [Fri, 10 Feb 2012 08:49:17 +0000 (19:49 +1100)]
i_psamp()/i_psampf() tests for 16-bit/sample images
Tony Cook [Fri, 3 Feb 2012 13:05:21 +0000 (00:05 +1100)]
coverage tests for 8-bits/sample psamp/psampf
Tony Cook [Fri, 3 Feb 2012 13:05:02 +0000 (00:05 +1100)]
allow imcover.perl to specify a subset of tests to run
and:
- allow for more verbose output
- disable duplicate gcov runs
Tony Cook [Mon, 30 Jan 2012 12:06:49 +0000 (23:06 +1100)]
initial XS interfaces to i_psamp()/i_psampf()
Tony Cook [Mon, 30 Jan 2012 12:04:27 +0000 (23:04 +1100)]
add psamp(), psampf() to C image API
Tony Cook [Mon, 30 Jan 2012 09:42:50 +0000 (20:42 +1100)]
re-work XS handling of channel lists
Tony Cook [Sun, 12 Feb 2012 00:40:30 +0000 (11:40 +1100)]
change note for getsamples(target) fix
Tony Cook [Sun, 12 Feb 2012 00:35:09 +0000 (11:35 +1100)]
[rt #74882] fix getsamples(target => \@foo, ...)
This was broken except for the i_gsamp_bits() case and was treating
the target parameter as a hashref instead of an array reference.
Tony Cook [Thu, 26 Jan 2012 03:35:33 +0000 (14:35 +1100)]
[rt #69244] remove long unused gif case from read
Tony Cook [Mon, 23 Jan 2012 07:55:31 +0000 (18:55 +1100)]
[rt #70014] re-work win32 text output to work in normal mode
Tony Cook [Mon, 23 Jan 2012 07:29:16 +0000 (18:29 +1100)]
ignore more generated files
Tony Cook [Fri, 20 Jan 2012 11:41:46 +0000 (22:41 +1100)]
note Debian inconsistency
Tony Cook [Fri, 20 Jan 2012 11:41:17 +0000 (22:41 +1100)]
add to the valid links list so the previous commit passes the link check
Tony Cook [Fri, 20 Jan 2012 11:29:35 +0000 (22:29 +1100)]
[rt #73906] add some other image processing modules to SEE ALSO
Tony Cook [Fri, 20 Jan 2012 11:16:47 +0000 (22:16 +1100)]
[rt #69261] fix incorrect image size and color calculations for rotate()
and matrix_transform().
For rotate() the calculation used integer abs() instead of fabs() when
calculating the size of the new image.
This made the image a little too small, hiding a bug at the corners
where a very low coverage pixel would reveal another bug where bad
rounding of the output level could result in all samples being maxed
out (8-bit images only.)
Tony Cook [Thu, 5 Jan 2012 23:39:39 +0000 (10:39 +1100)]
rotate.c -> rotate.im transition
convert to use the preprocessor to reduce code duplication