0.4.27

======

  - sse: preserve non volatile sse registers, needed for MSVC (Matej Knopp)

  - x86: don't hard-code register size to zero in orc_x86_emit_*() functions (Igor Rondarev)

  - Fix incorrect asm generation on 64-bit Windows when building with MSVC (Jan Schmidt)

  - Support build using the Meson build system (Nirbheek Chauhan, Tim-Philipp Müller)

0.4.26

======

  - Use 64 bit arithmetic to increment the stride if needed (Wim Taymans)

  - Fix generation of ModR/M / SIB bytes for the EBP, R12, R13 registers

    on X86/X86-64 (Sebastian Dröge)

  - Fix test_parse unit test if no executable backend is available (Pascal Terjan)

  - Add orc-test path to the -uninstalled .pc file (Josep Torra)

  - Fix compiler warnings in the tests on OS X (Josep Torra)

0.4.25

======

  - compiler: also prefer the backup function when no target, instead

    of trying to use emulation which is usually slower (Wim Taymans)

  - executor: fix load of parameters smaller than 64 bits, fixing crashes

    on ldresnearb and friends in emulated code (Wim Taymans)

  - test-limits: improve test without target (Wim Taymans)

  - Only check for Android's liblog on Android targets, so we don't accidentally

    pick up another liblog that may exist elsewhere (Sebastian Dröge)

  - Don't require libtool for uninstalled setups (-uninstalled pkg-config file)

    (Julien Isorce)

  - Make -Bsymbolic check in configure work with clang (Koop Mast)

  - Coverity code analyser fixes (Luis de Bethencourt)

  - docs: update generated opcode tables

  - add orc_version_string() function and make orcc check the liborc that is

    being picked up to make sure the right lib is being used (Tim-Philipp Müller)

0.4.24

======

  - Only reuse constants of the same size and value (Wim Taymans)

  - Fix reading of .orc files with Windows line endings on

    Windows (Tim-Philipp Müller)

  - Fix out of bounds array access in the tests (Luis de Bethencourt)

  - Remove duplicate code path in orcc (Edward Hervey)

  - Put a limit to the memcpy test (Edward Hervey)

  - Fix mmap leak on error path (Vincent Penquerc'h)

0.4.23

======

  - Various improvements to the NEON backend to bring it closer to the SSE

    backend (Wim Waymans)

  - Add support for setting a custom backup function (Wim Taymans)

  - Preserve NEON/VFP registers across subroutines (Jerome Laheurte)

  - Fix 64 bit parameter loading on big-endian systems (Tim-Philipp Müller)

  - Improved implementations for various opcodes (Wim Taymans)

  - Various improvements and fixes to constants handling (Wim Taymans)

  - Avoid some undefined operations on signed integers (Wim Taymans)

  - Prefer user specific directories over global ones for intermediate files

    to prevent name collisions (Fabian Deutsch)

0.4.22

======

Maintenance release:

  - Handle NOCONFIGURE=1 in autogen.sh (Colin Walters)

  - Some memory leak fixes in the compiler (Sebastian Dröge, Thiago Santos)

  - Fixes for compiler warnings on Win64 (Edward Hervey)

  - Properly detect CPU features on Android in non-debug build (Jan Schmidt)

  - Use Android logging system instead of stderr for debug output (Jan Schmidt)

0.4.21

======

Maintenance release:

  - Add libtool versioning to the linker flags again. This was accidentially

    removed in 0.4.20 but should not cause any problems on platforms other

    than OS X (Sebastian Dröge)

0.4.20

======

Maintenance release:

  - Fix list corruption when splitting code memory chunks, causing crashes

    when allocating a lot of code memory and trying to free it later

    (Tim-Philipp Müller)

  - Add some extra checks for the number of variables used in ORC code to

    prevent overflows and crashes in the compiler (Vincent Penquerc'h)

  - Various compiler warnings, coverity warnings and static code analysis

    fixes (Sebastian Dröge)

0.4.19

======

Maintenance release:

  - Fix out-of-tree builds (Edward Hervey)

  - Fix many memory leaks, compiler warnings and coverity warnings (Tim-Philipp Müller,

    Olivier Crête, Todd Agulnick, Sebastian Dröge, Vincent Penquerc'h, Edward Hervey)

  - Documentation fix for mulhsw, mulhuw (William Manley)

0.4.18

======

Maintenance release:

 - Important bugfix in reading constants from bytecode. (Tim-Philipp Müller

   and Sebastian Dröge)

 - Documentation and code cleanup (Stefan Sauer)

 - Fix cache flushing on iOS (Andoni Morales Alastruey)

0.4.17

======

Maintenance release:

 - Merged known distro patches.

