[Cs-dev] Towards the next release
Date | 2013-12-30 13:00 |
From | john ffitch |
Subject | [Cs-dev] Towards the next release |
We have planned the next release for Monday 6 Jan 2014 and to that end I am trying to create release notes. The main sources are the git log and my memory. I have described some changes but some are too cryptic for me and so I do not understand. So please if you have any suggestions, corrections, new information... please let me know on or off list. ==John ffitch ------------------------------------------------------------------------ TO BE ADDED TO RELEASE NOTES BELOW ================================== Fixed input device name for auhal xtratim opcode was using incorrect ekr value from csound instead of from instance; when used in conjunction with setksmps, was causing notes to have very long xtratim set and thus notes were effectively not getting turned off Better channel test. new nacl funcs array types for signalflowgraph. fixing merging of arrays Better string channel test. Minor change to alwayson. added support for playing CSDs directly in the pnacl module Fixes for vst4cs. allowing for cors access re-fixed python interface generation to not link to Python (allows loading of interface lib in any python) changes to file copy to local Fixed crash where no csoundSetHostImplementedMIDIIO is used on iOS and no _RTMIDI value is set fixed kcounter in UDOs with no local ksmps added experimental cuda opcodes to the repo fixed crashing bug with invalue channel callback due to wrong data object being pulled from csound host data ------------------------------------------------------------------------ ============================ CSOUND VERSION 6.02 RELEASE NOTES ============================ This new version has a large number of bug fixes (including clearing all general tickets on SourceForge). It also introduces some major new facilities such as use as a server, code to run Csound in a browser and a large generalisation of filter opcodes to have parameters changeable at audio rate. USER-LEVEL CHANGES ================== New opcodes: o instance opcode (**** UNDOCUMENTED ****) o kill opcode (**** UNDOCUMENTED ****) o strfromurl to set a string from a URL New Gen and Macros: None Orchestra: o If building supports it, a #include string can be a URL or a file o A space is again permitted between a function name and the opening bracket for all functions allowed in Csound5 (but not in general) o Csound can start with an empty orchestra Score: o If building supports it, a #include string can be a URL or a file Modified Opcodes and Gens: o Many filters generalised to allow k- or a-rate parameters. In particular it includes these: atonex butterworth filters lowres lpf18 mode moogladder tonex o The maximum number of presets in sfont increased to 16384 o cpsmidinn is now more accurate o max_k now behaves like the documentation. There were cases when it gave strange results. Utilities: None Frontends: o General usage: o o o o Bugs fixed: o adsynt2 opcode fixed o ftgentmp opcode fixed o dates opcode fixed o fixed a bug in pvsfilter o fixed stereo out in temposcal and mincer o pan2 opcode fixed o index overflow in randh and randi fixed o a number of fixes to CsoundVST o fixed pycalln for no inputs o fixed/revised setting and use of ksmps and kr in UDOs o fixed problem in sending a score event from max to csound vis csound~ (Ticket #58) o If itype in chn_k was set to 3 and values are set less than 1, Csound6 used to give an INIT Error. (Ticket #67) o A number of reported segfaults have been dealt with o o o o ==================== SYSTEM LEVEL CHANGES ==================== System changes: o A server mode is now available, accepting input via UDP o A longstanding bug in extract was detected and fixed. It does suggest that this facility is little used! o The way the external score generator was coded is substantially changed. In particular this should fix a very strange bug in Windows. o (Experimental) If the environment variable CS_UDO_DIR is set then any files in the directory that have an .udo extension are automatically included at the start of the orchestra. This facility needs review to see if it is what is required. Internal changes: o o o o API === Platform Specific ================= iOS --- Windows ------- OSX --- Linux ----- ======================================================================== ------------------------------------------------------------------------------ Rapidly troubleshoot problems before they affect your business. Most IT organizations don't have a clear picture of how application performance affects their revenue. With AppDynamics, you get 100% visibility into your Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net |
Date | 2013-12-30 13:46 |
From | Michael Gogins |
Subject | Re: [Cs-dev] Towards the next release |
Attachments | None None |
Here is more detail on some of my changes. The signal flow graph opcodes - started implementing use of a-rate array variables, but this is not finished. I may still be able to finish. this.
The vst4cs opcodes - factor out FLTK code, updated build system to get things working again for Csound 6. CsoundVST - updated Csound initialization, score handling, and MIDI driver initialization to get things working again for Csound 6.
