|
c77783ed
|
2020-06-25T21:28:33
|
|
BUILDING.md: Undocument 32-bit Mac build/packaging
(oversight from previous commit)
|
|
|
b797f700
|
2020-06-25T19:05:45
|
|
Build: Eliminate Cygwin packaging support
We haven't provided official Cygwin builds since 1.4.x, since Cygwin
now supplies its own libjpeg-turbo packages (although they apparently
haven't been updated past 1.5.3.)
|
|
|
9a2cf323
|
2020-02-11T13:41:43
|
|
Build: Enable separate iOS pkg/DMG w/ sim support
Refer to #406
|
|
|
ada6ea51
|
2020-01-16T20:52:15
|
|
BUILDING.md: Update Visual C++ recommendations
|
|
|
cbf0fcc8
|
2019-11-05T11:56:06
|
|
i386 SSE2 Huffman: Fix pointer arithmetic issue
Splitting the pointer arithmetic in GET_SYM() into a separate add and
sub instruction was an attempt to work around an error ("invalid operand
type") that occurred when assembling the file with NASM. However, this
created a link error on macOS ("ld: illegal text-relocation to
'_jconst_huff_encode_one_block' in
simd/CMakeFiles/simd.dir/i386/jchuff-sse2.asm.o from
'_jsimd_huff_encode_one_block_sse2' in
simd/CMakeFiles/simd.dir/i386/jchuff-sse2.asm.o for architecture i386")
and also changed the alignment of the code in ways that might have
affected the previous benchmark results (which took a great deal of time
to obtain.) Ultimately, the path of least resistance is just to
require NASM 2.13 or later.
|
|
|
bbedb4b5
|
2019-11-05T15:43:21
|
|
Merge branch 'master' into dev
|
|
|
cf54623b
|
2019-11-05T12:21:25
|
|
Mac: Support hiding SIMD fct symbols w/ NASM 2.14+
(NASM 2.14+ now supports the private_extern section directive, which was
previously only available with YASM.)
|
|
|
2ae2bd66
|
2019-10-04T16:05:11
|
|
Remove support for Apple's Java implementation
(AKA "Java for OS X systems.") This implementation of Java 1.6 is long
obsolete and not supported on any version of macOS past High Sierra.
Oracle no longer provides a 32-bit JVM on macOS, so it is no longer
necessary to provide a 32-bit version of the TurboJPEG Java wrapper on
macOS.
|
|
|
b7523059
|
2019-10-04T15:53:30
|
|
Merge branch 'master' into dev
|
|
|
ec4cacc1
|
2019-10-02T16:35:17
|
|
BUILDING.md: Document need for CRB/PowerTools repo
NASM and YASM are located in this non-default repository on
(respectively) RHEL and CentOS.
|
|
|
c055c880
|
2019-05-09T20:36:51
|
|
Discontinue support for 32-bit iOS builds
|
|
|
5308c1a0
|
2019-01-22T12:11:03
|
|
BUILDING.md: Update/simplify Android build instr.
Use the android.toolchain.cmake toolchain file in the NDK (v13b or
later), since this toolchain file generally takes care of setting the
approprate compiler flags and dealing with the differences between
GCC and Clang. Our custom Android build procedure did not work with
Clang-based NDK toolchains, which meant that it could not be made to
work with NDK v18b or later.
Fixes #309
|
|
|
c1f07a9f
|
2018-09-21T15:12:38
|
|
BUILDING.md: Correct/update Java information
- CMake 3.10.x or later must be used with JDK 11, or an error
("regex not supported") will occur when CMake tries to parse the Java
version number.
- The JDK is no longer available at java.com.
|
|
|
574f3a77
|
2018-07-27T11:44:46
|
|
Clarify Android Windows build instructions
(must add .exe to CMAKE_C_COMPILER)
Addresses a concern raised in #245, #260
|
|
|
f4b8a5cf
|
2018-07-02T11:42:06
|
|
BUILDING.md: Correct iOS/Android examples
toolchain.cmake should be created in the build directory, so the scripts
need to cd to that directory before generating that file.
Closes #254
|
|
|
367a8386
|
2018-02-26T19:41:59
|
|
Make SIMD syms private for x86[-64]/Mach-O builds
... if building with YASM. NASM doesn't currently support the necessary
directives.
