/[svn]/linuxsampler/trunk/ChangeLog

Diff of /linuxsampler/trunk/ChangeLog

Parent Directory | Revision Log | View Patch Patch

-revision 609 by schoenebeck,
Sat Jun  4 17:49:12 2005 UTC
+revision 3646 by schoenebeck,
Sun Dec  8 23:17:34 2019 UTC
 Line 1
- Version CVS HEAD (?)
+ Version SVN trunk (?)
+   * general changes:
+     - Fixed compiler error in Pool.h.
+     - Require C++14 compiler support.
+     - Autoconf: Added m4/ax_cxx_compile_stdcxx.m4 macro which is used
+       for checking in configure for C++14 support (as mandatory
+       requirement) and automatically adds compiler argument if required
+       (e.g. -std=C++14).
+     - RTMath: Implemented floating point comparison methods
+       fEqual32(float,float) and fEqual64(double,double) which take the
+       expected floating point tolerances into account.
+     - Added public API C++ class "LFO", which is a cluster class encapsulating
+       all the sampler's LFO implementations to be used by 3rd party applications
+       (e.g. by Gigedit).
+     - Added int math square LFO implementation.
+     - Added int math saw LFO implementation.
+     - Added numeric complex nr sine LFO implementation.
+     - Marked class LFOTriangleDiHarmonic as deprecated
+       (will be removed in future).
+     - Added LFOAll.h which includes all LFO implementation's header files.
+     - Refactored LFO class names and their file names (PulseLFO -> LFOPulse,
+       LFOSawIntMath -> LFOSawIntMathNew, SawLFO -> LFOSawIntMathOld,
+       SineLFO -> LFOSineBuiltinFn, LFOSine -> LFOSineNumericComplexNr,
+       SquareLFO -> LFOSquarePulse and separated the latter to its own header
+       file).
+     - Renamed type LFOSigned -> LFOTriangleSigned.
+     - Renamed type LFOUnsigned -> LFOTriangleUnsigned.
+     - Using now cubic interpolation as resampling algorithm by default; and
+       linear interpolation is deprecated now.
+   * Real-time instrument scripts:
+     - Added method ScriptVM::setExitResultEnabled() which allows to
+       explicitly enable the built-in exit() function to optionally accept
+       one function argument; the value of the passed exit() function
+       argument will then become available by calling
+       VMExecContext::exitResult() after script execution.
+     - 64 bit support for NKSP integer scripts variables (declare $foo).
+     - Variable names, function names and preprocessor condition names must start
+       with a regular character (a-z or A-Z); starting them with a digit or
+       underscore is not allowed.
+     - NKSP parser fix: equal comparison operator "=" and not equal comparison
+       operator "#" must only accept integer operands.
+     - NKSP language: Implemented support for standard units like Hertz, seconds,
+       Bel including support for metric unit prefixes; so one can now e.g.
+       conveniently use numbers in scripts like "5us" meaning "5 microseconds",
+       or e.g. "12kHz" meaning "12 kilo Hertz", or e.g. "-14mdB" meaning
+       "minus 14 Millidecibel", or e.g. "28c" meaning "28 cents" (for tuning).
+     - NKSP language: Introduced "final" operator "!" which is specifically
+       intended for synthesis parameter values to denote that the synthesis
+       parameter value is intended to be the "final" value for that synthesis
+       parameter that should explicitly be used by the engine and thus causing
+       the sampler engine to ignore all other modulation sources for the same
+       synthesis parameter (like e.g. LFO, EG); by simply prefixing a value,
+       variable or formula with this new "!" operator the expression is marked as
+       being "final".
+     - NKSP script editor API: Added support for detecting standard unit tokens
+       and their potential metric prefix token.
+     - NKSP language: Added support for NKSP real number literals and
+       arithmetic operations on them (e.g. "(3.9 + 2.9) / 12.3 - 42.0").
+     - NKSP language: Added support for NKSP real number (floating point) script
+       variables (declare ~foo).
+     - NKSP language: Added support for NKSP real number (floating point) array
+       script variables (declare ?foo[]).
+     - Built-in script function "message()" accepts now real number argument as
+       well.
+     - Added built-in script function "real_to_int()" and its short hand form
+       "int()" for casting from real number to integer in NKSP scripts.
+     - Added built-in script function "int_to_real()" and its short hand form
+       "real()" for casting from integer to real number in NKSP scripts.
+     - Allow built-in exit() function to potentially accept real number type
+       argument as well.
+     - Built-in script functions may have a different return type depending on
+       the arguments passed to the function.
+     - Built-in script function "abs()" optionally accepts and returns real
+       number.
+     - Built-in script functions "min()" and "max()" optionally accept real
+       number arguments and return real number as result in that case.
+     - NKSP VM API: Allow units and 'final'ness to be returned as result from
+       built-in functions (added methods VMFunction::returnUnitType() and
+       VMFunction::returnsFinal() for that purpose which must be implemented by
+       built-in function implementations).
+     - NKSP language: Allow metric unit prefixes of numeric scalar and array
+       variables to be changed freely at runtime (unlike unit types like Hz etc.
+       which are still sticky, parse-time features of variables which cannot be
+       changed at runtime for the intentional sake of determinism).
+     - NKSP language: 'final' values are prohibited for array variables for now
+       (attempt causes a parsers error).
+     - NKSP language: expressions with unit types (e.g. Hz) are prohibited for
+       conditions of runtime control structures like if(), while(), select()
+       (attempt causes a parser error).
+     - NKSP VM API: Allow built-in functions to perform their own, individual
+       parse time checks of arguments going to be passed to the function at
+       runtime (added method VMFunction::checkArgs() for that purpose).
+     - NKSP language: raise parser warning if only one operand of binary
+       operators (like logical 'or' comparison) contain a 'final' value (because
+       it would always yield in a 'final' result in such cases).
+     - NKSP language: Allow comparison (=, #, <, >, <=, >=) of values with
+       different metric unit prefixes, which will behave as expected (e.g.
+       result of expression '1000us < 2ms' is true).
+     - NKSP language: Allow adding values with different metric unit prefixes
+       (e.g. result of expression '100Hz + 5kHz' is '5100Hz').
+     - NKSP language: Allow subtracting values with different metric unit
+       prefixes (e.g. result of expression '1ms - 20us' is '980us').
+     - NKSP language: Allow multiplying with any metric unit prefixes
+       (e.g. result of expression '2k * 3ms' is '6s'), however multiplications
+       with unit types on both sides (e.g. '2s * 2s') is still prohibited since
+       we don't have any considerable practical use for a term like '4s^2'
+       (hence any attempt multiplying two unit types still causes parser error).
+     - NKSP language: Allow dividing by any metric unit prefixes and allow
+       division of same unit type on both sides (e.g. expression '8kHz / 1000Hz'
+       yields in unit free result '8'). So this is now a way to cast units away
+       e.g. for passing the result to other expressions, certain function calls
+       or variables which are not accepting any units (or that specific unit).
+     - NKSP language: integer arrays and real number arrays can now be converted
+       to strings (e.g. for dumping their content with message() calls for
+       script debugging purposes).
+     - NKSP language: expressions and variables with units are now correctly
+       casted to strings (e.g. with message() calls).
+     - NKSP language: comparing real numbers for equalness (e.g. '~foo = 3.1') or
+       unequalness (e.g. '~foo # 3.1') is now less strict and takes the expected
+       floating point tolerances into account.
+     - NKSP VM API: Added methods VMScalarNumberExpr::evalCastInt() and
+       VMScalarNumberExpr::evalCastReal().
+     - NKSP VM API: Added base class 'VMNumberArrayExpr' for classes
+       'VMIntArrayExpr' and 'VMRealArrayExpr'.
+     - NKSP VM API: replaced all unitPrefix() (parse time) methods by
+       unitFactor() (runtime) methods.
+     - Built-in function "exit()" supports now returning units and 'final'ness
+       for test cases.
+     - The following built-in functions support now units as well: "abs()",
+       "random()", "inc()", "dec()", "in_range()", "min()", "max()",
+       "real_to_int()", "int()", "int_to_real()" and "real()".
+     - Built-in functions "array_equal()", "search()" and "sort()" support now
+       real number arrays (correctly) as well.
+     - Added individual parse time checks of arguments to be passed to built-in
+       functions "random()", "inc()", "dec()", "in_range()", "min()", "max()",
+       "array_equal()" and "search()" specific for their individual purposes.
+     - NKSP VM refactoring: Renamed all methods, functions and classes matching
+       pattern *ScalarNumber* to simply *Number* (that is i.e. classes
+       VMScalarNumberExpr -> VMNumberExpr, ScalarNumberExpr -> NumberExpr,
+       ScalarNumberVariable -> NumberVariable, ScalarNumberBinaryOp ->
+       NumberBinaryOp, VMScalarNumberResultFunction -> VMNumberResultFunction,
+       method VMExpr::asScalarNumberExpr() -> VMExpr::asNumber(), function
+       isScalarNumber() -> isNumber()).
+     - NKSP VM API: Added 4 overridden methods to class VMNumberExpr:
+       evalCastInt(MetricPrefix_t), evalCastInt(MetricPrefix_t,MetricPrefix_t),
+       evalCastReal(MetricPrefix_t), evalCastReal(MetricPrefix_t,MetricPrefix_t)
+       as convenient methods for automatically converting values to expected
+       metric value basis.
+     - Built-in function "wait()" accepts now both integers and real numbers as
+       argument.
+     - NKSP VM API cleanup: Get rid of legacy method
+       VMFunction::argType(vmint iArg) which was already superseded by its new
+       replacement VMFunction::acceptsArgType(vmint iArg, ExprType_t type).
+     - NKSP parser: if wrong argument type was passed to a built-in function and
+       that built-in function accepts more than one data type for the argument,
+       then show all supported data types as parser error message.
+     - Built-in function "play_note()" accepts now real numbers and seconds as
+       unit type as well for its 3rd and 4th function arguments.
+     - The following built-in functions accept now real numbers as well for their
+nd function argument: "change_vol()", "change_tune()", "change_cutoff()",
+       "change_attack()", "change_decay()", "change_release()",
+       "change_sustain()", "change_cutoff_attack()", "change_cutoff_decay()",
+       "change_cutoff_sustain()", "change_cutoff_release()",
+       "change_amp_lfo_freq()", "change_cutoff_lfo_freq()",
+       "change_pitch_lfo_freq()", "change_vol_time()", "change_tune_time()",
+       "change_pan_time()", "fade_in()", "fade_out()", "change_play_pos()".
+     - Fixed built-in function "change_play_pos()" not having accepted metric
+       prefixes at all.
+     - Fixed the following built-in functions having misinterpreted values given
+       with unit type (for their 2nd argument) as if they were relative values
+       (that is as if they were passed without a unit type): "change_attack()",
+       "change_decay()", "change_release()", "change_cutoff_attack()",
+       "change_cutoff_decay()", "change_cutoff_release()".
+     - Fixed the following built-in functions having applied completely wrong
+       'final' values: "change_sustain()", "change_cutoff_sustain()" (since the
+       respective EGs being their modulation sink assume uint data type with
+       value range 0..1000 instead of 0.0..1.0.
+     - Added individual parse-time checks of function arguments for the following
+       built-in functions: "play_note()", "note_off()", "set_event_mark()",
+       "delete_event_mark()", "by_marks()", "change_cutoff()", "change_attack()",
+       "change_decay()", "change_release()", "change_cutoff_attack()",
+       "change_cutoff_decay()", "change_cutoff_release()",
+       "change_amp_lfo_freq()", "change_cutoff_lfo_freq()",
+       "change_pitch_lfo_freq()", "change_vol_time()", "change_tune_time()" and
+       "change_pan_time()".
