I've noticed that for many parameters on different plugins that there is no change event which is raised when the parameter is modified by the user. I've been reviewing the code trying to understand if this is intentional or a bug but I'm still not sure - I believe it has to do with slider type. This is important for my use of the YSFX library so that I can react to the user changes as the plugin host

Examples:

• Lava verb: https://github.com/JoepVanlier/JSFX
  -- The sliders in the "Ring" work notify host however the options in the middle do not
• Atlantis reverb: https://github.com/geraintluff/jsfx
  -- None of these raise changes

I'd be happy to make the changes on my side if you could advise and point me in the right direction.

Thanks!

The slider functions are valid to call from @gfx.

• make the operation atomic.
• handle the reset of change bits properly
• change API as necessary

Simple math functions

• sin(angle)
• cos(angle)
• tan(angle)
• asin(x)
• acos(x)
• atan(x)
• atan2(x,y)
• sqr(x)
• sqrt(x)
• pow(x,y)
• exp(x)
• log(x)
• log10(x)
• abs(x)
• min(x,y)
• max(x,y)
• sign(x)
• rand(x)
• floor(x)
• ceil(x)
• invsqrt(x)

Time functions

• time([v])
• time_precise([v])

Special variables

• spl0, spl1 ... spl63
• spl(channelindex)
• slider1, slider2, ... slider64
• slider(sliderindex)
• slider_next_chg(sliderindex,nextval)
• trigger
• srate
• num_ch
• samplesblock
• tempo
• play_state
• play_position
• beat_position
• ts_num
• ts_denom
• ext_noinit
• ext_nodenorm
• reg00-reg99
• _global.*
• pdc_delay
• pdc_bot_ch, pdc_top_ch
• pdc_midi

MIDI functions

• midisend(offset,msg1,msg2)
• midisend(offset,msg1,msg2 + (msg3 * 256))
• midisend(offset,msg1,msg2,msg3)
• midisend_buf(offset,buf, len)
• midisend_str(offset,string)
• midirecv(offset,msg1,msg23)
• midirecv(offset,msg1,msg2,msg3)
• midirecv_buf(offset,buf, maxlen)
• midirecv_str(offset, string)
• midisyx(offset,msgptr,len)
• ext_midi_bus
• midi_bus

File I/O and Serialization

• file_open(index or slider)
• file_close(handle)
• file_rewind(handle)
• file_var(handle,variable)
• file_mem(handle,offset, length)
• file_avail(handle)
• file_riff(handle,nch,samplrate)
• file_text(handle,istext)
• file_string(handle,str)

Memory/Slider/FFT/MDCT Functions

• mdct(start_index, size), imdct(start_index, size)
• fft(start_index, size), ifft(start_index, size)
• fft_real(start_index, size), ifft_real(start_index, size)
• fft_permute(index,size), fft_ipermute(index,size)
• convolve_c(dest,src,size)

Memory Utility

• freembuf(top)
• memcpy(dest,source,length)
• memset(dest,value,length)
• __memtop()

Stack

• stack_push(value)
• stack_pop(value)
• stack_peek(index)
• stack_exch(value)

Slider Functions

• sliderchange(mask or sliderX)
• slider_automate(mask or sliderX)
• slider_show(mask or sliderX[, value])

Atomic Variable Access

• atomic_setifequal(dest,value,newvalue)
• atomic_exch(val1,val2)
• atomic_add(dest_val1,val2)
• atomic_set(dest_val1,val2)
• atomic_get(val)

Pin Mapper Functions

• get_host_numchan()
• set_host_numchan(numchan)
• get_pin_mapping(inout,pin,startchan,chanmask)
• set_pin_mapping(inout,pin,startchan,chanmask,mapping)
• get_pinmapper_flags(no parameters)
• set_pinmapper_flags(flags)

String functions

• strlen(str)
• strcpy(str, srcstr)
• strcat(str, srcstr)
• strcmp(str, str2)
• stricmp(str, str2)
• strncmp(str, str2, maxlen)
• strnicmp(str, str2, maxlen)
• strncpy(str, srcstr, maxlen)
• strncat(str, srcstr, maxlen)
• strcpy_from(str,srcstr, offset)
• strcpy_substr(str,srcstr, offset, maxlen)
• str_getchar(str, offset[, type])
• str_setchar(str, offset, value[, type])
• strcpy_fromslider(str, slider)
• sprintf(str,format, ...)
• match(needle, haystack, ...)
• matchi(needle, haystack, ...)