 - Added MIPS backend (Guillaume Emont).

 - Disabled ARM backend because of poor coverage.

 - Added bytecode parsing and writing.  This can be used instead of

   manual creation of OrcPrograms.

0.4.16

======

Fix a few bugs people noticed in 0.4.15.

 - orc_init() tried to take the same mutex as generated C code that

   calls (indirectly) orc_init().

 - sse: Fixes for 64 bit pointers with any of the upper 32 bits set.

0.4.15

======

This should have been release much earlier.

 - Protect global resources with mutexes.  Duh.  This solves a bunch

   of bug reports.

 - Restore c64x-c backend.  Untested.

 - Convert MMX and SSE backends to a new instruction scheduler.

 - Add alignment and size hints to parser.

0.4.14

======

Yet more bug fixing.  Altivec should work again, OS/X should

work again.  MMX should work again.  Another codegen bug on

SSE fixed.

0.4.13

======

Fixes two serious code generation bugs in 0.4.12 on SSE and

Altivec.  Also added some compatibility code to mitigate

the previous automatic inclusion of stdint.h.

0.4.12

======

This is primarily a bug fixing release.

 - Fix gcc-4.6 warnings in generated code

 - Codegen fixes for Altivec.  Passes regression tests again.

 - More error checking for code allocation.

 - NEON: floating point improvements

 - Removed stdint.h from API.  This could theoretically cause

   breakage if you depended on orc to include stdint.h.

One new feature is the OrcCode structure, which keeps track of

compiled code.  This now allows applications to free unused code.

Internally, x86 code generation was completely refactored to add

an intermediate stage, which will later be used for instruction

reordering.  None of this is useful yet.

0.4.11

======

This is primarily a bug fixing release.

 - Fixes for CPUs that don't have backends

 - Fix loading of double parameters

 - mmx: Fix 64-bit parameter loading

 - sse/mmx: Fix x2/x4 with certain opcodes

There are still some issues with the ARM backend on certain

architecture levels (especially ARMv6).  Some assistance from

a user with access to such hardware would be useful.

0.4.10

======

Changes:

 - Added several simple 64-bit opcodes

 - Improved debugging by adding ORC_CODE=emulate

 - Allocation of mmap'd areas for code now has several fallback

   methods, in order to placate various SELinux configurations.

 - Various speed improvements in SSE backend

 - Add SSE implementations of ldreslinl and ldresnearl.

 - Update Mersenne Twister example

There was a bug in the calculation of maximum loop shift that, when

fixed, increases the speed of certain functions by a factor of two.

However, the fix also triggers a bug in Schroedinger, which is fixed

in the 1.0.10 release.

0.4.9

=====

This is primarily a bug fixing release.

Changes:

 - Added handling for 64-bit constants

 - Fix building and use of static library

 - Fix register allocation on Win64 (still partly broken, however)

 - Quiet some non-errors printed by orcc in 0.4.8.

 - Fix implementation of several opcodes.

Until this release, the shared libraries all had the same versioning

information.  This should be fixed going forward.

0.4.8

=====

Changes:

 - Fix Windows and OS/X builds

 - Improve behavior in failure cases

 - Major improvements for Altivec backend

 - Significant documentation additions

Memory for executable code storage is now handled in a much more

controlled manner, and it's now possible to reclaim this memory

after it's no longer needed.

A few more 64-bit opcodes have been added, mostly related to

arithmetic on floating point values.

The orcc tool now handles 64-bit and floating point parameters

and constants.

0.4.7

=====

Changes:

 - Lots of specialized new opcodes and opcode prefixes.

 - Important fixes for ARM backend

 - Improved emulation of programs (much faster)

 - Implemented fallback rules for almost all opcodes for

   SSE and NEON backends

 - Performance improvements for SSE and NEON backends.

 - Many fixes to make larger programs compile properly.

 - 64-bit data types are now fully implemented, although

   there are few operations on them.

Loads and stores are now handled by separate opcodes (loadb,

storeb, etc).  For compatibility, these are automatically

included where necessary.  This allowed new specialized

loading opcodes, for example, resampling a source array

for use in scaling images.

Opcodes may now be prefixed by "x2" or "x4", indicating that

a operation should be done on 2 or 4 parts of a proportionally

larger value.  For example, "x4 addusb" performs 4 saturated

unsigned additions on each of the four bytes of 32-bit

quantities.  This is useful in pixel operations.

The MMX backend is now (semi-) automatically generated from

the SSE backend.

The orcc tool has a new option "--inline", which creates inline

versions of the Orc stub functions.  The orcc tool also recognizes

a new directive '.init', which instructs the compiler to generate

an initialization function, which when called at application init

time, compiles all the generated functions.  This allows the

generated stub functions to avoid checking if the function has

already been compiled.  The use of these two features can

dramatically decrease the cost of calling Orc functions.