Closes #212
|
|
32eaeeac
|
2018-02-24T18:00:51
|
|
Fix BUILDING.md buglet ("--mfloat-abi=softfp")
(Introduced with the build system overhaul)
Closes #215
|
|
|
bd544e28
|
2017-12-16T09:34:28
|
|
Merge branch 'master' into dev
|
|
|
7d3feda7
|
2017-12-15T16:45:25
|
|
BUILDING.md: Adjust Android recipes for NDK r16+
NDK r16b moved some things around, so modify the Android build recipes
to take that into account while preserving compatibility with previous
NDK releases.
NOTE: the GCC 4.9 NDK toolchain is deprecated, so we will need to
develop new Android build recipes for libjpeg-turbo 1.6 that use the
Clang toolchain.
Closes #196
|
|
|
b9ab64d8
|
2017-05-11T21:02:29
|
|
TurboJPEG: Thread-safe error message retrieval
Introduce a new C API function (tjGetErrorStr2()) that can be used to
retrieve compression/decompression/transform error messages in a
thread-safe (i.e. instance-specific) manner. Retrieving error messages
from global functions is still thread-unsafe.
Addresses a concern expressed in #151.
|
|
|
2ac4e9d9
|
2017-06-26T21:58:32
|
|
Merge branch 'master' into dev
|
|
|
301ba4f3
|
2017-06-26T15:15:08
|
|
BUILDING.md: Include Android/x86 build recipes
Addresses a concern raised in #155.
|
|
|
2fb4d7e3
|
2016-12-11T22:38:30
|
|
BUILDING.md: Documentation buglet
`make cygwinpkg` was listed under "Mac".
|
|
|
c8358fcb
|
2016-12-08T14:43:59
|
|
Build: Various improvements to install/pkg system
- GNUInstallDirs: any directory variable can now reference any other
directory variable by including its name in angle brackets (<>).
- Changed the documentation of the directory variables in BUILDING.md
accordingly. This commit also includes some formatting tweaks to
that section (using boldface for directory names, as is our
convention.)
- Changed the package scripts such that they use
CMAKE_INSTALL_DATAROOTDIR rather than CMAKE_INSTALL_DATADIR.
- We no longer override the install dir. defaults on Windows unless
performing an official build. It may be useful, for instance, to
use the GNU defaults when installing into an MSYS environment.
|
|
|
b0fcd0cc
|
2016-12-07T19:14:20
|
|
Build: Minor tweaks to GNUInstallDirs defaults
It isn't actually necessary to specify `CMAKE_INSTALL_DEFAULT_MANDIR`
for our official build. Because `CMAKE_INSTALL_DEFAULT_DATAROOTDIR` is
blank for the official build, the default of "<DATAROOTDIR>/man" will
resolve to "man".
For the same reason, this commit changes the specification of
`CMAKE_INSTALL_DEFAULT_DOCDIR` and `CMAKE_INSTALL_DEFAULT_JAVADIR` in
the official build to be dependent on the data root directory (mainly to
make it obvious what we're doing.)
This commit also tweaks the example CMake command line in the directory
variable documentation so that it shows the correct location of the
CMake argument.
|
|
|
d681fa76
|
2016-12-07T10:54:54
|
|
Build: Set install dirs in a more GNU-friendly way
This builds upon the existing GNUInstallDirs module in CMake but adds
the following features to that module:
- The ability to override the defaults for each install directory
through a new set of variables (`CMAKE_INSTALL_DEFAULT_*DIR`).
Before operating system vendors began shipping libjpeg-turbo, it was
meant to be a run-time drop-in replacement for the system's
distribution of libjpeg, so it has traditionally installed itself
under /opt/libjpeg-turbo on Un*x systems by default. On Windows, it
has traditionally installed itself under %SystemDrive%\libjpeg-turbo*,
which is not uncommon behavior for open source libraries (open source
SDKs tend to install outside of the Program Files directory so as to
avoid spaces in the directory name.) At least in the case of Un*x,
the install directory behavior is based somewhat on the Solaris
standard, which requires all non-O/S packages to install their files
under /opt/{package_name}. I adopted that standard for VirtualGL and
TurboVNC while working at Sun, because it allowed those packages to be
located under the same directory on all platforms. I adopted it for
libjpeg-turbo because it ensured that our files would never conflict
with the system's version of libjpeg. Even though many Un*x
distributions ship libjpeg-turbo these days, not all of them ship the
TurboJPEG API library or the Java classes or even the latest version
of the libjpeg API library, so there are still many cases in which it
is desirable to install a separate version of libjpeg-turbo than the
one installed by the system. Furthermore, installing the files under
/opt mimics the directory structure of our official binary packages,
and it makes it very easy to uninstall libjpeg-turbo.
For these reasons, our build system needs to be able to use
non-GNU-compliant defaults for each install directory if
`CMAKE_INSTALL_PREFIX` is set to the default value.