Graphics

• gfx_set(r[g,b,a,mode,dest])
• gfx_lineto(x,y,aa)
• gfx_line(x,y,x2,y2[,aa])
• gfx_rectto(x,y)
• gfx_rect(x,y,w,h)
• gfx_setpixel(r,g,b)
• gfx_getpixel(r,g,b)
• gfx_drawnumber(n,ndigits)
• gfx_drawchar($'c')
• gfx_drawstr(str[,flags,right,bottom])
• gfx_measurestr(str,w,h)
• gfx_setfont(idx[,fontface, sz, flags])
• gfx_getfont()
• gfx_printf(str, ...)
• gfx_blurto(x,y)
• gfx_blit(source, scale, rotation)
• gfx_blit(source, scale, rotation[, srcx, srcy, srcw, srch, destx, desty, destw, desth, rotxoffs, rotyoffs])
• gfx_blitext(source, coordinatelist, rotation)
• gfx_getimgdim(image, w, h)
• gfx_setimgdim(image, w,h)
• gfx_loadimg(image, filename)
• gfx_gradrect(x,y,w,h, r,g,b,a[, drdx, dgdx, dbdx, dadx, drdy, dgdy, dbdy, dady])
• gfx_muladdrect(x,y,w,h, mul_r, mul_g, mul_b[, mul_a, add_r, add_g, add_b, add_a])
• gfx_deltablit(srcimg,srcx,srcy,srcw,srch, destx, desty, destw, desth, dsdx, dtdx, dsdy, dtdy, dsdxdy, dtdxdy[, usecliprect=1] )
• gfx_transformblit(srcimg, destx, desty, destw, desth, div_w, div_h, table)
• gfx_circle(x,y,r[,fill,antialias])
• gfx_roundrect(x,y,w,h,radius[,antialias])
• gfx_arc(x,y,r, ang1, ang2[,antialias])
• gfx_triangle(x1,y1,x2,y2,x3,y3[,x4,y4,...])
• gfx_getchar([char])
• gfx_showmenu("str")
• gfx_setcursor(resource_id[,"custom cursor name"])
• gfx_getdropfile(idx[,#str]) (not documented in JSFX)
• gfx_r, gfx_g, gfx_b, gfx_a
• gfx_w, gfx_h
• gfx_x, gfx_y
• gfx_mode
• gfx_clear
• gfx_dest
• gfx_texth
• gfx_ext_retina (lib OK, plugin needs work)
• mouse_x
• mouse_y
• mouse_cap
• mouse_wheel
• mouse_hwheel

Hi,

I'm not sure is it a real problem or not, but openSUSE build server complains on building that .so has a executable stack. I think it was started after b1921b6

The build log: https://build.opensuse.org/package/live_build_log/home:kill_it:JUCE/ysfx/openSUSE_Tumbleweed/x86_64
The error message:

[  165s] lv2-ysfx.x86_64: E: executable-stack (Badness: 10000) /usr/lib64/lv2/ysfx.lv2/ysfx.so
[  165s] vst-ysfx.x86_64: E: executable-stack (Badness: 10000) /usr/lib64/vst/libysfx.so
[  165s] vst3-ysfx.x86_64: E: executable-stack (Badness: 10000) /usr/lib64/vst3/ysfx.vst3/Contents/x86_64-linux/ysfx.so
[  165s] The binary declares the stack as executable. Executable stack is usually an
[  165s] error as it is only needed if the code contains GCC trampolines or similar
[  165s] constructs which uses code on the stack. One common source for needlessly
[  165s] executable stack cases are object files built from assembler files which don't
[  165s] define a proper .note.GNU-stack section.

Hi,

not really an issue but I made a package for archlinux: https://aur.archlinux.org/packages/ysfx-git ; maybe it can be useful to others who come across here and go in the readme or something.

cheers ! (and if you see anything that needs fixing / improving in the package, please tell me :-))

Not your problem, but documenting just in case others run into this
The duct-tape fix is to change the definition in win32_utf8.c to use struct _stat64 instead of struct stat

