source: trunk/libjpeg/change.log @ 470

Last change on this file since 470 was 470, checked in by Silvan Scherrer, 11 years ago

poppler, jpeg, freetype lib updates

File size: 13.0 KB
Line 
1CHANGE LOG for Independent JPEG Group's JPEG software
2
3
4Version 8b  16-May-2010
5-----------------------
6
7Repair problem in new memory source manager with corrupt JPEG data.
8Thank to Ted Campbell and Samuel Chun for the report.
9
10Repair problem in Makefile.am test target.
11Thank to anonymous user for the report.
12
13Support MinGW installation with automatic configure.
14Thank to Volker Grabsch for the suggestion.
15
16
17Version 8a  28-Feb-2010
18-----------------------
19
20Writing tables-only datastreams via jpeg_write_tables works again.
21
22Support 32-bit BMPs (RGB image with Alpha channel) for read in cjpeg.
23Thank to Brett Blackham for the suggestion.
24
25Improve accuracy in floating point IDCT calculation.
26Thank to Robert Hooke for the hint.
27
28
29Version 8  10-Jan-2010
30----------------------
31
32jpegtran now supports the same -scale option as djpeg for "lossless" resize.
33An implementation of the JPEG SmartScale extension is required for this
34feature.  A (draft) specification of the JPEG SmartScale extension is
35available as a contributed document at ITU and ISO.  Revision 2 or later
36of the document is required (latest document version is Revision 3).
37The SmartScale extension will enable more features beside lossless resize
38in future implementations, as described in the document (new compression
39options).
40
41Add sanity check in BMP reader module to avoid cjpeg crash for empty input
42image (thank to Isaev Ildar of ISP RAS, Moscow, RU for reporting this error).
43
44Add data source and destination managers for read from and write to
45memory buffers.  New API functions jpeg_mem_src and jpeg_mem_dest.
46Thank to Roberto Boni from Italy for the suggestion.
47
48
49Version 7  27-Jun-2009
50----------------------
51
52New scaled DCTs implemented.
53djpeg now supports scalings N/8 with all N from 1 to 16.
54cjpeg now supports scalings 8/N with all N from 1 to 16.
55Scaled DCTs with size larger than 8 are now also used for resolving the
56common 2x2 chroma subsampling case without additional spatial resampling.
57Separate spatial resampling for those kind of files is now only necessary
58for N>8 scaling cases.
59Furthermore, separate scaled DCT functions are provided for direct resolving
60of the common asymmetric subsampling cases (2x1 and 1x2) without additional
61spatial resampling.
62
63cjpeg -quality option has been extended for support of separate quality
64settings for luminance and chrominance (or in general, for every provided
65quantization table slot).
66New API function jpeg_default_qtables() and q_scale_factor array in library.
67
68Added -nosmooth option to cjpeg, complementary to djpeg.
69New variable "do_fancy_downsampling" in library, complement to fancy
70upsampling.  Fancy upsampling now uses direct DCT scaling with sizes
71larger than 8.  The old method is not reversible and has been removed.
72
73Support arithmetic entropy encoding and decoding.
74Added files jaricom.c, jcarith.c, jdarith.c.
75
76Straighten the file structure:
77Removed files jidctred.c, jcphuff.c, jchuff.h, jdphuff.c, jdhuff.h.
78
79jpegtran has a new "lossless" cropping feature.
80
81Implement -perfect option in jpegtran, new API function
82jtransform_perfect_transform() in transupp. (DP 204_perfect.dpatch)
83
84Better error messages for jpegtran fopen failure.
85(DP 203_jpegtran_errmsg.dpatch)
86
87Fix byte order issue with 16bit PPM/PGM files in rdppm.c/wrppm.c:
88according to Netpbm, the de facto standard implementation of the PNM formats,
89the most significant byte is first. (DP 203_rdppm.dpatch)
90
91Add -raw option to rdjpgcom not to mangle the output.
92(DP 205_rdjpgcom_raw.dpatch)
93
94Make rdjpgcom locale aware. (DP 201_rdjpgcom_locale.dpatch)
95
96Add extern "C" to jpeglib.h.
97This avoids the need to put extern "C" { ... } around #include "jpeglib.h"
98in your C++ application.  Defining the symbol DONT_USE_EXTERN_C in the
99configuration prevents this. (DP 202_jpeglib.h_c++.dpatch)
100
101
102Version 6b  27-Mar-1998
103-----------------------
104
105jpegtran has new features for lossless image transformations (rotation
106and flipping) as well as "lossless" reduction to grayscale.