- For each directory variable, the module now detects changes to
`CMAKE_INSTALL_PREFIX` and changes the directory variable accordingly,
if the variable has not been changed by the user.
This makes it easy to switch between our "official" directory
structure and the GNU-compliant directory structure "on the fly"
simply by changing `CMAKE_INSTALL_PREFIX`. Also, this new mechanism
eliminated the need for the crufty mechanism that previously did the
same thing just for the library directory variable.
How it should work:
- If a dir variable is unset, then the module will set an internal
property indicating that the dir variable was initialized to its
default value.
- If the dir variable ever diverges from its default value, then the
internal property is cleared, and it cannot be set again without
unsetting the dir variable.
- If the install prefix changes, and if the internal property
indicates that the dir variable is still set to its default value,
and if the dir variable's value is not being manually changed at the
same time that the install prefix is being changed, then the dir
variable's value is automatically changed to the new default value
for that variable (as determined by the new install prefix.)
- The directory variables are now always cached, regardless of whether
they were set on the command line or not. This ensures that they can
easily be examined and modified after being set, regardless of how they
were set.
This was made possible by the introduction of the aforementioned
`CMAKE_INSTALL_DEFAULT_*DIR` variables.
- Improved directory variable documentation (based on descriptions at
https://www.gnu.org/prep/standards/html_node/Directory-Variables.html)
- The module now allows "<DATAROOTDIR>" to be used as a placeholder in
relative directory variables.
It is replaced "on the fly" with the actual path of
`CMAKE_INSTALL_DATAROOTDIR`.
This should more closely mimic the behavior of the old autotools build
system while retaining our customizations to it, and it should retain
the behavior of the old CMake build system.
Closes #124
|
|
|
6abd3916
|
2016-11-15T08:47:43
|
|
Unified CMake-based build system
See #56 for discussion.
Fixes #21, Fixes #29, Fixes #37, Closes #56, Fixes #58, Closes #73
Obviates #82
See also:
https://sourceforge.net/p/libjpeg-turbo/feature-requests/5/
https://sourceforge.net/p/libjpeg-turbo/patches/5/
|
|
|
9df7ac2e
|
2016-11-18T08:31:49
|
|
BUILDING.md: NASM 2.10+/YASM 1.2.0+ always needed
... for all x86[-64] builds, because we now have both 64-bit and 32-bit
AVX2 SIMD extensions.
|
|
|
9fdb8f85
|
2016-11-22T09:33:19
|
|
Merge branch 'master' into dev
|
|
|
f9f0c75b
|
2016-11-20T20:59:55
|
|
BUILDING.md: Clarifications and wordsmithing
Updated out-of-date information, wordsmithed and clarified many
sections, and generally cleaned up the build recipes (including a
complete overhaul of the iOS recipes.)
|
|
|
ab664e35
|
2016-11-18T09:09:41
|
|
BUILDING.md/README.md: Increment libjpeg SO age
Documentation buglet. This should have been changed in
6ed4d9d11085acd04dc7f2f899848693976dc010 to reflect the addition of
libjpeg API functions in libjpeg-turbo 1.5.
|
|
|
a280fa63
|
2016-11-15T09:52:49
|
|
BUILDING.md: Don't use trailing spaces as line break
Makes it easier to maintain this file using editors that automatically
remove trailing spaces.
|
|
|
6c365686
|
2016-09-20T18:09:15
|
|
Merge branch 'master' into dev
|
|
|
e9d9c31f
|
2016-09-19T22:47:18
|
|
Build: Remove ARMv6 support from 'make iosdmg'
The last iDevice to require ARMv6 was the iPhone 3G, which required iOS
4.2.1 or older. Our binaries have always required iOS 4.3 or newer,
so I'm not sure if the ARMv6 fork of our binaries was ever useful to
begin with. In any case, if it ever was useful, it no longer is. Fat
binaries can still be generated with ARMv6 support by invoking
{build_directory}/pkgscripts/makemacpkg manually.
|
|
|
426d787c
|
2016-05-28T16:42:44
|
|
64-bit AVX2 impl. of RGB->YCC color conversion
|
|
|
1d50a8cd
|
2016-05-31T22:48:52
|
|
BUILDING.md: More NASM/YASM clarifications
28d1a1300c6be7fc8614ed827eb56cd97cf84e76 introduced the line
"nasm.exe should be in your PATH". This commit corrects an oversight in
8f1c0a681cd34e8e80ba7b06f356d6080a7172c9 /
e5091f2cf3b6ba747907012146df93df0d01ec85 whereby this line should have
been extended to include yasm.exe.