+     - Don't abort function call if unit type was used and at the same time
+       'final' operator was omitted for the primary value argument of the
+       following built-in functions: "change_cutoff()", "change_attack()",
+       "change_decay()", "change_release()", "change_cutoff_attack()",
+       "change_cutoff_decay()", "change_cutoff_release()",
+       "change_amp_lfo_freq()", "change_cutoff_lfo_freq()",
+       "change_pitch_lfo_freq()", "change_vol_time()", "change_tune_time()",
+       "change_pan_time()", instead imply 'final'ness at runtime and raise an
+       appropriate parser warning at parse time.
+     - Added built-in real number functions "round()", "ceil()", "floor()",
+       "sqrt()", "log()", "log2()", "log10()", "exp()", "pow()", "sin()",
+       "cos()", "tan()", "asin()", "acos()", "atan()".
+     - Added built-in script real number constant "~NI_MATH_PI".
+     - Added built-in script real number constant "~NI_MATH_E".
+     - NKSP language: Allow unary '+' operator.
+   * test cases:
+     - Fixed compiler errors in test cases.
+     - Updated README for how to compile & run test cases.
+     - Updated test case
+       MutexTest::testDoubleLockStillBlocksConcurrentThread() to latest
+       expected behaviour of the Mutex class implementation (recursive
+       mutex type).
+     - Added test cases for NKSP core language aspects and core built-in
+       functions.
+     - Fixed thread tests segfaulting on Linux.
+     - NKSP: Added real number test cases for built-in functions exit(),
+       int_to_real(), real(), real_to_int() and int(), as well as for the
+       plus, minus and negate language operators.
+     - Added massive amount of NKSP test cases for standard measuring units and
+       'final' operator usage cases.
+     - Added NKSP test cases for (floating point tolerance aware) real number
+       equalness / unequalness comparison.
+     - Added NKSP int array and real array tests for value assignment and
+       initialization of arrays.
+     - Added NKSP test cases for built-in functions "round()", "ceil()",
+       "floor()", "sqrt()", "log()", "log2()", "log10()", "exp()", "pow()",
+       "sin()", "cos()", "tan()", "asin()", "acos()", "atan()".
+     - Added NKSP test cases for unary '+' operator.
+   * GigaStudio/Gigasampler format engine:
+     - LFOTriangleIntMath and LFOTriangleIntAbsMath: Fixed FlipPhase=true
+       behaviour for start_level_mid.
+     - Changed LFO start levels: LFO1 and LFO2 both to mid, LFO3 to max. start
+       level (see discussion "GigaStudio LFO compatibility" on mailing list
+       from 2019-09-26 for details).
+     - Changed default wave form for all 3 LFOs to sine (instead of triangle;
+       see discussion "GigaStudio LFO compatibility" on mailing list from
+-09-26 for details).
+     - Format extension: Added support for different LFO wave forms (currently
+       either sine [default], triangle, saw or square).
+     - Format extension: Added support for LFO phase displacement (0°..360°).
+     - Format extension: Added support for flipping LFO polarity on LFO 3
+       (in the original gig format this was only available for LFO 1 and LFO 2).
+     - Format extension: Support for additional filter types:
+       lowpass 1/2/4/6-pole, highpass 1/2/4/6-pole, bandpass 2-pole,
+       bandreject 2-pole.
+   * SFZ format engine:
+     - Fixed support for regions with loccN/hiccN conditions on more than one
+       MIDI controller.
+   * Benchmarks:
+     - Fixed benchmarks/triang.cpp falsely having favoured "int math abs"
+       algorithm (since result of 2nd run was not accumulated).
+     - Added benchmark for saw wave (benchmarks/saw.cpp).
+     - Added benchmark for sine wave (benchmarks/sine.cpp).
+     - Added benchmark for square wave (benchmarks/square.cpp).
+     - Increased amount of benchmarks runs by factor 6 to achieve benchmark times
+       which are large enough on modern systems.
+ Version 2.1.1 (27 Jul 2019)
+   * Real-time instrument scripts:
+     - Fixed behavior of built-in NKSP functions change_sustain(),
+       change_cutoff_attack(), change_cutoff_decay(), change_cutoff_sustain()
+       and change_cutoff_release().
+   * general changes:
+     - Only play release trigger samples on sustain pedal up if this behaviour
+       was explicitly requested by the instrument (otherwise only on note-off).
+     - Fixed compiler warnings.
+     - Fixed compilation error when cross-compiling to Mac.
+     - FX Sends: Provide more useful error messages on routing problems
+       (see bug #169).
+     - LSCP doc: Be more clear describing the two distinct approaches
+       of using external vs. internal effects (see bug #169).
+     - "optional" class: Fixed comparison operators.
+   * Gigasampler/GigaStudio format engine:
+     - Format extension: If requested by instrument then don't play release
+       trigger sample on note-off events.
+   * SFZ format engine:
+     - Fixed memory leak when releasing samples
+       (fixes bug #307, patch by Jacek Roszkowski)
+     - Fixed potential crash when a sample is shared by more than one region
+       (fixes bug #308, patch by Jacek Roszkowski).
+     - Opcode 'sample': Added support for built-in sample '*silence'
+       (fixes bug #310, patch by Jacek Roszkowski).
+ Version 2.1.0 (25 Nov 2017)
+   * SFZ format engine:
+     - added support for <global>, <master> and #define (patch by Alby M)
+     - Removed code duplication in SFZ file loading code.
+     - Added support for sfz extension opcode 'script' which may be used to
+       load real-time instrument script file (NKSP script language).
+     - Implemented opcode set_ccN (initial patch by Giovanni Senatore).
+     - Fixed unintended volume fade-in of voices under certain conditions.
+     - sfz parser: allow missing space between header and opcode
+   * Gigasampler/GigaStudio format engine:
+     - Fixed clicks and pumping noise with Lowpass Turbo filter on very low
+       cutoff settings.
+     - Got rid of resembling an ancient GSt misbehavior which did not pitch at
+       all if an up-pitch of more than 40 semi tones was requested (I don't
+       think there is any stock gig sound that requires this behavior to
+       resemble its original sound).
+     - Added support for controlling whether the individual EGADSR stages may
+       be aborted (as LinuxSampler extension to the original GigaStudio 4
+       format).
+   * general changes:
+     - fixed printf type errors (mostly in debug messages)
+     - use unique_ptr instead of auto_ptr when building with C++11
+     - Added RTAVLTree class which is a real-time safe ordered multi-map, thus
+       allowing to sort data efficiently in real-time safe manner.
+     - RTList class: added methods for moving/inserting elements to arbitrary
+       position within a list.
+     - RTList class: added method fromPtr() for creating an Iterator object from
+       a raw element pointer.
+     - Attempt to partly fix resetting engine channels vs. resetting engine, an
+       overall cleanup of the Reset*(), ConnectAudioDevice(),
+       DisconnectAudioDevice() API methods would still be desirable though,
+       because the current situation is still inconsistent and error prone.
+     - Active voices are now internally grouped to "Note" objects, instead of
+       being directly assigned to a keyboard key. This allows more fine graded
+       processing of voices, which is i.e. required for certain instrument
+       script features.
+     - Fix: Release trigger voices were not spawned on sustain pedal up
+       (CC #64) events.
+     - Fix: Release trigger voices can now also distinguish correctly between
+       note-off and sustain pedal up events.
+     - All engines: Increased ramp speed of volume smoother and pan smoother
+       (while slow rate of crossfade smoother is preserved) to allow quick
+       volume and pan changes by instrument scripts for instance.
+     - gig/sf2/sfz: Fixed aftertouch channel pressure events being processed
+       properly and thus were ignored for certain uses.
+     - Added new C++ API method
+       VirtualMidiDevice::SendChannelPressureToSampler().
+     - windows, 32-bit: fixed potential crashes by making sure the stack in
+       sub threads is 16-byte aligned
+     - fixed numerous compiler warnings
+     - Fixed invalid (note-on) event ID being assigned to new Note objects.
+     - Revised fundamental C++ classes "Thread", "Mutex" and "Condition" which
+       fixes potential undefined behavior.
+     - Fixed Note object leak when triggering notes on keys which did not
+       have a valid sample mapped (fixes bug #252).
+     - Fixed compilation errors when compiling with CONFIG_DEVMODE enabled.
+     - linuxsampler binary fix: option --create-instruments-db ignored
+       subsequent optional argument due to glibc's implementation oddity
+       which expects a "=" sign, but no space between them.
+   * packaging changes:
+     - removed unnecessary dependency to libuuid
+       (originated by libgig's usage of it)
+     - Automake: set environment variable GCC_COLORS=auto to allow GCC to
+       auto detect whether it (sh/c)ould output its messages in color.
+     - Debian: Fixed packaging error about invalid "Source-Version"
+       substitution variable.
+     - Debian: Raised Debian compatibility level to Debian 9 "Stretch".
+     - Debian: Added build dependency to libsqlite3-dev for building
+       linuxsampler with instruments DB support.
+   * Real-time instrument scripts:
+     - Implemented scheduler for delayed MIDI events and for suspended scripts.
+     - Built-in script function "wait()": implemented support for function's
+       "duration-us" argument, thus scripts using this function are now
+       correctly resumed after the requested amount of microseconds.
+     - Built-in script function "play_note()": implemented support for
+       function's "duration-us" argument, thus notes triggered with this
+       argument are now correctly released after the requested amount of
+       microseconds.
+     - Fix: script events were not cleared when engine channel was reset,
+       potentially causing undefined behavior.
+     - Fixed crash which happened when trying to reference an undeclared
+       variable.
+     - Built-in script function "play_note()": Added support for passing
+       special value -1 for "duration-us" argument, which will cause the
+       triggered note to be released once the original note was released.
+     - Instrument script classes now exported with the liblinuxsampler C++ API.
+     - Added new API method ScriptVM::syntaxHighlighting() which provides
+       a convenient syntax highlighting backend for external instrument
+       script editor applications.
+     - Added new C++ API class "ScriptVMFactory".
+     - Extended parser issues to provide not only first line and first
+       column, but also last line and last column of issue (thus marking
+       the precise span of the issue within the source code).
+     - Reload script automatically after being modified by an instrument
+       editor.
+     - NKSP language grammar correction: allow empty event handler bodies
+       like "on note end on".
+     - Implemented built-in script function "change_vol()".
+     - Implemented built-in script function "change_tune()".
+     - Implemented built-in script function "change_pan()".
+     - Implemented built-in script function "change_cutoff()".
+     - Implemented built-in script function "change_reso()".
+     - Implemented built-in script function "event_status()".
+     - Added built-in script constants "$EVENT_STATUS_INACTIVE" and
+       "$EVENT_STATUS_NOTE_QUEUE" both for being used as flags for
+       "event_status()" function.
+     - NKSP language: Added support for bitwise operators ".or.", ".and."
+       and ".not.".
+     - NKSP language scanner: Fixed IDs matching to require at least one
+       character (i.e. when matching function names or variable names).
+     - NKSP language scanner: disabled unusued rules.
+     - Fixed behavior of play_note() and note_off() functions which must
+       be distinguished engine internally from "real" MIDI note on/off
+       events in order to avoid misbehaviors like hanging notes.
+     - Implemented built-in script variable "$KSP_TIMER".
+     - Implemented built-in script variable "$NKSP_REAL_TIMER".
+     - Implemented built-in script variable "$NKSP_PERF_TIMER".
+     - Implemented built-in script variable "$ENGINE_UPTIME".
+     - Implemented built-in script function "inc()".
+     - Implemented built-in script function "dec()".
+     - NKSP language fix: division expressions were evaluated too often.
