Luis Ibanez [Wed, 26 May 2010 23:44:07 +0000 (23:44 +0000)]
ENH: Added custom configuration for CTest. This allows to fine tune
the selection of files for code coverage computation, and also
allows to filter out acceptable compiler warnings.
MACOSX : Fixed the XCode project file to link with the PNG library. Library and header should be located in /usr/local/lib/ and /usr/local/include/, respectively. Check http://ethan.tira-thompson.org/Mac_OS_X_Ports.html if it is not the case on your mac.
[v1.4]Fixed issue 6 on google code. Variable "pi" not freed or pointed-to in function "pi_create_encode". Thanks to Kent Mein for reporting this.
[v1.4]Fixed problem with Borland C++ Builder (Borland C do not have lrintf). Thanks Marek Mauder for this fix.
Added support for PNG image format [Not yet functional under WIN32]. Thanks to Winfried for this patch. See details here http://code.google.com/p/openjpeg/issues/detail?id=16
[FOD] Created a new constant in openjpeg.h to differentiate the case when the colorspace is not supported by the library and when it is not specified in the codestream. Suggested by Matteo Italia.
openjpeg.c : fixed initialization of parameters->cp_disto_alloc/fixed_quality/fixed_alloc/tcp_numlayers because they were preventing to use the -q and -f options.
Antonin Descampe [Fri, 22 Aug 2008 16:18:17 +0000 (16:18 +0000)]
found a bug in tcd.c that was preventing to find the correct threshold in tcd_rateallocate.c for high-precision images. Applied a temporary patch but a better solution should be found.
Antonin Descampe [Thu, 21 Aug 2008 16:00:44 +0000 (16:00 +0000)]
fixed a bug in image_to_j2k.c that was preventing the 'r' option to work properly (everything was compressed lossless regardless of the specified rate).
Initial commit of openjpeg version 2. Temprarily added as a separate directory in the trunk (will maybe be moved to a branch of version 1.3, this has to be investigated).
Fixed openjpeg.c for proper initialization of codec context structures (dinfo in opj_create_compress() and opj_create_decompress(). Bug fix suggested by Andrey V. Kiselev
Clean up of opj_aligned_malloc(), to just forgo the use of posix_memalign(), as apparently memalign() is what is working better for everyone. Patch by Callum.
Giuseppe Baruffa [Mon, 11 Feb 2008 14:43:11 +0000 (14:43 +0000)]
Minor style modifications to comply with c99 compiler flag; removed Microsoft-specific "union-in-struct" syntax; Re-enabled cstr_info struct creation when -W switch is specified. Changed a number of things in opjviewer (e.g., decoding thread does not execute GUI calls anymore), to have it running under linux --> it is better than before, but still crashes
In image.c, changed the opj_image_create0() memory allocation from malloc() to calloc() in order to avoid segfaults when freeing the memory allocated for the coding of bad images. Thanks to Christopher Layne for this improvement.
First import of JAVAOpenJPEG, a Java wrapper of OpenJPEG, developed by Patrick Piscaglia of Telemis (www.telemis.com). Thank you Patrick for this new module !
Created the file index.c in the codec directory. This file handles the creation of index files, at encoding and decoding.
Fixed bugs during the creation of the index (PCRL progression order)
Patch by Callum Lerwick. The opj_tcd_cblk array is one of the largest allocations, because it contains a bunch of static buffers. This also makes it a major source of cache thrashing. This patch allocates the buffers from the heap, and dynamically sizes them in the decoder. I have not yet managed to dynamically size them in the encoder, getting the decoder to do it was tricky enough... I also split opj_tcd_cblk_t into separate encode and decode versions. A lot of fields were not used by both, so this cuts its size even further.
Patch by Callum Lerwick. This patch rearranges the largest memory allocations so they're allocated as late as possible, and freed as soon as possible. This cuts memory usage by about half on two large test images.
Patch by Dzonatas and Callum Lerwick. Fp/vectorization patch which basically converts most of the irreversible decode codepath to floating point, eliminating a few rounds of int/fp conversion, resulting in a vast performance improvement, and an increase in accuracy.
Fixed a bug which prevented JPWL from working on multi-tiled images; added some more fields in the interface info structures (keep a list of markers, save start packet number for each tile)
Changed the way the image structure is allocated when the decoding parameters include some resolutions to discard. This should have a significant impact for the decoding of huge images when some resolutions are discarder (-r parameter)
Warning: The output image size is now reduced when discarding resolutions !
BUG: Fix previous patch from Callum Lerwick. I have no clue what CMAKE_INSTALL_LIBDIR refers too. Bump up cmake 2.2 restriction to cmake 2.4 because of previous patch not backward compatible. Properly set the SOVERSION in a cross plateform way (yes WIN32 is a plateform)
Fixed issues with Reading and Writing TIF images in convert.c to avoid segmentation fault.
Fixed issues relating to using user specified rates for CINEMA option for multiple images.
OpenJPEG library interface modified to retain compatibility with version 1.2. Sorry if some of you already adapted their code to the previous interface, but we want to avoid a ABI break....
Patch from Callum Lerwick. Fixed the library linking so it links with the soversion properly set. Fixes up the install targets so that it interacts properly with RPM. Install target for MJ2. Sets some necessary and useful CFLAGS if gcc is in use.
Giuseppe Baruffa [Tue, 11 Sep 2007 15:21:12 +0000 (15:21 +0000)]
JPWL encoding is finalized correctly into the JP2 file format; added an additional structure in opj_codestream_info, to keep a record of the written markers
projects / openjpeg / log