====================[ Build | eel2 | RelWithDebInfo-MinGW ]=====================
C:\Users\rayga\AppData\Local\JetBrains\Toolbox\apps\CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe --build C:\Users\rayga\Projects\tmp\ysfx\cmake-build-relwithdebinfo-mingw --target eel2
[0/2] C:\Users\rayga\AppData\Local\JetBrains\Toolbox\apps\CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe -P C:\Users\rayga\Projects\tmp\ysfx\cmake-build-relwithdebinfo-mingw\CMakeFiles\VerifyGlobs.cmake
[1/2] C:\msys64\ucrt64\bin\cc.exe -DEELSCRIPT_NO_LICE -DEELSCRIPT_NO_NET -DNSEEL_ATOF=nseel_atof -DWDL_FFT_REALSIZE=8 -D_FILE_OFFSET_BITS=64 -IC:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source -O2 -g -DNDEBUG -fvisibility=hidden -Wall -MD -MT CMakeFiles/eel2.dir/thirdparty/WDL/source/WDL/win32_utf8.c.obj -MF CMakeFiles\eel2.dir\thirdparty\WDL\source\WDL\win32_utf8.c.obj.d -o CMakeFiles/eel2.dir/thirdparty/WDL/source/WDL/win32_utf8.c.obj -c C:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source/WDL/win32_utf8.c
FAILED: CMakeFiles/eel2.dir/thirdparty/WDL/source/WDL/win32_utf8.c.obj 
C:\msys64\ucrt64\bin\cc.exe -DEELSCRIPT_NO_LICE -DEELSCRIPT_NO_NET -DNSEEL_ATOF=nseel_atof -DWDL_FFT_REALSIZE=8 -D_FILE_OFFSET_BITS=64 -IC:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source -O2 -g -DNDEBUG -fvisibility=hidden -Wall -MD -MT CMakeFiles/eel2.dir/thirdparty/WDL/source/WDL/win32_utf8.c.obj -MF CMakeFiles\eel2.dir\thirdparty\WDL\source\WDL\win32_utf8.c.obj.d -o CMakeFiles/eel2.dir/thirdparty/WDL/source/WDL/win32_utf8.c.obj -c C:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source/WDL/win32_utf8.c

C:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source/WDL/win32_utf8.c: At top level:
C:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source/WDL/win32_utf8.c:803:5: error: conflicting types for 'statUTF8'; have 'int(const char *, struct stat *)'
  803 | int statUTF8(const char *filename, struct stat *buffer)
      |     ^~~~~~~~

In file included from C:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source/WDL/win32_utf8.c:6:
C:/Users/rayga/Projects/tmp/ysfx/thirdparty/WDL/source/WDL/win32_utf8.h:64:25: note: previous declaration of 'statUTF8' with type 'int(const char *, struct _stat64 *)'
   64 | WDL_WIN32_UTF8_IMPL int statUTF8(const char *filename, struct stat *buffer);
      |                         ^~~~~~~~

ninja: build stopped: subcommand failed.

Bundled VST3 breaks:

In file included from /disk-samsung/freebsd-ports/audio/ysfx/work/.build/_deps/juce-src/modules/juce_audio_plugin_client/VST3/juce_VST3_Wrapper.cpp:42:
/disk-samsung/freebsd-ports/audio/ysfx/work/.build/_deps/juce-src/modules/juce_audio_processors/format_types/juce_VST3Headers.h:29:3: error: To build JUCE VST3 plug-ins on BSD you must use an external BSD-compatible VST3 SDK with JUCE_CUSTOM_VST3_SDK=1
 #error To build JUCE VST3 plug-ins on BSD you must use an external BSD-compatible VST3 SDK with JUCE_CUSTOM_VST3_SDK=1
  ^

This problem is definitely solved in the VST3 port.

How to use external packages?
vst3sdk-3.7.4.b.25
juce-6.0.8_4

FreeBSD 13.1

I receive several "Unable to build: error C3646: 'effect': unknown override specifier" when attempting to build after a fresh clone on my win 10 machine. This happens right after cmake --build .

D:\BeatConnect\Plugin\ysfx\plugin\info.h(23,18): error C3646: 'effect': unknown override specifier (compiling source fi
le D:\BeatConnect\Plugin\ysfx\plugin\editor.cpp) [D:\BeatConnect\Plugin\ysfx\build\ysfx_plugin.vcxproj]
D:\BeatConnect\Plugin\ysfx\plugin\info.h(23,18): error C4430: missing type specifier - int assumed. Note: C++ does not
support default-int (compiling source file D:\BeatConnect\Plugin\ysfx\plugin\editor.cpp) [D:\BeatConnect\Plugin\ysfx\bu
ild\ysfx_plugin.vcxproj]
D:\BeatConnect\Plugin\ysfx\plugin\editor.cpp(130,30): error C2039: 'effect': is not a member of 'YsfxInfo' [D:\BeatConn
ect\Plugin\ysfx\build\ysfx_plugin.vcxproj]

Instead of waiting for more typos in other peoples code,
I tried to do it all wrong myself and share the results here
hoping it is somewhat helpful.

All the following examples compile and show in reaper.

These are 10 identical sliders:

slider1:official=0<-150,12,1>official
slider2:0<-150,12,1>official no var.name
slider3:=0<-150,12,1>=value
slider4:<-150,12,1>no default
slider5:0<-150,12,1,,,>toomanycommas
slider6:0<-150,12,1,2,3,4>toomanyvalues
slider7:0time<-150kilo,12uhr,1euro>strings
slider8:0*2<-150-151,12=13,1+3>math?
slider9:+/-0a0<-150<<-149<,12...13,1 3><v<<al..u e>
slider10:a1?+!%&<-150%&=/?+!,12!%/&?+=,1=/?+!%&>?+!%&=/

What (i think) i see is

• default_value: Falls back to zero if value is nonsense ,empty var.
  or non existant (=value is allowed here).
• <min,max,increment> : If field starts with valid value,
  reaper takes that and ignores the rest.
• " and ' are a problem if not closed.
  You can < multiple times but > only once.
  You can't open with <<.
• Additional subsequent fields are ignored.