+     - NKSP language fix: string concatenation operator was right
+       associative instead of left (to right).
+     - NKSP language correction: allow empty statements in entire language
+       (i.e. "if end if").
+     - Implemented built-in script function "stop_wait()".
+     - Implemented built-in script variable "$NI_CALLBACK_ID".
+     - Implemented built-in script variable "$NI_CALLBACK_TYPE".
+     - Implemented built-in script variable "$NKSP_IGNORE_WAIT".
+     - Added support for read-only built-in variables (respectively
+       handled by the script parser).
+     - Added built-in script constant "$NI_CB_TYPE_INIT".
+     - Added built-in script constant "$NI_CB_TYPE_NOTE".
+     - Added built-in script constant "$NI_CB_TYPE_RELEASE".
+     - Added built-in script constant "$NI_CB_TYPE_CONTROLLER".
+     - NKSP Language: Added support for user defined script functions.
+     - Implemented built-in script function "change_attack()".
+     - Implemented built-in script function "change_decay()".
+     - Implemented built-in script function "change_release()".
+     - Fixed all change_*() built-in script functions to apply their
+       synthesis parameter changes immediately in case the respective note
+       was triggered at the same time, instead of scheduling the parameter
+       change, especially because it would cause some parameter types's
+       changes either to be ramped (i.e. change_vol()) or other types even
+       to have not effect at all (i.e. change_attack()).
+     - Implemented built-in script function "sh_left()".
+     - Implemented built-in script function "sh_right()".
+     - Implemented built-in script function "min()".
+     - Implemented built-in script function "max()".
+     - NKSP Fix: "init" event handler was not always executed when a script
+       was loaded.
+     - NKSP built-in wait() script function: abort script execution if
+       a negative or zero wait time was passed as argument (since this is
+       a common indication of a bug either of the script or even of the
+       engine, which could lead to RT instability or even worse).
+     - ScriptVM: Implemented automatic suspension of RT safety
+       threatening scripts.
+     - Provide more user friendly error messages on syntax errors.
+     - Fixed NKSP parser warning "Not a statement" when assigning an
+       initializer list to an array variable.
+     - Implemented built-in script array variable "%ALL_EVENTS".
+     - Implemented built-in script function "in_range()".
+     - Implemented built-in script function "change_amp_lfo_depth()".
+     - Implemented built-in script function "change_amp_lfo_freq()".
+     - Implemented built-in script function "change_pitch_lfo_depth()".
+     - Implemented built-in script function "change_pitch_lfo_freq()".
+     - Implemented built-in script function "change_vol_time()".
+     - Implemented built-in script function "change_tune_time()".
+     - Implemented built-in script function "fade_in()".
+     - Implemented built-in script function "fade_out()".
+     - Fixed acceptance of wrong data type of parameters passed to built-in
+       script functions "change_vol()", "change_tune()", "change_pan()",
+       "change_cutoff()", "change_reso()", "change_attack()", "change_decay()",
+       "change_release()", "change_amp_lfo_depth()", "change_amp_lfo_freq()",
+       "change_pitch_lfo_depth()" and "change_pitch_lfo_freq()".
+     - Added built-in script function "get_event_par()" and implemented some
+       of its possible parameter selections.
+     - Added built-in script function "set_event_par()" and implemented some
+       of its possible parameter selections.
+     - Fixed a bunch of scheduler time related bugs.
+     - Fixed polyphonic variables not being reset to zero after usage.
+     - Built-in "ignore_event()" function: argument is now optional, like with
+       built-in function "ignore_controller()".
+     - Implemented built-in script function "change_velo()".
+     - Implemented built-in script function "change_note()".
+     - Adjusted behavior of "change_vol()" and "change_tune()" to a more
+       intuitive behavior if used in combination with "change_vol_time()" or
+       "change_tune_time()" respectively: now tuning/volume changes are only
+       assigned (without delay) immediately to a new note if the respective
+       timing function has not been called before, otherwise the volume/tuning
+       changes are automatically faded (before, only the event's time stamp was
+       relevant).
+      - Implemented built-in script function "array_equal()".
+      - Implemented built-in script function "search()".
+      - Implemented built-in script function "sort()".
+      - NKSP Fix: Never suspend "init" event handlers.
+      - Implemented built-in script function "same_region()" (currently only
+        available for gig format engine).
+      - Added built-in script constant "$NKSP_LINEAR".
+      - Added built-in script constant "$NKSP_EASE_IN_EASE_OUT".
+      - Implemented built-in script function "change_vol_curve()".
+      - Implemented built-in script function "change_tune_curve()".
+      - built-in "play_note()" function now supports a sample playback start
+        offset with argument 3, where special value -1 means to use the regular
+        sample offset as defined by the instrument file.
+      - Built-in array variable %KEY_DOWN[] is now a read-only variable.
+      - Built-in variable $EVENT_NOTE is now a read-only variable.
+      - Built-in variable $EVENT_VELOCITY is now a read-only variable.
+      - built-in "play_note()" function now accepts -2 for its fourth argument
+        (note duration) which means the life time of the note shall be sticked
+        to the requested note number of argument 1.
+      - Fix: built-in "play_note()" function now returns 0 as result value if
+        -1 was passed for its fourth argument (note duration) and the respective
+        parent note is already gone.
+     - Implemented built-in script function "change_play_pos()".
+     - NKSP language: Added support for user declared const array variables.
+     - NKSP language: Raise parser warning if array variable is accessed with
+       an index that exceeds the array's size.
+     - NKSP language fix: Unknown characters were not handled correctly.
+     - NKSP language: Added support for "synchronized .. end synchronized"
+       code blocks.
+     - Implemented built-in script function "abort()" which allows to abort
+       another script handler by passing its callback ID.
+     - Fixed potential memory access bug and potential undefined behavior of
+       "init" event handlers.
+     - Print a time stamp along to each call of built-in function "message()".
+     - ScriptVM API: Added VMParserContext::preprocessorComments() which allows
+       to retrieve all code blocks filtered out by the preprocessor.
+     - Added built-in script function "fork()".
+     - Added built-in array variable %NKSP_CALLBACK_CHILD_ID[].
+     - Added built-in variable $NKSP_CALLBACK_PARENT_ID.
+     - Fixed potential crash when accessing dynamic built-in array variables.
+     - Added built-in script function "callback_status()".
+     - Added built-in constant $CALLBACK_STATUS_TERMINATED.
+     - Added built-in constant $CALLBACK_STATUS_QUEUE.
+     - Added built-in constant $CALLBACK_STATUS_RUNNING.
+     - Removed max. value limitation of built-in functions "change_attack()",
+       "change_decay()" and "change_release()" to i.e. allow passing 2000000
+       for doubling the respective time.
+     - NKSP script editor syntax highlighting API: Fixed app termination due
+       to a lexer start condition stack underrun.
+     - NKSP preprocessor: Fixed wrong behavior on nested USE_CODE_IF() and
+       USE_CODE_IF_NOT() preprocessor statements.
+     - NKSP: Added built-in preprocessor condition NKSP_NO_MESSAGE, which
+       can be set to disable all subsequent built-in "message()" function calls
+       on preprocessor level.
+     - Implemented built-in script function "change_sustain()".
+     - NKSP script editor syntax highlighting API: catch all fatal lexer errors,
+       to avoid the editor app to crash on ill-formed text input.
+     - Added built-in script function "change_pan_time()".
+     - Added built-in script function "change_pan_curve()".
+     - Added built-in script function "change_cutoff_attack()".
+     - Added built-in script function "change_cutoff_decay()".
+     - Added built-in script function "change_cutoff_sustain()".
+     - Added built-in script function "change_cutoff_release()".
+     - Added built-in script function "change_cutoff_lfo_depth()".
+     - Added built-in script function "change_cutoff_lfo_freq()".
+   * Instruments DB:
+     - Fixed memory access bug of general DB access code which lead to
+       undefined behavior.
+     - Cleanup of instruments DB file creation and opening code.
+     - The instrument DB path of linuxsampler's --create-instruments-db argument
+       is now optional, if it is missing, then a default location is used.
+     - Added support for scanning SFZ (.sfz) files.
+     - Added support for scanning Sound Font (.sf2) files.
+     - Fixed undefined DB transaction behavior.
+ Version 2.0.0 (15 July 2015)
+   * packaging changes:
+     - fixed building with newer MinGW-w64
+     - Mac OS X: support the new dir for Core Audio SDK
+     - Mac OS X: fixed building outside source directory
+     - made sure all source files for hostplugins are included when
+       doing "make dist"
+     - removed empty directories left from the cvs to svn migration
+     - fixed compilation with gcc 4.6.1
+     - another "make dist" fix, for the LV2 plugin
+     - made --enable-pthread-testcancel default on Mac OS X
+     - minor Makefile fix for building lscpparser when build and source
+       directories are separate
+     - Mac OS X: made it possible to specify plugin installation dir to
+       configure
+     - Mac OS X: Makefile fix for the install-strip target
+     - fixed compilation with gcc 4.7
+     - fixed configure script error with old autoconf versions
+     - lsatomic.h: use gcc provided atomic functions if building with
+       gcc 4.7 and C++11
+     - modernized configure script
+     - fixed linkage error when building with
+       LDFLAGS="-Wl,--no-undefined" (#190)
+     - fixed compilation with Clang 3.2
+     - removed usage of deprecated Automake variable INCLUDES
+     - fixed building with C++11
+     - build fix: ChangeFlagRelaxed.h was missing in makefile
+     - build fix: libsndfile compiler flags were missing in some
+       makefiles
+     - fix for building with bison 3.0 (#202)
+     - Mac OS X: added temporary hack allowing to spawn gigedit as callback
+       on the process's main thread
+     - fixed build error on newer MinGW
+     - support building with older jack versions
+     - support building with spaces in vst sdk path
+     - enabled automake 'subdir-objects' option and moved external
+       source references (vst, au, asio) from makefiles to cpp files,
+       in order to get rid of warnings from automake 1.14
+   * general changes:
+     - Refactoring: moved the independent code from
+       the Gigasampler format engine to base classes
+     - command line option '--profile' is currently disabled, since the
+       respective profiling code is currently broken
+     - Introduced support for optional environment variable
+       "LINUXSAMPLER_PLUGIN_DIR" which allows to override the directory
+       where the sampler shall look for instrument editor plugins
+       (patch by Luis Garrido, slightly modified).
+     - implemented Roland GS NRPN 1ArrH which allows to set volume per note
+     - implemented Roland GS NRPN 1CrrH which allows to set pan per note
+     - implemented Roland GS NRPN 1DrrH which allows to set reverb send per
+       note (in this implementation of the sampler its simply hard routed to
+       the 1st effect send of the sampler channel, no matter what the actual
+       effect type is)
+     - implemented Roland GS NRPN 1ErrH which allows to set chorus send per
+       note (in this implementation of the sampler its simply hard routed to
+       the 2nd effect send of the sampler channel, no matter what the actual
+       effect type is)
+     - implemented support for internal LADSPA effects
+     - raised limit of program change queue from 100 to 512
+       (as suggested by Alex Stone)
+     - implemented sine LFO, pulse LFO and saw LFO
+     - experimental support for per voice equalization
+     - added command line option --exec-after-init
+     - Introduced new C++ API method:
+       EngineChannel::InstrumentFileName(int index)
+       allowing to retrieve the whole list of files used for the loaded
+       instrument on an engine channel (a.k.a. part). Some GigaStudio
+       instruments for example are splitted over several files like
+       "Foo.gig", "Foo.gx01", "Foo.gx02", ...
+     - Added new C++ API method Sampler::GetGlobalMaxVoices().
