| When should I schedule to make a release branch?
On Fri, May 13, 2016 at 7:10 AM, jpff wrote:
> 16 May 2016
>
>
> On Fri, 13 May 2016, Victor Lazzarini wrote:
>
>> Which Monday?
>> ========================
>> Dr Victor Lazzarini
>> Dean of Arts, Celtic Studies and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>>
>>> On 13 May 2016, at 11:37, jpff wrote:
>>>
>>> I meant 16 may (this keyboard is a pain).
>>> The Tuesday was in case anyone had an issue with Monday.
>>>
>>> Let us release MONDAY
>>> ==John
>>>
>>>
>>> On Fri, 13 May 2016, Victor Lazzarini wrote:
>>>
>>>> Are we agreed on Monday 6 or Tues 17?
>>>> ========================
>>>> Dr Victor Lazzarini
>>>> Dean of Arts, Celtic Studies and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>>
>>>>> On 12 May 2016, at 15:55, jpff wrote:
>>>>>
>>>>> After this delay it seems that we should be able to release next week
>>>>> -- say Monday 6 May or Tuesday 17 May. Are there any objections?
>>>>>
>>>>> Latest version of release notes appended below
>>>>>
>>>>> ==John ffitch
>>>>>
>>>>> ========================================================================
>>>>>
>>>>> ============================
>>>>> CSOUND VERSION 6.07
>>>>> RELEASE NOTES VERSION 6.07
>>>>> ============================
>>>>>
>>>>> A large number of bug fixes, some quite major, some internal are
>>>>> included, as well as some new facilities and extensions. As ever there
>>>>> are coding improvements as well.
>>>>>
>>>>> -- The Developers
>>>>>
>>>>>
>>>>> USER-LEVEL CHANGES
>>>>> ==================
>>>>>
>>>>> New opcodes:
>>>>>
>>>>> o compress2: like compress but using a more normal use of dB
>>>>> (0.0 for full scale).
>>>>>
>>>>> o (Experimental, source code-only) New cuda opcodes cudasynth2
>>>>> and cudanal2
>>>>>
>>>>> o directory opcode: Reads a directory and outputs to a string
>>>>> array a list of file names.
>>>>>
>>>>> o ftsamplebank:to load a sample library from a directory.
>>>>>
>>>>> o mvclpf1, mvclpf2, mvclpf3, mvclpf4, mvchpf: Moog
>>>>> voltage-controlled filter emulations from Fons Andriaensen.
>>>>>
>>>>> o S() converter from k-rate and i-time number to a string.
>>>>>
>>>>> o cepsinv opcode to calculate the inverse cepstrum of an array.
>>>>>
>>>>> o moogladder2 is a faster, less accurate implementation of
>>>>> moogladder.
>>>>>
>>>>> o paulstretch opcode is a lightweight implementation of the
>>>>> PaulStretch time-stretching algorithm by Paul Nasca
>>>>> Octavian. It is ideal for timestretching a signal by very
>>>>> large amounts.
>>>>>
>>>>> o mp3scal implements phase-locked vocoder processing from
>>>>> mp3-format disk files, resampling if necessary.
>>>>>
>>>>> o filescal implements phase-locked vocoder processing from disk
>>>>> files, resampling if necessary.
>>>>>
>>>>> New Gen and Macros:
>>>>>
>>>>> o NONE
>>>>>
>>>>> Orchestra:
>>>>>
>>>>> o The boolean ?..:.. construction can now have string results,
>>>>> while previously it only allowed numbers.
>>>>>
>>>>> o The line number reported when an error is detected at the end
>>>>> of a line is now correct. Also more improvements in line
>>>>> numbers
>>>>> for complex syntax
>>>>>
>>>>> o While loop improved/fixed.
>>>>>
>>>>> o Better and consistent reading of comments.
>>>>>
>>>>> o Continuation lines handled better, especially respecting line
>>>>> numbers.
>>>>>
>>>>> o opcode:S(.) syntax now allowed.
>>>>>
>>>>> Score:
>>>>>
>>>>> o The list of tempo points in score opcode t is now arbitrarily
>>>>> long.
>>>>>
>>>>> o A stupidity in r and { opcodes fixed.