|
|
|
e5091f2c
|
2016-05-28T18:19:45
|
|
Merge branch '1.4.x'
|
|
|
0d61e80a
|
2016-05-01T12:07:05
|
|
Merge branch '1.4.x'
|
|
|
2628c562
|
2016-04-14T14:19:19
|
|
BUILDING.md: Fix "... OR ..." indentation again
<sigh> GitHub doesn't render indented text the same as my local MarkDown
viewer (MacDown), so it's necessary to indent "... OR ..." by 3 spaces
so both will display it on the same indentation level as "Visual C++
2005 or later" and "MinGW".
|
|
|
28d1a130
|
2016-04-14T14:12:46
|
|
BUILDING.md: Fix confusing Windows build reqs
Indent "... OR ..." to make it clear that the choice is between Visual
C++ and MinGW, not Visual C++ and MinGW + NASM. Move NASM to the top of
the list to make that even more clear. Make it clear that nasm.exe
should be in the PATH.
Addresses concerns raised in #70
|
|
|
4f581231
|
2016-03-09T17:23:45
|
|
BUILDING.md and README.md: Cosmetic tweaks
|
|
|
ab80273b
|
2016-02-08T14:41:07
|
|
BUILDING.md: Update OS X Java information
The Apple Java Developer Package is only necessary on OS X < 10.7.
When building on Lion and later, the Oracle JDK is preferred.
|
|
|
8ff67fdb
|
2016-02-06T12:17:40
|
|
Update MinGW Linux build recipe
|
|
|
6c3fc97b
|
2016-02-01T11:12:22
|
|
Update Android build instr. for ARMv8, PIE, etc.
* Include information on how to do a 64-bit ARMv8 build with the latest
NDK
* Suggest -fPIE and -pie as default CFLAGS (required for android-16 and
later.
* Remove -fstrict-aliasing flag (-Wall already includes it)
|
|
|
4e356a92
|
2016-01-13T12:25:03
|
|
BUILDING.md: Restore autotools processing instr.
Partially reverts 54014d9c2a41905b7b766057af6728834da64b59. When
building from a git sandbox, as opposed to from an official source
tarball, it is still necessary to run autoreconf.
Closes #48
|
|
|
f3a8684c
|
2016-01-07T00:19:43
|
|
SSE2 SIMD implementation of Huffman encoding
Full-color compression speedups relative to libjpeg-turbo 1.4.2:
2.8 GHz Intel Xeon W3530, Linux, 64-bit: 2.2-18% (avg. 9.5%)
2.8 GHz Intel Xeon W3530, Linux, 32-bit: 10-25% (avg. 17%)
2.3 GHz AMD A10-4600M APU, Linux, 64-bit: 4.9-17% (avg. 11%)
2.3 GHz AMD A10-4600M APU, Linux, 32-bit: 8.8-19% (avg. 15%)
3.0 GHz Intel Core i7, OS X, 64-bit: 3.5-16% (avg. 10%)
3.0 GHz Intel Core i7, OS X, 32-bit: 4.8-14% (avg. 11%)
2.6 GHz AMD Athlon 64 X2 5050e:
Performance-neutral (give or take a few percent)
Full-color compression speedups relative to IPP:
2.8 GHz Intel Xeon W3530, Linux, 64-bit: 4.8-34% (avg. 19%)
2.8 GHz Intel Xeon W3530, Linux, 32-bit: -19%-7.0% (avg. -7.0%)
Refer to #42 for discussion. Numerous other approaches were attempted,
but this one proved to be the most performant across all platforms.
This commit also fixes #3 (works around, really-- the clang-compiled version
of jchuff.c still performs 20% worse than its GCC-compiled counterpart, but
that code is now bypassed by the new SSE2 Huffman algorithm.)
Based on:
https://github.com/mayeut/libjpeg-turbo/commit/2cb4d41330e1edc4469f6b97ba73b73abfbeb02f
https://github.com/mayeut/libjpeg-turbo/commit/36c94e050d117912adbff9fbcc6fe307df240168
|
|
|
17de5183
|
2015-10-10T10:34:55
|
|
Markdown versions of README, LICENSE, BUILDING
|
|
|
7e3acc0e
|
2015-10-10T10:25:46
|
|
Rename README, LICENSE, BUILDING text files
The IJG README file has been renamed to README.ijg, in order to avoid
confusion (many people were assuming that that was our project's README
file and weren't reading README-turbo.txt) and to lay the groundwork for
markdown versions of the libjpeg-turbo README and build instructions.
|