+     - Added new C++ API method Sampler::GetGlobalMaxStreams().
+     - Added new C++ API method Sampler::SetGlobalMaxVoices().
+     - Added new C++ API method Sampler::SetGlobalMaxStreams().
+     - Various "const" and "restrict" optimizations.
+     - all engines: add pan CC value to instrument pan parameter before
+       applying panning, instead of using two separate pan functions in
+       series (#182)
+     - added a lock guard class for exception safe mutex handling and
+       used it everywhere appropriate
+     - Immediately apply scale tuning changes to active voices.
+     - Exposed scale tuning to C++ API (along to the already existing standard
+       SysEx way).
+     - lsatomic.h fixes: seq_cst load and store were suboptimal for x86
+       and broken for ppc64. (Seq_cst loads and stores are actually not
+       used in LS, so the bug wasn't noticable.)
+     - lsatomic.h: added ARMv7 support
+     - Added support for multiple MIDI input ports per sampler channel (and
+       added various new C++ methods for this new feature / design change, old
+       C++ API methods for managing SamplerChannel's MIDI inputs are now marked
+       as deprecated but are still there and should provide full behavior
+       backward compatibility).
+     - AbstractEngine::GSChecksum(): don't allocate memory on the stack (was
+       unsafe and caused compilation error with clang 2.x).
+     - Bugfix: only process the latest MIDI program change event.
+     - Introducing the LSCP shell, which provides convenient control of the
+       sampler from the command line by providing LSCP aware features.
+     - VirtualMidiDevice: Added support for program change.
+     - VirtualMidiDevice: Added support for bank select (MSB & LSB).
+     - VirtualMidiDevice: Added support for pitch bend.
+     - Aftertouch: extended API to explicitly handle channel pressure and
+       polyphonic key pressure events (so far polyphonic pressure was not
+       supported at all, and channel pressure was rerouted as CC128 but not
+       used so far).
+     - Added initial support for real-time instrument scripts. The script VM
+       code is shared by all sampler engine implemementations, however only the
+       gig file format currently provides support for storing instrument scripts
+       (as LinuxSampler extension to the original GigaStudio 4 format).
+   * Gigasampler format engine:
+     - implemented the "round robin keyboard" dimension
+     - fixed round robin and random dimensions for cases when number of
+       dimension zones is not a power of two
+     - made round robin use a counter for each region instead of each
+       key
+     - bugfix: pitch LFO controller "internal+aftertouch" was broken
+     - bugfix: filter keyboard tracking was broken
+     - filter performance fix (an unnecessary copy was made of the
+       filter parameters in each sub fragment)
+     - handle special case when pan parameter in gig file has max or
+       min value
+     - Exclusive Groups: don't ever stop voices of the same note,
+       doesn't sound naturally with a drumkit
+     - fixed EG1 modulation when attack or release is zero
+     - Fixed support for 'aftertouch' attenuation controller.
+     - Fixed crash that happened with velocity split sounds under certain
+       conditions (see also previous commit on libgig).
+     - fixed behaviour of filter LFO
+   * SFZ format engine:
+     - Initial implementation (not usable yet)
+     - added support for v2 multiple stage envelope generators
+     - added a fine-tuned v1 envelope generator instead of using the
+       one from the gig engine
+     - fixed memory leak and memory handling errors
+     - added support for trigger=first, trigger=legato and sw_previous
+     - allow non-numerical key values ("C#4" for example)
+     - "key" opcode now sets pitch_keycenter too
+     - fixed error when unloading instrument with same sample used by
+       multiple regions
+     - added some opcode aliases, like loopmode for loop_mode, to be
+       more compatible
+     - added support for trigger=release and rt_decay
+     - added support for off_mode=normal
+     - added support for random, seq_position, seq_length and volume
+     - added v1 LFO opcodes to parser (no support in engine yet)
+     - added support for amp_veltrack and amp_velcurve_N
+     - fine-tuned the default velocity curve
+     - added support for transpose
+     - fixed crash when using sw_down/up
+     - improved logic for sw_lokey/hikey/up/down/last
+     - added more v1 aliases to parser
+     - reduced memory usage for sfz data
+     - RT-safeness: avoid malloc in audio thread
+     - fixed a bug that could cause voice stealing to fail
+     - optimized sample lookup
+     - improved support for exclusive groups (group, off_by and
+       off_mode)
+     - added support for controller triggered regions (on_locc/on_hicc)
+     - added support for loop_mode=one_shot
+     - fixed playback of 16 bit wav files on big endian CPUs
+     - added support for Ogg Vorbis sample files
+     - use loop markers from sample file if loop_start and loop_end are
+       not set in sfz file
+     - implemeted filters. Filter types: lowpass, bandpass, bandreject
+       and highpass. 1, 2, 4 and 6 pole filters. Opcodes: fil_type,
+       cutoff, resonance, fil_veltrack, fil_keytrack, fil_keycenter,
+       cutoff_cc, cutoff_chanaft.
+     - bugfix: zero ampeg_sustain didn't work
+     - added support for velocity effect on amplifier envelope time
+       (ampeg_vel2attack, ampeg_vel2decay, ampeg_vel2sustain and
+       ampeg_vel2release)
+     - added support for EG hold (ampeg_hold)
+     - added support for sample offset (offset)
+     - use the newly introduced signal units model
+     - implemented opcodes lfoN_delay,
+       lfoN_freq, lfoN_pan, lfoN_cutoff, lfoN_resonance
+     - implemented opcodes ampeg_delay, ampeg_vel2delay
+       pitcheg_delay, pitcheg_start, pitcheg_attack, pitcheg_hold,
+       pitcheg_decay, pitcheg_sustain, pitcheg_release, pitcheg_vel2delay,
+       pitcheg_vel2attack, pitcheg_vel2hold, pitcheg_vel2decay,
+       pitcheg_vel2sustain, pitcheg_vel2release, pitcheg_depth
+     - implemented opcodes pitchlfo_delay, pitchlfo_freq,
+       pitchlfo_depth, fillfo_delay, fillfo_freq, fillfo_depth,
+       amplfo_delay, amplfo_freq, amplfo_depth
+     - implemented opcodes fileg_delay, fileg_start, fileg_attack,
+       fileg_hold, fileg_decay, fileg_sustain, fileg_release, fileg_vel2delay,
+       fileg_vel2attack, fileg_vel2hold, fileg_vel2decay, fileg_vel2sustain,
+       fileg_vel2release
+     - implemented opcode lfoN_wave
+     - implemented opcode pitchlfo_depthccN
+     - implemented opcodes lfoN_phase, lfoN_phase_onccX,
+       lfoN_pitch, lfoN_pitch_onccX
+     - implemented opcodes pitchlfo_fade,
+       fillfo_fade, amplfo_fade, lfoN_fade, lfoN_fade_onccX
+     - implemented opcodes fillfo_freqccN,
+       pitchlfo_freqccN, amplfo_freqccN, lfoN_freq_onccX
+     - implemented opcodes ampeg_delayccN, ampeg_startccN,
+       ampeg_attackccN, ampeg_holdccN, ampeg_decayccN, ampeg_sustainccN,
+       ampeg_releaseccN, egN_timeX_onccY, egN_levelX_onccY
+     - lfoN_* and egN_* opcodes defined in group sections
+       are now taken into account
+     - implemented curves
+     - implemented opcodes volume_onccN, volume_curveccN
+     - implemented opcode volume_smoothccN
+     - implemented opcodes fillfo_depthccN, amplfo_depthccN,
+       lfoN_volume, lfoN_volume_onccX, lfoN_volume_smoothccX,
+       lfoN_freq_smoothccX, lfoN_pitch_smoothccX, lfoN_pan_onccX,
+       lfoN_pan_smoothccX, lfoN_cutoff_onccX, lfoN_cutoff_smoothccX,
+       lfoN_resonance_onccX, lfoN_resonance_smoothccX, lfoN_delay_onccX
+     - implemented opcode end
+     - implemented opcodes egN_amplitude_onccX,
+       egN_volume, egN_volume_onccX, egN_cutoff, egN_cutoff_onccX,
+       egN_pitch, egN_pitch_onccX, egN_resonance, egN_resonance_onccX
+     - implemented opcodes xfin_lokey, xfin_hikey,
+       xfout_lokey, xfout_hikey, xf_keycurve, xfin_lovel, xfin_hivel,
+       xfout_lovel, xfout_hivel, xf_velcurve, xfin_loccN, xfin_hiccN,
+       xfout_loccN, xfout_hiccN, xf_cccurve
+     - implemented opcodes pan_onccN, pan_smoothccN,
+       pan_curveccN, egN_pan, egN_pan_curve, egN_pan_onccX,
+       egN_pan_curveccX, pitch_veltrack
+     - when failed to parse a sfz file
+       print the line number on which the error occurs
+     - use common pool of CC objects to minimize RAM usage
+     - implemented opcodes amplfo_delay_onccN,
+       amplfo_fade_onccN, fillfo_delay_onccN, fillfo_fade_onccN,
+       pitchlfo_delay_onccN, pitchlfo_fade_onccN
+     - implemented opcodes fileg_delay_onccN,
+       fileg_start_onccN, fileg_attack_onccN, fileg_hold_onccN,
+       fileg_decay_onccN, fileg_sustain_onccN, fileg_release_onccN,
+       fileg_depth_onccN, pitcheg_delay_onccN, pitcheg_start_onccN,
+       pitcheg_attack_onccN, pitcheg_hold_onccN, pitcheg_decay_onccN,
+       pitcheg_sustain_onccN, pitcheg_release_onccN, pitcheg_depth_onccN
+     - implemented automatic aliasing *ccN <-> *_onccN
+     - *lfo_freqccN wasn't working when the respective *lfo_freq
+       was not set or was set to zero
+     - lfoN_freq_onccX wasn't working when lfoN_freq
+       was not set or was set to zero
+     - implemented opcodes resonance_onccN, resonance_smoothccN,
+       resonance_curveccN, cutoff_smoothccN, cutoff_curveccN
+     - implemented opcodes fillfo_depthchanaft,
+       fillfo_freqchanaft, amplfo_depthchanaft, amplfo_freqchanaft,
+       pitchlfo_depthchanaft, pitchlfo_freqchanaft
+     - implemented opcodes pitch_onccN,
+       pitch_curveccN, pitch_smoothccN, pitch_stepccN
+     - implemented opcodes volume_stepccN, pan_stepccN,
+       cutoff_stepccN, resonance_stepccN, lfoN_freq_stepccX,
+       lfoN_volume_stepccX, lfoN_pitch_stepccX, lfoN_pan_stepccX,
+       lfoN_cutoff_stepccX, lfoN_resonance_stepccX
+     - implemented opcodes eq1_freq, eq2_freq, eq3_freq,
+       eq1_freqccN, eq2_freqccN, eq3_freqccN, eq1_bw, eq2_bw, eq3_bw,
+       eq1_bwccN, eq2_bwccN, eq3_bwccN, eq1_gain, eq2_gain, eq3_gain,
+       eq1_gainccN, eq2_gainccN, eq3_gainccN
+     - implemented opcodes delay, delay_onccN, delay_random,
+       delay_samples, delay_samples_onccN
+     - implemented opcodes egN_eq1gain, egN_eq2gain, egN_eq3gain,
+       egN_eq1gain_onccX, egN_eq2gain_onccX, egN_eq3gain_onccX, egN_eq1freq,
+       egN_eq2freq, egN_eq3freq, egN_eq1freq_onccX, egN_eq2freq_onccX,
+       egN_eq3freq_onccX, egN_eq1bw, egN_eq2bw, egN_eq3bw, egN_eq1bw_onccX,
+       egN_eq2bw_onccX, egN_eq3bw_onccX, lfoN_eq1gain, lfoN_eq2gain,
+       lfoN_eq3gain, lfoN_eq1gain_onccX, lfoN_eq2gain_onccX, lfoN_eq3gain_onccX,
+       lfoN_eq1gain_smoothccX, lfoN_eq2gain_smoothccX, lfoN_eq3gain_smoothccX,
+       lfoN_eq1gain_stepccX, lfoN_eq2gain_stepccX, lfoN_eq3gain_stepccX,
+       lfoN_eq1freq, lfoN_eq2freq, lfoN_eq3freq, lfoN_eq1freq_onccX,
+       lfoN_eq2freq_onccX, lfoN_eq3freq_onccX, lfoN_eq1freq_smoothccX,
+       lfoN_eq2freq_smoothccX, lfoN_eq3freq_smoothccX, lfoN_eq1freq_stepccX,
+       lfoN_eq2freq_stepccX, lfoN_eq3freq_stepccX, lfoN_eq1bw, lfoN_eq2bw,
+       lfoN_eq3bw, lfoN_eq1bw_onccX, lfoN_eq2bw_onccX, lfoN_eq3bw_onccX,
+       lfoN_eq1bw_smoothccX, lfoN_eq2bw_smoothccX, lfoN_eq3bw_smoothccX,
+       lfoN_eq1bw_stepccX, lfoN_eq2bw_stepccX, lfoN_eq3bw_stepccX
+     - implemented opcodes eq1_vel2freq, eq2_vel2freq,
+       eq3_vel2freq, eq1_vel2gain, eq2_vel2gain, eq3_vel2gain
+     - sfz parser: allow double spaces in sample filenames
+     - sfz parser: allow absolute paths for sample filenames
+     - use linear decay and release for filter and pitch EG
+     - bugfix: only the first amp_veltrack definition in a file was
+       used
+     - bugfix: looping was disabled if loop_start was set to 0
+     - allow regions with end=-1 to turn off other regions using the
+       group and off_by opcodes (#168)
+     - made end=0 play the whole sample
+     - fixed support for lochan and hichan opcodes (#155)
+     - fixed crash when using lochan/hichan opcodes (#187)
+     - sfz parser: allow -200 to 200 for pan_oncc opcode (#182)
+     - added FLAC support (#191)
+     - sfz parser bugfix: lines starting with whitespace were ignored
+     - added amplitude opcode
+     - added support for "#include" instruction
+       (modified patch which was originally posted by Sergey on LS mailing list)
+     - bugfix: generation of velocity curves etc should not be done
+       after each "#include", only after the main file is parsed
+     - bugfix: line numbers in error messages were wrong after "#include"
+     - added support for float and 32 bit sample files
+   * SoundFont format engine:
+     - Initial implementation (not usable yet)
+     - RT-safeness: avoid malloc in audio thread
+     - fixed a bug that could cause voice stealing to fail
+     - fine-tuned amplitude EG (by switching from gig to sfz EG)
+     - initial implementation of Vibrato LFO and Modulation LFO
+     - initial implementation of cutoff filter
+     - use linear decay and release for filter and pitch EG
+   * Host plugins (VST, AU, LV2, DSSI):
+     - AU bugfix: failed to destroy its audio/MIDI devices
+     - Listen to all interfaces on Mac OS X (INADDR_ANY)
+     - VST bugfix: If the host called resume() before and after
+       changing sample rate or block size, the number of channels was
+       incorrectly set to two. This caused silence in Cubase 5.