>>>>>
>>>>> Options:
>>>>>
>>>>> o The -z option now suppresses deprecated opcodes, unless given
>>>>> a 2 or 3 argument.
>>>>>
>>>>> o The new option --fftlib controls which real fft library to
>>>>> use internally (FFTLIB = 0, PFFFT = 1, vDSP = 2)
>>>>>
>>>>>
>>>>>
>>>>> Modified Opcodes and Gens:
>>>>>
>>>>> o In OSC opcodes it is now possible to send and receive arrays,
>>>>> tables and audio.
>>>>>
>>>>> o Better diagnostic if diskin2 fails.
>>>>>
>>>>> o Rezzy now checks for unstable filter and modifies to close
>>>>> stable version.
>>>>>
>>>>> o adsr rewritten so it gives an error if the segments are
>>>>> longer than p3.
>>>>>
>>>>> o Use of diskin to an array ow resizes the output array if
>>>>> necessary.
>>>>>
>>>>> o chnget now checks for a change in chanel name as well as data.
>>>>>
>>>>> o interp can take an optional extra argument to give an initial
>>>>> value.
>>>>>
>>>>> o oscilikts uses a more liberal table lookup.
>>>>>
>>>>> o opcode in can read mono or stereo, obviating the in/ins
>>>>> distinction.
>>>>>
>>>>> o sensekey rewritten to provide better diagnostics and fixes a
>>>>> minor bug.
>>>>>
>>>>> o Fix to a case of defining a macro with arguments.
>>>>>
>>>>> o sockrecv now works at a and k rate.
>>>>>
>>>>> o GEN49 now works from ftgen calls.
>>>>>
>>>>> o GEN34 liberalised in tables it accepts.
>>>>>
>>>>> o chnget now allows channel names to be changed at perf time.
>>>>>
>>>>> o iceps has been renamed cepsinv to avoid name clashes.
>>>>>
>>>>> o mp3 support improved in a number of minor ways.
>>>>>
>>>>> o a minor fix to allow a ladspa plugin to be re-loaded by the
>>>>> host.
>>>>>
>>>>>
>>>>> Utilities:
>>>>>
>>>>> o
>>>>>
>>>>> Frontends:
>>>>>
>>>>> pnacl:
>>>>>
>>>>> icsound:
>>>>>
>>>>> csound~:
>>>>>
>>>>> Emscripten:
>>>>>
>>>>> csdebugger:
>>>>>
>>>>> * Some memory issues fixed.
>>>>>
>>>>> HTML5:
>>>>>
>>>>> CsoundQT has its own notes at
>>>>>
>>>>> https://github.com/CsoundQt/CsoundQt/blob/master/release_notes/Release%20notes%200.9.2.md
>>>>>
>>>>> General usage:
>>>>>
>>>>> o The multicore options somewhat improved.
>>>>>
>>>>> o When replacing instruments the new version inherits maxalloc
>>>>> and active flags.
>>>>>
>>>>> o Multicore code now works with midi instruments.
>>>>>
>>>>> o MIDI operations now available via a new rtmidi jack module
>>>>> (-+rtmidi=jack)
>>>>>
>>>>>
>>>>> Bugs fixed:
>>>>>
>>>>> o Fix to trigseq.
>>>>>
>>>>> o Major error in rezzy fixed.
>>>>>
>>>>> o p() fixed for high numbers of p-arguments.
>>>>>
>>>>> o p() now works from MIDI events.
>>>>>
>>>>> o The 31-bit random number generator could give a very short
>>>>> loop if seeded with zero; fixed.
>>>>>
>>>>> o Macros in .orc files now work.
>>>>>
>>>>> o A long-standing bug in display of graphs fixed.
>>>>>
>>>>> o Fixes to the envelope opcodes linen, expsegr, linsegr,
>>>>> cossegr, transegr, envlpx, including new warnings.
>>>>>
>>>>> o Many fixes in string opcodes (strsub, strcpy and others).
>>>>>
>>>>> o Fixed bugs in print formats (sprintf printf prints).
>>>>>
>>>>> o pvsmooth had a missing initialisation.
>>>>>
>>>>> o Array initialisation now robust.
>>>>>
>>>>> o Bug in copya2ftab fixed.
>>>>>
>>>>> o Fix in cudapvsanal (EXPERIMENTAL).