107
108jpegtran now copies comments by default; it has a -copy switch to enable
109copying all APPn blocks as well, or to suppress comments.  (Formerly it
110always suppressed comments and APPn blocks.)  jpegtran now also preserves
111JFIF version and resolution information.
112
113New decompressor library feature: COM and APPn markers found in the input
114file can be saved in memory for later use by the application.  (Before,
115you had to code this up yourself with a custom marker processor.)
116
117There is an unused field "void * client_data" now in compress and decompress
118parameter structs; this may be useful in some applications.
119
120JFIF version number information is now saved by the decoder and accepted by
121the encoder.  jpegtran uses this to copy the source file's version number,
122to ensure "jpegtran -copy all" won't create bogus files that contain JFXX
123extensions but claim to be version 1.01.  Applications that generate their
124own JFXX extension markers also (finally) have a supported way to cause the
125encoder to emit JFIF version number 1.02.
126
127djpeg's trace mode reports JFIF 1.02 thumbnail images as such, rather
128than as unknown APP0 markers.
129
130In -verbose mode, djpeg and rdjpgcom will try to print the contents of
131APP12 markers as text.  Some digital cameras store useful text information
132in APP12 markers.
133
134Handling of truncated data streams is more robust: blocks beyond the one in
135which the error occurs will be output as uniform gray, or left unchanged
136if decoding a progressive JPEG.  The appearance no longer depends on the
137Huffman tables being used.
138
139Huffman tables are checked for validity much more carefully than before.
140
141To avoid the Unisys LZW patent, djpeg's GIF output capability has been
142changed to produce "uncompressed GIFs", and cjpeg's GIF input capability
143has been removed altogether.  We're not happy about it either, but there
144seems to be no good alternative.
145
146The configure script now supports building libjpeg as a shared library
147on many flavors of Unix (all the ones that GNU libtool knows how to
148build shared libraries for).  Use "./configure --enable-shared" to
149try this out.
150
151New jconfig file and makefiles for Microsoft Visual C++ and Developer Studio.
152Also, a jconfig file and a build script for Metrowerks CodeWarrior
153on Apple Macintosh.  makefile.dj has been updated for DJGPP v2, and there
154are miscellaneous other minor improvements in the makefiles.
155
156jmemmac.c now knows how to create temporary files following Mac System 7
157conventions.
158
159djpeg's -map switch is now able to read raw-format PPM files reliably.
160
161cjpeg -progressive -restart no longer generates any unnecessary DRI markers.
162
163Multiple calls to jpeg_simple_progression for a single JPEG object
164no longer leak memory.
165
166
167Version 6a  7-Feb-96
168--------------------
169
170Library initialization sequence modified to detect version mismatches
171and struct field packing mismatches between library and calling application.
172This change requires applications to be recompiled, but does not require
173any application source code change.
174
175All routine declarations changed to the style "GLOBAL(type) name ...",
176that is, GLOBAL, LOCAL, METHODDEF, EXTERN are now macros taking the
177routine's return type as an argument.  This makes it possible to add
178Microsoft-style linkage keywords to all the routines by changing just
179these macros.  Note that any application code that was using these macros
180will have to be changed.
181
182DCT coefficient quantization tables are now stored in normal array order
183rather than zigzag order.  Application code that calls jpeg_add_quant_table,
184or otherwise manipulates quantization tables directly, will need to be
185changed.  If you need to make such code work with either older or newer
186versions of the library, a test like "#if JPEG_LIB_VERSION >= 61" is
187recommended.
188
189djpeg's trace capability now dumps DQT tables in natural order, not zigzag
190order.  This allows the trace output to be made into a "-qtables" file
191more easily.
192
193New system-dependent memory manager module for use on Apple Macintosh.
194
195Fix bug in cjpeg's -smooth option: last one or two scanlines would be
196duplicates of the prior line unless the image height mod 16 was 1 or 2.
197
198Repair minor problems in VMS, BCC, MC6 makefiles.
199
200New configure script based on latest GNU Autoconf.
201
202Correct the list of include files needed by MetroWerks C for ccommand().
203
204Numerous small documentation updates.