+     - save engine type (gig, sfz or sf2) in plugin state
+     - VST: when opening Fantasia, look for both 32 and 64 bit Java on
+bit Windows
+     - AU: changed number of output channels from one stereo to 16
+       stereo
+     - VST: made it possible to build the VST plugin for Mac
+     - AU: link AU plugin dynamically if --disable-shared isn't
+       specified
+     - LV2 "state" extension support (patch by David Robillard)
+     - VST bugfix: instrument loading hang and crashed the host when
+       the plugin was loaded a second time (#174)
+     - plugin bugfix: instrument loading hang when the plugin was
+       loaded a second time (this time it's for Linux and Mac, previous
+       similar fix was for Windows)
+     - thread safety fixes for the instrument loading thread
+     - LV2: use the new lv2 package if present
+     - VST: try to open Fantasia automatically on Linux and Mac too (on
+       Linux, the Fantasia jar should be placed in <prefix>/share/java)
+     - VST: fixed crashes on Linux Ardour and EnergyXT
+     - DSSI bugfix: it wasn't possible to change engine type. The MIDI
+       port and audio channel routing for DSSI plugins are now visible.
+     - LV2: use urid and atom extensions instead of deprecated uri-map
+       and event
+     - LV2: lv2 package 1.0 is now required to build the LV2 plugin
+     - LV2: changed number of output channels to 16 stereo, just like
+       the VST and AU plugins
+     - LV2: fixed save/restore of SFZ state (patch by David Robillard)
+     - LV2: made LV2 plugin buildable on Windows and Mac
+     - VST: implemented retrieval and switching of programs using the
+       sampler's internal MIDI instrument mapping system
+   * Instrument editor interface:
+     - Changed instrument editor plugin interface, providing additional
+       informations like the EngineChannel for which the instrument editor was
+       spawned for. This allows the instrument editors to interact more actively
+       with the sampler.
+   * MIDI driver:
+     - ALSA MIDI driver supports now "NAME" device parameter, for overriding
+       the ALSA sequencer client name
+     - removed limit of maximum amount of MIDI ports per MIDI device, since
+       there is no reason for this limit
+     - MME: fixed memory handling bug found with cppcheck
+     - MME: removed compiler warning
+     - CoreMIDI: implemented driver specific port parameter "CORE_MIDI_BINDINGS",
+       which allows to retrieve the list of CoreMIDI clients / ports and to connect
+       to them a la JACK, via the usual sampler APIs
+     - CoreMIDI: added driver specific port parameter "AUTO_BIND", if enabled
+       the driver will automatically connect to other CoreMIDI clients' ports
+       (e.g. external MIDI devices being attached to the Mac)
+     - added support for MIDI note on velocity filter
+     - CoreMIDI: fixed memory deallocation error
+     - Fixed variable underflow in VirtualMidiDevice, which caused graphical
+       virtual keyboards in frontends / instrument editors being stuck.
+     - Bugfix in VirtualMidiDevice: process note on with velocity 0 as note off.
+     - Implemented missing handling of MIDI "running status".
+     - CoreMIDI fix: a MIDIPacket can contain more than one event per packet.
+     - MME bugfix: driver wasn't closed properly
+     - CoreMIDI: automatically connect to all input sources by default (driver
+       parameter "AUTO_BIND").
+     - CoreMIDI: fixed auto bind feature to CoreMIDI ports that go online
+     - CoreMIDI: fixed minor memory leak
+   * audio driver:
+     - ASIO driver fixes for newer gcc versions (fix from PortAudio)
+     - JACK audio: react on sample rate changes.
+     - JACK audio: react on buffer size changes.
+     - JACK audio: jack_port_get_buffer() was cached and called outside
+       RT context.
+     - ASIO driver: removed compiler warnings
+     - CoreAudio: fixed minor error handling bug
+     - ASIO driver: be more verbose when no ASIO card could be found (fixes #203)
+     - JACK audio: return the JACK server's current sample rate as default value
+       for audio device parameter "SAMPLERATE" (fixes #166).
+   * LSCP server:
+     - added support for sending MIDI CC messages via LSCP command
+       "SEND CHANNEL MIDI_DATA CC <sampler-chan> <ctrl> <val>"
+     - added LSCP command "GET AVAILABLE_EFFECTS"
+     - added LSCP command "LIST AVAILABLE_EFFECTS"
+     - added LSCP command "GET EFFECT INFO <effect-index>"
+     - added LSCP command "CREATE EFFECT_INSTANCE <effect-index>"
+     - added LSCP command
+       "CREATE EFFECT_INSTANCE <effect-system> <module> <effect-name>"
+     - added LSCP command "DESTROY EFFECT_INSTANCE <effect-instance>"
+     - added LSCP command "GET EFFECT_INSTANCES"
+     - added LSCP command "LIST EFFECT_INSTANCES"
+     - added LSCP command "GET EFFECT_INSTANCE INFO <effect-instance>"
+     - added LSCP command
+       "GET EFFECT_INSTANCE_INPUT_CONTROL INFO <effect-instance> <input-control>"
+     - added LSCP command "SET EFFECT_INSTANCE_INPUT_CONTROL VALUE
+       <effect-instance> <input-control> <value>"
+     - added LSCP command "GET SEND_EFFECT_CHAINS <audio-device>"
+     - added LSCP command "LIST SEND_EFFECT_CHAINS <audio-device>"
+     - added LSCP command "ADD SEND_EFFECT_CHAIN <audio-device>"
+     - added LSCP command
+       "REMOVE SEND_EFFECT_CHAIN <audio-device> <effect-chain>"
+     - added LSCP command
+       "GET SEND_EFFECT_CHAIN INFO <audio-device> <effect-chain>"
+     - added LSCP command "APPEND SEND_EFFECT_CHAIN EFFECT <audio-device>
+       <effect-chain> <effect-instance>"
+     - added LSCP command "INSERT SEND_EFFECT_CHAIN EFFECT <audio-device>
+       <effect-chain> <effect-chain-pos> <effect-instance>"
+     - added LSCP command "REMOVE SEND_EFFECT_CHAIN EFFECT <audio-device>
+       <effect-chain> <chain-pos>"
+     - added LSCP command "SET FX_SEND EFFECT <sampler_channel>
+       <fx_send_id> <effect_chain> <chain_pos>"
+     - added LSCP command "REMOVE FX_SEND EFFECT <sampler_channel> <fx_send_id>"
+     - added LSCP commands "SUBSCRIBE EFFECT_INSTANCE_COUNT",
+       "SUBSCRIBE EFFECT_INSTANCE_INFO", "SUBSCRIBE SEND_EFFECT_CHAIN_COUNT",
+       "SUBSCRIBE SEND_EFFECT_CHAIN_INFO"
+     - provide comprehensive error messages on LSCP syntax errors
+       (suggesting expected next non-terminal symbols)
+     - Fixed client connection not being closed after network errors.
+   * LSCP shell:
+     - Added support for auto correction of obvious and trivial syntax mistakes.
+     - Added support for auto completion by tab key.
+     - Show currently available auto completion while typing.
+     - Added support for browsing command history with up / down keys.
+     - Show all possible next symbols immediately right to the current command
+       line while typing (no double tab required for this feature, as it would
+       be the case in other shells).
+     - Added support for moving cursor left/right with arrow keys.
+     - Added support for built-in LSCP reference documentation, which will
+       automatically show the relevant LSCP reference section on screen as soon
+       as one specific LSCP command was detected while typing on the command
+       line.
+   * Real-time instrument scripts:
+     - Implemented built-in script array variable %CC.
+     - Implemented built-in script int variable $CC_NUM.
+     - Implemented built-in script int variable $EVENT_NOTE.
+     - Implemented built-in script int variable $EVENT_VELOCITY.
+     - Implemented built-in script constant variable $VCC_MONO_AT.
+     - Implemented built-in script constant variable $VCC_PITCH_BEND.
+     - Implemented execution of script event handler "init".
+     - Implemented execution of script event handler "controller".
+     - Implemented execution of script event handler "note".
+     - Implemented execution of script event handler "release".
+     - Implemented built-in script function "play_note()" (only two of the
+       max. four function arguments are currently implemented yet though).
+     - Implemented built-in script int variable $EVENT_ID.