Known Bugs: Orc generates code that crashes on 64-bit OS/X.

Plans for 0.4.8: (was 2.5 for 4 this time around, not too bad!)

Document all the new features in 0.4.7.  Instruction scheduler.

Code and API cleanup.

0.4.6

=====

Changes:

 - Various fixes to make Orc more portable

 - Major performance improvements to NEON backend

 - Minor performance improvements to SSE backend

 - Major improvements to ARM backend, now passes regression

   tests.

The defaults for floating point operations have been changed

somewhat: NANs are handled more like the IEEE 754 standard,

and denormals in operations are treated as zeros.  The NAN

changes causes certain SSE operations to be slightly slower,

but produce less surprising results.  Treating denormals as

zero has effects ranging from "slightly faster" to "now possible".

New tool: orc-bugreport.  Mainly this is to provide a limited

testing tool in the field, especially for embedded targets

which would not have access to the testsuite that is not

installed.

The environment variable ORC_CODE can now be used to adjust

some code generation.  See orc-bugreport --help for details.

orcc has a new option to generate code that is compatible

with older versions of Orc.  For example, if your software

package only uses 0.4.5 features, you can use --compat 0.4.5

to generate code that run on 0.4.5, otherwise it may generate

code that requires 0.4.6.  Useful for generating source code

for distribution.

New NEON detection relies on Linux 2.6.29 or later.

Plans for 0.4.7: (not that past predictions have been at all

accurate) New opcodes for FIR filtering, scaling and compositing

of images and video.  Instruction scheduler, helpful for non-OOO

CPUs.  Minor SSE/NEON improvements.  Orcc generation of inline

macros.

0.4.5

=====

This release contains many small improvements related to

converting GStreamer from liboil to Orc.

The major addition in this release is the mainstreaming of

the NEON backend, made possible by Nokia.

There is a new experimental option to ./configure,

--enable-backend, which allows you to choose a single code

generation backend to include in the library.  This is mostly

useful for embedded systems, and is not recommended in general.

The upcoming release will focus on improving code generation

for the SSE and NEON backends.

0.4.4

=====

This is almost entirely a cleanup and bug fix release.

 - fix register copying on x86-64

 - better checking for partial test failures

 - fix documention build

 - fix build on many systems I don't personally use

 - various fixes to build/run on Win64 (Ramiro Polla)

 - add performance tests

Next release will merge in the new pixel compositing opcodes

and the SSE instruction scheduler.

0.4.3

=====

New opcodes: all the 32-bit float opcodes from the orc-float

library have been moved into the core library.

New opcodes: splitlw and splitwb, which are equivalent to

select0lw, select1lw, select0wb, and select1wb, except that

the new opcodes split a value into two destinations in one

opcode.

New backend: c64x-c, for the TI C64x+ DSP.  This backend only

produces source code, unlike other backends which can produce

both source and binary code.  Generating code for this backend

can be done using 'orcc --assembly --target=c64x-c'.

Orc now understands and can generate code for two-dimensional

arrays.  If the size of the array is known at compile time,

this information can be used to improve generated code.

Various improvements to the ARM backend by Wim Taymans.  The

ARM backend is still experimental.

0.4.2

=====

Bug fixes to C backend.  Turns out this is rather important on

CPUs that don't have a native backend.

New features have been postponed to 0.4.3.

0.4.1

=====

This release introduces the orcc program, which parses .orc files and

outputs C source files to compile into a library or application.  The

main source file implements the functions described by the .orc source

code, by creating Orc programs and compiling them at runtime.  Another

source file that it outputs is a test program that can be compiled and

run to determine if Orc is generating code correctly.  In future

releases, the orcc tool will be expanded to output assembly code, as

well as make it easier to use Orc in a variety of ways.

Much of Schroedinger and GStreamer have been converted to use Orc

instead of liboil, as well as converting code that wasn't able to

use liboil.  To enable this in Schroedinger, use the --enable-orc

configure option.  The GStreamer changes are in the orc branch in

the repository at http://cgit.freedesktop.org/~ds/gstreamer

Scheduled changes for 0.4.2 include a 2-D array mode for converting

the remaining liboil functions used in Schroedinger and GStreamer.

Major changes:

 - Add the orcc compiler.  Generates C code that creates Orc programs

   from .orc source files.

 - Improved testing

 - Fixes in the C backend

 - Fix the MMX backend to emit 'emms' instructions.

 - Add a few rules to the SSE backend.

0.4.0

=====

Stuff happened.