The signal flow graph opcodes - replaced OpenMP multi-threading with pthreads, using one-time initialization of static structures. The alwayson opcode - better handling of pfields, more reliable insert of the instrument instance for repeating or re-started score sections. The cpsmidinn opcode - more precise calculation. Regards, Mike ----------------------------------------------------- Michael GoginsIrreducible Productions http://michaelgogins.tumblr.com Michael dot Gogins at gmail dot com On Mon, Dec 30, 2013 at 8:00 AM, john ffitch <jpff@codemist.co.uk> wrote: We have planned the next release for Monday 6 Jan 2014 and to that end |
Date | 2013-12-30 13:58 |
From | Felipe Sateler |
Subject | Re: [Cs-dev] Towards the next release |
On Mon, Dec 30, 2013 at 10:00 AM, john ffitch |
Date | 2013-12-30 15:27 |
From | Victor Lazzarini |
Subject | Re: [Cs-dev] Towards the next release |
There is a bug in the software bus with string channels, which makes chnexport unusable for strings. I guess we need to fix this before the release. Victor On 30 Dec 2013, at 13:00, john ffitch |
Date | 2013-12-30 16:00 |
From | Michael Gogins |
Subject | Re: [Cs-dev] Towards the next release |
Attachments | None None |
This has been bothering me, and fixing it would be great. Is it just strings? -----------------------------------------------------
Michael GoginsIrreducible Productions http://michaelgogins.tumblr.com Michael dot Gogins at gmail dot com On Mon, Dec 30, 2013 at 10:27 AM, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote: There is a bug in the software bus with string channels, which makes chnexport unusable for strings. I guess |
Date | 2013-12-30 17:45 |
From | Steven Yi |
Subject | Re: [Cs-dev] Towards the next release |
I'm just back from a family vacation. I had done some looking into this and think I'll have some time to look at this today. I certainly want to have this fix for 6.02 as it affects Blue users. On Mon, Dec 30, 2013 at 10:27 AM, Victor Lazzarini |
Date | 2013-12-30 18:11 |
From | Victor Lazzarini |
Subject | Re: [Cs-dev] Towards the next release |
Just strings, as far as I know. On 30 Dec 2013, at 16:00, Michael Gogins |
Date | 2013-12-31 14:32 |
From | jpff |
Subject | Re: [Cs-dev] Towards the next release |
Attachments | None None |
Thanks Mike; incorporated. There will be an opportunity for yu to review soon. ==John ff On Mon, 30 Dec 2013, Michael Gogins wrote: > Here is more detail on some of my changes. > The signal flow graph opcodes - started implementing use of a-rate array > variables, but this is not finished. I may still be able to finish. this. > > The vst4cs opcodes - factor out FLTK code, updated build system to get things > working again for Csound 6. > > CsoundVST - updated Csound initialization, score handling, and MIDI driver > initialization to get things working again for Csound 6. > > The signal flow graph opcodes - replaced OpenMP multi-threading with pthreads, > using one-time initialization of static structures. > > The alwayson opcode - better handling of pfields, more reliable insert of the > instrument instance for repeating or re-started score sections. > The cpsmidinn opcode - more precise calculation. > > Regards, > Mike > > > ----------------------------------------------------- > Michael Gogins > Irreducible Productions > http://michaelgogins.tumblr.com > Michael dot Gogins at gmail dot com > > > On Mon, Dec 30, 2013 at 8:00 AM, john ffitch |
Date | 2013-12-31 14:33 |
From | jpff |
Subject | Re: [Cs-dev] Towards the next release |
On Mon, 30 Dec 2013, Felipe Sateler wrote: > On Mon, Dec 30, 2013 at 10:00 AM, john ffitch |
Date | 2014-01-01 01:06 |
From | Andres Cabrera |
Subject | Re: [Cs-dev] Towards the next release |
Attachments | None None |
Thanks John,
On Mon, Dec 30, 2013 at 5:00 AM, john ffitch <jpff@codemist.co.uk> wrote: I have described some changes but some are too cryptic I think we should all make a point of writing clear commit messages, as they are not meant for us but for John who has to write the Changelog :)
Cheers, Andrés ==John ffitch |
Date | 2014-01-01 12:20 |
From | Victor Lazzarini |
Subject | Re: [Cs-dev] Towards the next release |
The instance and kill opcodes are now documented. I have a some edits: instead of "Csound can start with an empty orchestra” (which it can since 6.00, but not the command-line program): - The Csound command can start with an empty CSD in daemon mode (--daemon daemon mode: do not exit if CSD/orchestra is not given, is empty or does not compile) ----- - A server mode is now available, accepting input via UDP (with --port option). —— With regards to commit messages, I can clarify these: new nacl funcs: this is to do with the NaCl platform, which you already noted in a different email. added support for playing CSDs directly in the pnacl module: ditto changes to file copy to local: ditto fixed kcounter in UDOs with no local ksmps: - fixed bug in UDOs with no local ksmps where kcounter was being used incorrectly. added experimental cuda opcodes to the repo: - New experimental cuda GPGPU opcodes (source only): cudasynth (3 versions for additive synthesis, additive synthesis of fsigs and phase vocoder resynthesis) and cudanal (a GPGPU version of pvsanal). Victor On 30 Dec 2013, at 13:00, john ffitch |
Date | 2014-01-01 13:06 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Cs-dev] Towards the next release |
Attachments | None |