+     - Implemented built-in script function "ignore_event()"
+     - Implemented built-in script function "ignore_controller()" (may have one
+       or no argument).
+     - Implemented built-in script function "set_controller()".
+     - Added extended script VM for the Gigasampler/GigaStudio format sampler
+       engine, which extends the general instrument script VM with Giga format
+       specific variables and functions.
+     - Giga format scripts: added built-in script int constant variables
+       $GIG_DIM_CHANNEL, $GIG_DIM_LAYER, $GIG_DIM_VELOCITY, $GIG_DIM_AFTERTOUCH,
+       $GIG_DIM_RELEASE, $GIG_DIM_KEYBOARD, $GIG_DIM_ROUNDROBIN, $GIG_DIM_RANDOM,
+       $GIG_DIM_SMARTMIDI, $GIG_DIM_ROUNDROBINKEY, $GIG_DIM_MODWHEEL,
+       $GIG_DIM_BREATH, $GIG_DIM_FOOT, $GIG_DIM_PORTAMENTOTIME, $GIG_DIM_EFFECT1,
+       $GIG_DIM_EFFECT2, $GIG_DIM_GENPURPOSE1, $GIG_DIM_GENPURPOSE2,
+       $GIG_DIM_GENPURPOSE3, $GIG_DIM_GENPURPOSE4, $GIG_DIM_SUSTAIN,
+       $GIG_DIM_PORTAMENTO, $GIG_DIM_SOSTENUTO, $GIG_DIM_SOFT,
+       $GIG_DIM_GENPURPOSE5, $GIG_DIM_GENPURPOSE6, $GIG_DIM_GENPURPOSE7,
+       $GIG_DIM_GENPURPOSE8, $GIG_DIM_EFFECT1DEPTH, $GIG_DIM_EFFECT2DEPTH,
+       $GIG_DIM_EFFECT3DEPTH, $GIG_DIM_EFFECT4DEPTH, $GIG_DIM_EFFECT5DEPTH.
+     - Giga format scripts: Implemented built-in script function
+       "gig_set_dim_zone(event_id, dimension, zone)".
+     - Implemented built-in script int array variable %KEY_DOWN.
+     - Implemented built-in script function "abs()".
+     - Implemented built-in script function "random()".
+     - Implemented built-in script function "num_elements()".
+     - Implemented built-in script function "note_off()".
+     - Implemented built-in script function "set_event_mark()".
+     - Implemented built-in script function "delete_event_mark()".
+     - Implemented built-in script function "by_marks()".
+     - Added built-in script int const variables $MARK_1 to $MARK_28.
+     - Built-in script functions "ignore_event()", "note_off()" and
+       "gig_set_dim_zone()" now also accept an array of event IDs as argument
+       (i.e. return value of new script function "by_marks()").
+     - Pass/preserve polyphonic variable data from respective "note" event
+       handler to "release" event handler.
+     - Fixed crash when using built-in script function "by_marks()".
+   * Bug fixes:
+     - Fixed crash which may occur when MIDI key + transpose is out of range
+     - minor valgrind fixes
+     - fixed crash which occurred when changing an already deployed sampler
+       channel to a different engine type
+     - fixed crash when deleting a sampler channel or changing engine
+       type while an instrument load was in progress
+     - bugfix: playing a note while changing the instrument could cause
+       a crash, or give "Handing back unknown region" error messages
+     - bugfix: calling SET VOICES (which Fantasia does on start and
+       refresh) could cause instruments to be unloaded
+     - Mac OS X: fixed name collision of enum in EffectControl and
+       wrong name of destructor in AudioOutputDeviceCoreAudio.cpp
+     - Mac OS X: fixed hanging threads
+     - Windows: disabled the previous thread fix on non-Mac systems, as
+       it caused hanging threads on Windows
+     - Fixed possible crashes due to corrupted MIDI/audio device list
+       after MIDI/audio device creation failure
+     - When creating MIDI instrument map entries with "PERSISTENT" type, the
+       instruments were uselessly precached with zero samples, however it still
+       took the full preloading time and on 1st program change the respective
+       instrument was completely reloaded again.
+     - fixed handling of rapid bank select and program change messages
+       sent to the same sampler channel (patch from the Open Octave
+       project, slightly adjusted)
+     - fixed crash when trying to create an effect instance with controls
+       which min and/or max values depend on the sample rate
+     - fixed bug #162
+     - bugfix: LADSPA_PATH was not evaluated correctly when containing
+       multiple paths (#165)
+     - thread safety fixes for the instrument loading thread
+     - bugfix: instrument loading crashed for sfz and sf2 in Ardour
+       (#176)
+     - more thread safety fixes for the instrument loading thread
+     - sfz/sf2 engine: fixed crash when using small audio fragment size
+     - Mac OS X: fixed crash when unloading plugin on 10.7 and later
+     - Mac OS X: fixed process hang when unloading 32-bit plugin (bug
+       introduced in previous fix)
+     - fixed crash when a channel received a program change while
+       playing a note in a key group
+     - fixed erroneous error message piping in VoiceBase.h
+       (labelled "Disk stream not available in time")
+     - Update effects on sample rate & period size changes (to avoid
+       crashes and noise on such transitions).
+     - ignore missing LADSPA paths without ignoring valid LADSPA paths
+       (fixes #208)
+     - Fixed bug in LSCP grammar definition which caused a statement like
+       "GET SERVER INFOasdf\n" to be accepted as valid statement (was so far
+       practically irrelevant, however it caused problems with the new LSCP
+       shell's auto completion feature).
+     - Fixed MIDI program change messages being ignored if quickly executed
+       after each other (fixes #231).
+ Version 1.0.0 (31 July 2009)
+   * packaging changes:
+     - autoconf bugfix: the PKG_CONFIG variable wasn't initialized properly,
+       causing e.g. the libgig test to fail when
+       "./configure --disable-jack-driver" was used
+       (patch by Alexis Ballier)
+     - fixed compilation with gcc 4.3
+     - fixes for building on OS X (thanks to Ebrahim Mayat for testing)
+     - fixed configure so it detects x86_64 (#107)
+     - fixes for building with newer MinGW versions
+     - fix for building with bison 2.4 (#111)
+     - fixed building with libgig installed in a non-standard directory
+     - minor fix in configure for mmsystem.h detection on MinGW
+     - Windows: look for editor plugins and Fantasia using base
+       directory of liblinuxsampler dll (look in the same directory and one
+       directory above)
+     - configure script fix: removed unconditional use of SSE
+     - fixed building with sqlite installed in a non-standard directory
+     - when cross-compiling, don't try to create instruments.db
+     - fix for new mingw-w64 version, which has usleep
+   * general changes:
+     - bugfix: on some POSIX systems instrument editor plugins refused to
+       load as we used a non-portable Linux specific struct field
+       (fixes bug #70, patch by Ronald Baljeu)
+     - fixed endless loop which occured when loading an instrument editor
+       plugin DLL on Windows systems failed
+     - fixed memory leaks that occurred when liblinuxsampler was unloaded
+     - optimized the SynchronizedConfig class so it doesn't wait
+       unnecessarily long after an update
+     - added support for notifying instrument editors on note-on / note-off
+       events (e.g. to highlight the pressed keys on the virtual keyboard
+       of gigedit)
+     - added support for triggering notes by instrument editors (see above)
+     - be verbose on DLL load errors (on Linux)
+     - fixed di-harmonic triangle LFO implementation (this LFO implementation
+       is only used by very few systems, most pick the int math
+       implementation, which is usually faster)
+     - fixes for audio drivers with varying buffer sizes
+     - experimental support for running LinuxSampler as a DSSI, LV2 and
+       VST plugin
+     - notification events for stream/voice count statistics are now sent
+       only when there are actual changes
+     - added memory ordering constraints to improve stability on
+       multi-core and multi-cpu systems
+     - maximum voices and disk streams can now be altered at runtime
+     - fixed CPU feature detection on x86_64 (maybe fixes #108)
+     - automatic stacktrace mechanism is now turned off by default and can
+       be switched on by command line option "--stacktrace" (the automatic
+       stacktrace mechanism seems to be broken on most systems at the moment)
+     - C++ API method InstrumentManager::LaunchInstrumentEditor() now returns
+       a pointer to the launched InstrumentEditor object
+     - added optional 3rd party user data parameter for following
+       liblinuxsampler C++ API methods: InstrumentEditor::Main(),
+       InstrumentEditor::Launch(),
+       InstrumentManager::LaunchInstrumentEditor()
+     - theoretical fix: made SynchronizedConfig follow C++0x memory
+       model more strictly
+     - fixes for using large audio device buffers
+     - Windows: add the installation directory to the DLL search path
+       when loading an editor plugin (solves problems with VST and
+       gigedit on systems with other GTK versions installed)
+     - updated linuxsampler man page
+   * audio driver:
+     - removed the nonsense audio channel constraint (which was hard coded to
+       max. 100 audio channels) for most audio drivers
+     - JACK audio driver did not offer a device parameter "SAMPLERATE" as
+       opposed to the LSCP specs
+     - bugfix: the SAMPLERATE parameter of some drivers (e.g. JACK)
+       reflected the wrong value
+     - fixed a memory management bug in ASIO driver
+     - Makefile fix: JACK_CFLAGS wasn't used
+     - JACK: use jack_client_open instead of the deprecated
+       jack_client_new
+     - added (experimental) CoreAudio driver
+     - applied old fixes to the ASIO driver that were included in the
+       previous binary release but accidentally never committed to CVS
+       (fixes #117)
+     - fixes for ASIO on mingw-w64 (iasio wrapper is not needed on
+       win64)
+     - VST: added support for sample rate and buffer size changes
+     - VST: close editor (Fantasia) when the VST is removed
+     - VST: avoid opening Fantasia more than once for each VST instance
+     - VST: export main function as "main" on Linux too (fix for
+       energyXT)
+     - VST: changed number of output channels from one stereo to 16
+       stereo
+     - added channel routing, fxsends and midi maps to the settings
+       stored in the plugin state
+     - performance optimization of AudioChannel::MixTo() and
+       AudioChannel::CopyTo() methods using GCC vector exensions
+       (if available)
+     - ASIO fixes: avoid initializing the device twice, avoid throwing
+       exception when getting parameters from a disconnected device
+   * MIDI driver:
+     - added JACK MIDI driver
+     - dispatch bank select as ordinary CC as well, the user might seriously
+       want to (mis)use it for some purpose ("fixed" in all current MIDI
+       input drivers: ALSA, CoreMIDI, JACK, MidiShare, MME)
+     - bugfix: pitch bend wasn't working with JackMidi, VST, LV2, MME,
+       CoreMidi or AU
+     - fixed mingw-w64 compilation error in MME driver
+     - made program change handling in MIDI thread real-time safe by
+       moving the logic to a non-RT thread
+     - fixed minor memory leak in ALSA driver
+   * instruments database:
+     - avoid time consuming samples scanning when adding instruments
+       to the instruments database
+     - added support for handling lost files in the instruments database
+     - Implemented option for adding instruments in separate directories
+       in the instruments database
+       (patch by Chris Cherrett & Andrew Williams, a bit adjusted)
+     - work-around for missing fnmatch function on Windows to make
+       instrument database compilable
+     - added instrument database support on Windows
+       works with both standalone mode and VST plugin.