>>>>>
>>>>> o partikkel fix to subsample grain placement.
>>>>>
>>>>> o Opcodes that require an odd number of arguments are now properly
>>>>> checked.
>>>>>
>>>>> o pvswrite now takes proper notice of 0dbfs.
>>>>>
>>>>> o GEN34 now allowing non-power-of-two source tables.
>>>>>
>>>>> o vst4cs will now work on 64bit architecture.
>>>>>
>>>>> o strcat mended.
>>>>>
>>>>> o nstance mended.
>>>>>
>>>>> o fixed small issues in rtauhal module.
>>>>>
>>>>> o Windows installer for 64 bit CPU architecture now sets
>>>>> environment variables on system level as it should.
>>>>>
>>>>> o transeg at k-rate with itype non-zero had an initial delay if one
>>>>> cycle
>>>>>
>>>>> o log function on arrays fixed
>>>>>
>>>>> o Rescaling of named GENs fixed
>>>>>
>>>>> o A fencepost error in reading ATS files believed fixed; it
>>>>> generated spurious errors about file finishing early.
>>>>>
>>>>> o Initialisation error in pvbandp fixed.
>>>>>
>>>>> o readfi string allocation fixed
>>>>>
>>>>> ====================
>>>>> SYSTEM LEVEL CHANGES
>>>>> ====================
>>>>>
>>>>> System changes:
>>>>>
>>>>> o Extensive use made of in-memory files. Orc and sco are always
>>>>> thus, as are .csound6rc. This simplifies much internally.
>>>>>
>>>>> o Nested longjmps fixed which affected some API use.
>>>>>
>>>>> o If a non-existent .csd file is presented Csound no longer
>>>>> crashes.
>>>>>
>>>>> o The kcounter value is now consistently an unsigned 64bit
>>>>> value, allowing for long performances.
>>>>>
>>>>>
>>>>> API
>>>>> ===
>>>>>
>>>>> * csoundCompileCsdText: New function to compile a CSD from a string
>>>>> of text.
>>>>>
>>>>> * The function call csound->GetKcounter(csound) now returns an
>>>>> unsigned 64 bit integer. Previously it returned a long which is
>>>>> unclear.
>>>>>
>>>>> * ctcsound.py is a new wrapper file to the Csound API for the
>>>>> Python language. It is a pure Python file using the FFI (Fast
>>>>> Foreign Interface) module ctypes. It does not depends on
>>>>> Swig, and it works with Python2 and Python3 as well. It uses
>>>>> numpy for its data structures, numpy being the fundamental
>>>>> package for scientific computing with Python.
>>>>>
>>>>> Platform Specific
>>>>> =================
>>>>>
>>>>> iOS
>>>>> ---
>>>>>
>>>>> *
>>>>>
>>>>>
>>>>> Android
>>>>> -------
>>>>>
>>>>> *
>>>>>
>>>>> Windows
>>>>> -------
>>>>>
>>>>> * The Windows installer for Csound now contains all executables
>>>>> built for 64 bit CPU architecture.
>>>>>
>>>>> * The Windows installer for 64 bit CPU architecture now provides
>>>>> NW.js
>>>>> "out of the box." It runs all features of
>>>>> Csound, as well as all features of HTML5, in a browser-like runtime
>>>>> with JavaScript scripting, and includes a Csound editor implemented
>>>>> in HTML. The editor will run pieces either as .csd files with
>>>>> embedded
>>>>> or as .html files calling the csound object, and includes
>>>>> JavaScript debugging.
>>>>>
>>>>> * A new simplified build system has been implemented in the
>>>>> csound/mingw64
>>>>> directory for building Csound using the MSYS2/mingw64 toolchain.
>>>>>
>>>>> * Linking to pthreads and other libraries now static.
>>>>>
>>>>> OSX
>>>>> ---
>>>>>
>>>>> * Installer fixed for link names for _csnd6 and _CsoundAC.
>>>>>
>>>>> GNU/Linux
>>>>> ---------
>>>>>
>>>>> * date opcode more accurate.
>>>>>
>>>>>
>>>>> ========================================================================
>>>>>
>>>>>
>>>>> ========================================================================
>>>>> Date: Tue May 10 14:49:44 2016 +0100
>>>>
>>>>
>> |