205
206
207Version 6  2-Aug-95
208-------------------
209
210Progressive JPEG support: library can read and write full progressive JPEG
211files.  A "buffered image" mode supports incremental decoding for on-the-fly
212display of progressive images.  Simply recompiling an existing IJG-v5-based
213decoder with v6 should allow it to read progressive files, though of course
214without any special progressive display.
215
216New "jpegtran" application performs lossless transcoding between different
217JPEG formats; primarily, it can be used to convert baseline to progressive
218JPEG and vice versa.  In support of jpegtran, the library now allows lossless
219reading and writing of JPEG files as DCT coefficient arrays.  This ability
220may be of use in other applications.
221
222Notes for programmers:
223* We changed jpeg_start_decompress() to be able to suspend; this makes all
224decoding modes available to suspending-input applications.  However,
225existing applications that use suspending input will need to be changed
226to check the return value from jpeg_start_decompress().  You don't need to
227do anything if you don't use a suspending data source.
228* We changed the interface to the virtual array routines: access_virt_array
229routines now take a count of the number of rows to access this time.  The
230last parameter to request_virt_array routines is now interpreted as the
231maximum number of rows that may be accessed at once, but not necessarily
232the height of every access.
233
234
235Version 5b  15-Mar-95
236---------------------
237
238Correct bugs with grayscale images having v_samp_factor > 1.
239
240jpeg_write_raw_data() now supports output suspension.
241
242Correct bugs in "configure" script for case of compiling in
243a directory other than the one containing the source files.
244
245Repair bug in jquant1.c: sometimes didn't use as many colors as it could.
246
247Borland C makefile and jconfig file work under either MS-DOS or OS/2.
248
249Miscellaneous improvements to documentation.
250
251
252Version 5a  7-Dec-94
253--------------------
254
255Changed color conversion roundoff behavior so that grayscale values are
256represented exactly.  (This causes test image files to change.)
257
258Make ordered dither use 16x16 instead of 4x4 pattern for a small quality
259improvement.
260
261New configure script based on latest GNU Autoconf.
262Fix configure script to handle CFLAGS correctly.
263Rename *.auto files to *.cfg, so that configure script still works if
264file names have been truncated for DOS.
265
266Fix bug in rdbmp.c: didn't allow for extra data between header and image.
267
268Modify rdppm.c/wrppm.c to handle 2-byte raw PPM/PGM formats for 12-bit data.
269
270Fix several bugs in rdrle.c.
271
272NEED_SHORT_EXTERNAL_NAMES option was broken.
273
274Revise jerror.h/jerror.c for more flexibility in message table.
275
276Repair oversight in jmemname.c NO_MKTEMP case: file could be there
277but unreadable.
278
279
280Version 5  24-Sep-94
281--------------------
282
283Version 5 represents a nearly complete redesign and rewrite of the IJG
284software.  Major user-visible changes include:
285  * Automatic configuration simplifies installation for most Unix systems.
286  * A range of speed vs. image quality tradeoffs are supported.
287    This includes resizing of an image during decompression: scaling down
288    by a factor of 1/2, 1/4, or 1/8 is handled very efficiently.
289  * New programs rdjpgcom and wrjpgcom allow insertion and extraction
290    of text comments in a JPEG file.
291
292The application programmer's interface to the library has changed completely.
293Notable improvements include:
294  * We have eliminated the use of callback routines for handling the
295    uncompressed image data.  The application now sees the library as a
296    set of routines that it calls to read or write image data on a
297    scanline-by-scanline basis.
298  * The application image data is represented in a conventional interleaved-
299    pixel format, rather than as a separate array for each color channel.
300    This can save a copying step in many programs.
301  * The handling of compressed data has been cleaned up: the application can
302    supply routines to source or sink the compressed data.  It is possible to
303    suspend processing on source/sink buffer overrun, although this is not
304    supported in all operating modes.
305  * All static state has been eliminated from the library, so that multiple
306    instances of compression or decompression can be active concurrently.
307  * JPEG abbreviated datastream formats are supported, ie, quantization and
308    Huffman tables can be stored separately from the image data.
309  * And not only that, but the documentation of the library has improved
310    considerably!
311
312
313The last widely used release before the version 5 rewrite was version 4A of
31418-Feb-93.  Change logs before that point have been discarded, since they
315are not of much interest after the rewrite.
Note: See TracBrowser for help on using the repository browser.