+       the instruments DB file is located in
+       %USERPROFILE%\.linuxsampler\instruments.db which allows different
+       databases for each windows user
+       if no DB is present it automatically creates the .linuxsampler subdir
+       and then creates an empty DB
+     - fixed recursive import, which was broken on Windows
+   * Gigasampler format engine:
+     - fixed a memory leak that could happen when a channel was deleted
+       while notes were playing
+     - made it possible to load an instrument even if the audio thread
+       isn't running
+     - added partial support for the "Controller Triggered" MIDI rule,
+       enough for piano gigs with pedal noise samples
+     - minor fix: only mark FX sends as being modified if really the
+       respective FX send MIDI controller was used
+     - added support for GM global device volume SysEx message
+       (can be selected at compile time to either apply globally to the
+       whole sampler [default] or only to the sampler channels that are
+       connected to the respective MIDI input port the SysEx message arrived
+       on)
+     - bugfix: notes triggered at position 0 in the audio buffer were
+       sometimes wrongly killed in the same buffer, causing no sound to
+       be played
+     - added support for chromatic / drumkit mode Roland GS Sysex message
+       which is usally used to switch a part between chromatic sounds and
+       drumkit sounds (as opposed to their common default setting of MIDI
+       part 10 being a drumkit part and all other ones chromatic parts), in
+       LS however one can switch between the first 16 MIDI instrument maps
+       defined for the sampler
+     - made it possible to create multiple sample channels even if the
+       audio thread isn't running
+     - fixed crash when removing channel with active voices (#116)
+     - bugfix: on sample reference changes (instrument editor), only
+       un-cache the respective sample if it's really not used by any
+       sampler engine anymore
+     - re-cache samples in case they were changed by an instrument editor,
+       e.g. when a sample was added while playing (#82)
+     - fixed hanging notes which occured when note-off event had the exact
+       same time stamp as the voice's note-on event and occured both in the
+       same audio fragment cycle (fixes bug #112)
+     - added support for the "fine tune" and "pitch bend range"
+       instrument-level gig parameters
+     - fixed minor artifacts in pitch bend handling
+     - added support for GS Reset SysEx message
+     - allow gig files to use unlimited downward pitch shifting
+     - added a limit check for upward pitch shifting
+     - bugfix: sometimes, when playing a note twice fast, the second
+       note was silent
+     - fixed crash happening when a pitch bend event arrived at the
+       same time a new instrument was loading
+   * LSCP server:
+     - added new LSCP event "CHANNEL_MIDI" which can be used by frontends to
+       react on MIDI data arriving on certain sampler channels (so far only
+       Note-On and Note-Off events are sent via this LSCP event)
+     - added new LSCP event "DEVICE_MIDI" which can be used by frontends to
+       react on MIDI data arriving on certain MIDI input devices (so far only
+       Note-On and Note-Off events are sent via this LSCP event)
+     - added new LSCP commands: FIND LOST DB_INSTRUMENT_FILES and
+       SET DB_INSTRUMENT FILE_PATH
+     - added new LSCP command: SEND CHANNEL MIDI_DATA which can be used by
+       frontends to send MIDI messages to specific sampler channel
+     - added two additional fields to GET FILE INSTRUMENT INFO command -
+       KEY_BINDINGS and KEYSWITCH_BINDINGS
+     - bugfix: the bank number provided by MIDI_INSTRUMENT_INFO notifications
+       was incorrect
+     - Optimized the retrieval of the MIDI instrument mappings
+     - added new LSCP command "SET VOICES" to globally alter the maximum
+       amount of voices, added new LSCP event "GLOBAL_INFO:VOICES" which will
+       be triggered respectively
+     - added new LSCP command "SET STREAMS" to globally alter the maximum
+       amount of disk streams, added new LSCP event "GLOBAL_INFO:STREAMS"
+       which will be triggered respectively
+     - bugfix: retry if "select" returns EINTR (this fixes a crash when
+       a gigedit file dialog is opened)
+     - close all connections when LSCPServer is deleted
+     - hard close of all sockets on exit
+     - bugfix: SET CHANNEL MIDI_INPUT_TYPE didn't work with the MME
+       driver
+   * Bug fixes:
+     - fixed a crash which occurs when removing a sampler channel waiting
+       to start instrument loading after another channel
+     - fixed a crash which occurs when removing a sampler channel with
+       instrument loading in progress (bug #113)
+     - fixed termination caused by uncaught exception when adding MIDI
+       instrument with PERSISTENT load mode
+     - fixed possible iterator invalidations when resetting the sampler
+     - fixed memory leaks when issuing the following LSCP commands:
+       GET AUDIO_OUTPUT_DRIVER INFO
+       GET AUDIO_OUTPUT_DRIVER_PARAMETER INFO
+       GET MIDI_INPUT_DRIVER INFO
+       GET MIDI_INPUT_DRIVER_PARAMETER INFO
+     - fixed possible compilation error when sqlite is not present
+     - fixed orphaned pointers when setting maximum voices limit (bug #118)
+     - fixed crash when changing the audio output device of a sampler
+       channel with loaded instrument and start playing notes
+     - fixed endless loop in Engine::SuspendAll() (bug #120)
+     - fixed a low-level atomic load function that was broken on 64-bit
+       PowerPC, which probably could cause crashes on that platform
+     - fixed a memory management error which could cause a crash when a
+       plugin was unloaded
+     - bugfix: two private structs had the same name, which could cause
+       problems if the linker chose the wrong constructor
+     - fixed low-level ConditionServer usage bug that caused lockups on
+       Windows
+ Version 0.5.1 (6 December 2007)
+   * packaging changes:
+     - added autoconf checks for pthread library
+     - added autoconf check for pthread bug found on certain NPTL-enabled
+       glibc versions (see Gentoo bug report #194076)
+     - added autoconf checks for MS Windows
+   * general changes:
+     - bugfix: the thread used by an editor plugin didn't die when the
+       editor closed
+     - bugfix: Ringbuffer.h: fill_write_space_with_null() did not zero
+       out all the space. operator--() did not apply size_mask after
+       decrementing the read_ptr. DEFAULT_WRAP_ELEMENTS set to 0 in
+       order to avoid problems with the _NonVolatileReader functions.
+     - bugfix: Stream.h: added a missing divide by BytesPerSample in
+       GetWriteSpace(). Since this function is currently only used in
+       the stream's qsort() compare function, it didn't trigger any
+       bugs.
+     - Resampler.h, Synthesizer.h: cubic interpolation now works in
+bit mode too. Faster method to read 24bit words on little
+       endian machines (x86): replaced 3 byte reads + shifts with a 1
+       unaligned 32bit read and shift
+     - experimental support for MS Windows (MIDI input via MME, AUDIO
+       output via ASIO)
+     - made handling of SIGINT signal (Ctrl-C) a bit more robust
+     - support for monitoring the total number of active disk streams
+       (new LSCP commands: GET TOTAL_STREAM_COUNT,
+       SUBSCRIBE TOTAL_STREAM_COUNT, UNSUBSCRIBE TOTAL_STREAM_COUNT)
+   * AUDIO driver:
+     - added Windows ASIO low latency audio driver
+   * MIDI driver:
+     - added MME Windows MIDI driver
+   * LSCP server:
+     - added support for Windows style path / filenames, however with
+       forward slash path separators instead of backslash
+       (i.e. "C:/foo/bar.gig")
+     - allow naughty liblscp to send non-string device parameters within
+       apostrophes as well
+     - added new LSCP commands: "GET FILE INSTRUMENTS <file>",
+       "LIST FILE INSTRUMENTS <file>" and
+       "GET FILE INSTRUMENT INFO <file> <index>" for retrieving informations
+       about an arbitrary instrument file on the system where the sampler is
+       running on
+ Version 0.5.0 (15 October 2007)
+   * packaging changes:
+     - config.h is not going to be installed along with liblinuxsampler's
+       API header files anymore
+     - only the API relevant header (and source) files will be exported to
+       the Doxygen API documentation (explicitly listed in Doxyfile.in)
+     - added completely new XCode project files for Mac OSX which is now
+       capable to execute our autoconf environment, thus no need anymore
+       to maintain the compile time configuration file (osx/version.h) for
+       OSX manually (patch by Toshi Nagata)
+     - fixed buggy boolean --enable-foo / --disable-foo configure script
+       parameters
+     - global.h now only covers global definitions that are needed for the
+       C++ API header files, all implementation internal global definitions
+       are now in global_private.h
+     - atomic.h is not exposed to the C++ API anymore
+     - no need to include config.h anymore for using LS's API header files
+     - fixed warnings in API doc generation
+   * general changes:
+     - replaced the old, confusing MIDI program change mechanism by a
+       flexible MIDI instrument mapper which allows to map arbitrary
+       (MIDI bank, MIDI program) pairs with arbitrary
+       (engine type, instrument file, file index) triplets which will be
+       loaded on the respective channel when such MIDI program change
+       messages arrive, beside that, each entry allows to define a life-time
+       strategy for the instrument, so the user can define whether the
+       instrument should i.e. be permanently loaded (to be able to switch
+       quickly among sounds) or i.e. loaded just on demand when the
+       respective program change arrives (to safe RAM space), as well as a
+       global volume factor for each entry, so the user can adjust the volume
+       dynamic of his mapped instrument collection without having to modify
+       the instrument files, also one can manage arbitrary amount of such
+       MIDI instrument maps and assign each sampler channel individually
+       a certain map, so that i.e. separation of normal instruments and
+       drumkits is possible
+     - new notification events for tracking changes to audio/MIDI devices,
+       MIDI instrument maps, MIDI instruments, FX sends, global volume.
+     - sampler was limited to load max. 200 instruments in the background
+       due to a constant size RingBuffer FIFO which is now replaced by a
+       dynamic (unlimited) size std::list FIFO
+     - added FX sends, these allow to route audio signals to arbitrary audio
+       output channels for being processed by external effect processors
+       (i.e. jack-rack), the send levels are controllable via arbitrary MIDI
+       controllers
+     - global (sampler wide) volume can now be controlled at runtime
+     - Implemented new, improved notification system
+     - fixed compilation errors regarding OSX
+       (patch by Toshi Nagata)
+     - implemented instruments database
+     - added support for escape sequences to the instruments database
+     - added highly experimental support for on-the-fly instrument editing
+       within the sampler's process (by using instrument editor plugins),
+       you'll notice the new "Registered instrument editors:" message on
+       startup, added a new LSCP command:
+       "EDIT CHANNEL INSTRUMENT <sampler-channel>"
+       to spawn a matching instrument editor for the instrument on the
+       given sampler channel, the plugin path can be overridden at compile
+       time with ./configure --enable-plugin-dir=/some/dir
+     - added experimental code for synchronizing instrument editors hosted
+       in the sampler's process to safely edit instruments while playing
+       without a crash (hopefully) by either suspending single regions
+       wherever possible or if unavoidable whole engine(s)
+     - fixed several issues in fundamental "Thread" class: set scheduling
+       policy and priority on thread level, set a minimum stack size for
+       thread (TODO: a reasonable value yet to be tested), bugfix: non-RT
+       threads simply inherited properties of starting thread instead of
+       setting their own policy and priority
+     - minor fix in our automatic stack trace mechanism on crashes, the
+       main process did not wait for the stack trace process to finish
+       its output
+     - fixed some minor memory leaks
+     - reenabled assembly features support, at the moment only for
+       enabling a fast denormal FPU mode (x86 platforms supporting SSE2)
+     - minor assembly fix in x86 features detection (don't use the PIC
+       register, to avoid relocations in the text segment at runtime)
+     - POSIX callback functions of Thread.h are hidden
+   * MIDI driver:
+     - dispatch bank select (MSB and LSB) messages
+   * audio driver:
+     - the ALSA audio output driver parameters now reflect the correct
+       parameter value ranges for the respective selected sound card
+       (patch by Till Wimmer, a bit fixed and extended)
+   * Gigasampler format engine:
+     - if a filter is used and EG2 finishes before EG1, let the voice
+       die when EG2 has finished (this fixes a problem with clicks and
+       voice starvation for some gigs)
+     - playback is no longer disabled during instrument loading
+     - all notes playing on a channel that changes its instrument keep
+       playing with the old instrument until they get a note off
+       command
+     - EG fix: a release value of zero could cause noises or crash
+     - handle MIDI coarse tuning messages (MIDI RPN #0 MSB #2 LSB)
+     - EG fine tuning: when attack is zero the EG starts at a level
+       above max sustain level, which means that there is a decay phase
+       even if sustain is 100%
+     - more EG fixes: the level could sometimes go below zero and cause
+       noises or crashes
+     - minor fix of EGDecay (patch by Toshi Nagata)
+     - fixed compiler error when --enable-override-filter-type was
+       supplied to the configure script (fixes #46)
+     - disk thread: queue sizes are now proportional to CONFIG_MAX_STREAMS
+       instead of a fix value
+     - behavior fix: on MIDI CC# 65 (portamento on / off), 126 (mono mode),
+(solo mode) only kill voices if the respective mode really
+       changed
+   * LSCP server:
+     - fixed compile time error for old Bison versions
+       (i.e. v1.28 found on MacOS 10.4, patch by Toshi Nagata)
+     - parser now supports extended ASCII character set
+       (up to ASCII code 255, i.e. includes now umlauts and accents)
+     - filename arguments in LSCP commands now allow to use escape
+       sequences, that is directly literal as one of: \', \", \\, \n, \r,
+       \f, \t, \v, or as octal ASCII code value like \132, or as hex ASCII
+       code value like \xf2) (fixes bug #24)
+     - the following LSCP commands now also support escape sequences for at
+       least one of their text-based parameters (i.e. name, description):
+       "ADD MIDI_INSTRUMENT_MAP", "MAP MIDI_INSTRUMENT",
+       "SET MIDI_INSTRUMENT_MAP NAME", "SET FX_SEND NAME", "CREATE FX_SEND",
+       "SET DB_INSTRUMENT_DIRECTORY NAME",
+       "SET DB_INSTRUMENT_DIRECTORY DESCRIPTION", "SET DB_INSTRUMENT NAME",
+       "SET DB_INSTRUMENT DESCRIPTION", "FIND DB_INSTRUMENTS",
+       "FIND DB_INSTRUMENT_DIRECTORIES"
+     - returns verbose syntax errors (line and column where syntax error
+       occured, the unexpected character and the actually expected, possible
+       character(s), the latter only if less than 5 possibilities)
+     - made sure that LSCP syntax is not affected by gigedit locale
+       settings
+     - bugfix regarding strings parameter lists: all comma separated lists
+       of strings were treated as being one string containing commas
+       (fixes #57)
+   * Bug fixes:
+     - fixed crash occurring on certain LSCP scripts (Bug 39)
+     - another thread safety fix for lscp "load engine" and "set
+       channel audio output device"
+     - fixed a crash which occurs when reassigning the same engine
+       on a sampler channel with connected MIDI device
+     - fixed a crash which occurs when changing the number of ports of a MIDI
+       device connected to a sampler channel to number less then or equal
+       to the index of the port to which the sampler channel is connected.
+     - The previous bindings were not been disconnected when altering
+       the ALSA_SEQ_BINDINGS parameter. Introduced a NONE keyword for
+       unsubscribing from all bindings (e.g. ALSA_SEQ_BINDINGS=NONE).
+     - The active stream/voice count statistic was incorrect.
+     - notification events were not been sent for some sampler
+       channel changes
+     - added default min and max values to restrict the number of allowed
+       audio output channels and MIDI input ports
+     - the connection to the PCM interface is now closed when destroying
+       an audio output device
+     - files with slash in their path or filename could not be loaded
+   * test cases:
+     - updated and fixed (haven't been touched in a while)
+ Version 0.4.0 (24 November 2006)
+   * packaging changes:
+     - changed deprecated copyright attribute to license; added ldconfig
+       to post-(un)install steps; added devel package for liblinuxsampler;
+       to linuxsampler.spec (RPM)
+     - install necessary development header files for allowing 3rd party
+       applications to link against liblinuxsampler
+     - liblinuxsampler's API documentation can be generated with 'make docs'
+       (Doxygen required)
+     - added benchmark to automatically detect the best triangle LFO
+       implementation (currently either an integer math solution or a
+       di-harmonic approximation), automatic detection can be overriden
+       with --enable-signed-triang-algo=x and --enable-unsigned-triang-algo=x
+       configure script argument though (mandatory for cross-compilation)
+     - do not automatically pick optimized gcc flags if the user already
+       provided some on his own (as CXXFLAGS)
+     - added compile time option to disable processing of All-Notes-Off MIDI
+       messages
+     - added compile time options to allow disabling the various audio and
+       MIDI drivers
+     - fixed automatic GCC CPU switch detection on PPC
+       (patch by Ebrahim Mayat)
+   * Gigasampler format engine:
+     - extensive synthesis optimization
+       (reimplementation of EGs and LFO(s), removed synthesis parameter
+       prerendering and the synthesis parameter matrix in general, splitting
+       each audio fragment into subfragments now where each subfragment uses
+       constant pitch and filter coefficients. The volume coefficient is
+       linearly interpolated inside a subfragment, unless
+       --disable-interpolate-volume is set.)
+     - fine tuning of the EG modulation parameters
+     - improved filter cutoff calculation by adding support for the
+       following gig parameters: Cutoff freq (used when no cutoff
+       controller is defined), Control invert, Minimum cutoff, Velocity
+       curve and Velocity range. The keyboard tracking now scales
+       cutoff frequency instead of resonance.
+     - added support for gig parameter Resonance.
+     - fixed bug in sysex handling (patch by Juan Linietsky)
+     - added global volume attenuation of -9 dB (0.35f) to prevent clipping
+       which can be overridden with --enable-global-attenuation
+     - EG fixes: made the length of "attack hold" stage more
+       accurate. Release stage can now start before attack stage
+       ends. Cancel release didn't work when sustain was zero. Attack
+       time now has a minimal value to prevent clicks.
+     - fixed pitch changes larger than one octave
+     - fixed EG3 (pitch envelope) synthesis which was neutral all the time
+     - implemented portamento mode and solo mode (a.k.a 'mono mode'):
+       all modes can be altered via standard GM messages, that is CC5 for
+       altering portamento time, CC65 for enabling / disabling portamento
+       mode, CC126 for enabling solo mode and CC127 for disabling solo mode
+     - fine tuning of the curves for volume (CC7), pan (CC10 and gig
+       parameter) and crossfade
+     - added support for the "attenuation controller threshold" gig
+       parameter
+     - added smoothing of volume changes caused by control change
+       messages
+     - sample loop parameters are now taken from the DimensionRegion
+       instead of the wave chunk
+     - fixed keyswitching for v3 gigs with a number of keyswitch splits
+       not equal to a power of two
+     - reimplementation of the filter algorithm. The new filter is more
+       accurate and supports all gig filter types, including bandreject
+       and lowpass turbo.
+     - real support for 24 bit samples - samples are not truncated to
+bits anymore
+     - support for aftertouch (channel pressure, not polyphonic
+       aftertouch)
+     - LFO1 behaviour fixed (dampening from max volume instead of
+       amplifying from 0)
+   * LSCP server:
+     - fixed application exit on broken pipe error (fixes bug #20)
+     - fixed the notification messages delay due to lack of
+       network activity (fixes bug #26)
+     - fixed parser bug which occured on space(s) within device parameters
+       (e.g. "SET AUDIO_OUTPUT_CHANNEL_PARAMETER 0 0 SOMEPARAM='foo bar'")
+   * audio driver:
+     - added aRts audio output driver (by no means RT safe)
+   * MIDI driver:
+     - fixed legacy sysex code which caused dispatching of MIDI SysEx
+       messages several times instead of once
+     - API extension for MIDI drivers which already supply exact time stamps
+       for events (i.e. for offline rendering based MIDI drivers)
+     - added checks for bad MIDI data
+   * linuxsampler application:
+     - show available sampler engine types on startup
+   * general changes:
+     - support for muting sampler channels and solo mode of the same, two new
+       LSCP commands ("SET CHANNEL MUTE" and "SET CHANNEL SOLO") and two new
+       fields ("MUTE" and "SOLO") for command "GET CHANNEL INFO" were
+       introduced for this, the behavior is the same like on a mixer console
+       (patch by Grigor Iliev, a bit adjusted). Also added configure option
+       --enable-process-muted-channels which can be used to enable the
+       processing of muted channels.
+     - support for sostenuto pedal
+     - support for monitoring the total number of active voices
+       (new LSCP commands: GET TOTAL_VOICE_COUNT, GET TOTAL_VOICE_COUNT_MAX,
+       SUBSCRIBE TOTAL_VOICE_COUNT, UNSUBSCRIBE TOTAL_VOICE_COUNT)
+     - fixed some memory management errors
+     - fixed some concurrency problems that could lead to crashes when
+       LSCP commands were executed
+     - fixed crash when instrument loading failed previously on the same
+       sampler channel (fixes bug #36)
+ Version 0.3.3 (15 July 2005)
+   * packaging changes:
+     - fixed compilation with gcc 4.0
+   * Gigasampler format engine:
+     - LFO preprocessing was performed even though the respective LFO was
+       disabled by the instrument patch (minor efficiency fix)
+     - if period time of chosen audio device is too small (< MIN_RELEASE_TIME)
+       for volume ramp downs in the same fragment (needed for current voice
+       stealing implementation) then simply inform the user with a warning
+       about possible click sounds and reduce the volume ramp down
+       appropriately instead of cancelling the audio device connection
+ Version 0.3.2 (24 June 2005)
    * packaging changes:
      - updated autotools build files to compile on OS X
        (conditional compilation of CoreMIDI and MidiShare drivers)
+     - hand-crafted assembly optimization code can be disabled with
+       './configure --disable-asm' (definitely not recommended)
+     - fixed 'make dist' rule to include all necessary files
+     - require automake (>= 1.5) for 'make -f Makefile.cvs'
+       (due to 'dist-bzip2' automake option)
+   * Gigasampler format engine:
+     - support for the gig parameters for "release velocity response" curves
+     - fine tuning of the EGADSR envelope
+     - volume of release triggered samples now depends on note-on velocity,
+       note length and gig parameter "release trigger decay" instead of
+       note-off velocity.
+     - revised voice stealing
+       (fixes crash and endless loop caused by voice stealing)
+     - don't reset scale tuning on instrument or audio output device change
+     - handle key group conflicts right at the beginning of each fragment
+       instead of when the respective voice is actually launched
+       (fixes undefined behavior if stolen voices belonged to a key group -
+       this case was followed by a "killed voice survived" error message)
+     - fixed minor issue with null/silence samples
+       (those stole voices even though they don't need a voice at all which
+       resulted in "voice stealing didn't work out" messages)
+     - don't reset volume, pan, pitch and MIDI controller values on
+       instrument or audio output device change
+   * LSCP server:
+     - fixed some crashes (patch by Grigor Iliev, fixes #19)
+     - fixed LSCP event "CHANNEL_INFO" notification
+       (e.g. did not notify on volume changes or MIDI program change events)
+   * linuxsampler application:
+     - added command line parameters --lscp-addr and --lscp-port to override
+       default IP address and port of LSCP server
  Version 0.3.1 (24 May 2005)

 Legend:



Removed from v.609
 


changed lines


 
Added in v.3646
 Legend:



Removed from v.609
 


changed lines


 
Added in v.3646
-Removed from v.609
+Added in v.3646

	ViewVC Help
Powered by ViewVC