If fields are empty:

//(if either min or max >=10 then increment=0.1 (for following 3 examples))

slider11:<-6>   //-> 0<-6,0,0.01>
slider12:<,6>   //-> 0<0,6,0.01>
slider13:<6,>   //-> 0<6,0,0.01>

slider14:<3,,>   //-> 0<3,0,0>
slider15:<,3,>   //-> 0<0,3,0>
slider16:<3,,1>  //-> 0<3,0,1>
slider17:<,3,1>  //-> 0<0,3,1>

//all nonsense -> all zero:
slider18:a1<b2,c3,d4>  //-> 0<0,0,0> (without string here, its not accessible)

Similar for menu slider:

• increment is always 1 as long as there are 2 commas.
• min/max values don't matter, the number is defined by the {L,I,S,T}

slider20:0<0,5,1{0,1,2,3,4,5}>official menu
slider21:<,,{0,1,2,3,4,5}>empty fields
slider22:0<0.2,64,12{0,1,2,3,4,5}>wrong values
slider23:0<0?,<5a<,b1c<{0,1,2,3,4,5}><?abc

That's what I found so far.

This would speed up signal processing on ARM machines (by @falkTX's request)
This needs to write the ASM equivalents and possibly it's not difficult at all.

(currently I don't have a machine at hand that lets me test this)

Due to some WDL things, build error being:

thirdparty/WDL/source/WDL/wdltypes.h:203:69: error: narrowing conversion of '-1' from 'int' to 'long unsigned int' [-Wnarrowing]
  203 | typedef char wdl_assert_failed_unsigned_char[((char)-1) > 0 ? -1 : 1];
      |                                                                     ^
thirdparty/WDL/source/WDL/wdltypes.h:203:61: error: size '-1' of array 'wdl_assert_failed_unsigned_char' is negative
  203 | typedef char wdl_assert_failed_unsigned_char[((char)-1) > 0 ? -1 : 1];
      |                                              ~~~~~~~~~~~~~~~^~~~~~~~

Shortly after clicked, the toggle goes back to the initial position.

On CentOS 7 I'm facing:

target_compile_definitions(ysfx-private
    PRIVATE
        "_FILE_OFFSET_BITS=64")

which makes the build fail with:

[ 98%] Building CXX object CMakeFiles/ysfx-private.dir/sources/ysfx_eel_utils.cpp.o
In file included from /image/ysfx/sources/ysfx_utils.cpp:31:
/usr/include/fts.h:41:3: error: "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
# error "<fts.h> cannot be used with -D_FILE_OFFSET_BITS==64"
  ^

Do you think I can safely comment that on that platform ? Or I should just update to a system with a more recent fts.h ? (if so, if you know what would be the minimal requirement..)

Lots of these *.rpl files in the Geraint Luff collection

When testing on Win32 MSVC 2022 x64 build in Debug mode (and thus with std::vector assertions, etc):

As you can see, it tries to access data[pos] for pos == 482 when data.size() == 482, which is definitely not ok.

I'm not sure what should happen in that case though : from the comment it seems that it's expected than data would contain additional data.

I suppose maybe I am supposed to regenerate the SHA or something?
Unsure why this would happen on first build, as I haven't touched anything though.

Compiled with MinGW GCC 11 on Windows 11

====================[ Build | eel2 | RelWithDebInfo-MinGW ]=====================
CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe --build Projects\tmp\ysfx\cmake-build-relwithdebinfo-mingw --target eel2
[0/2] CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe -P Projects\tmp\ysfx\cmake-build-relwithdebinfo-mingw\CMakeFiles\VerifyGlobs.cmake
[1/12] cmd.exe /C "cd /D Projects\tmp\ysfx\cmake-build-relwithdebinfo-mingw && CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe -DSOURCE_FILE=Projects/tmp/ysfx/thirdparty/WDL/source/WDL/eel2/asm-nseel-x64-sse.asm -DCHECKSUM_FILE=Projects/tmp/ysfx/sources/eel2-gas/sources/ref-hash-sha512.txt -P Projects/tmp/ysfx/cmake/ValidateSHA512.cmake && CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe -E touch Projects/tmp/ysfx/cmake-build-relwithdebinfo-mingw/eel2-gas-validate-stamp.txt"
FAILED: eel2-gas-validate-stamp.txt Projects/tmp/ysfx/cmake-build-relwithdebinfo-mingw/eel2-gas-validate-stamp.txt 
cmd.exe /C "cd /D Projects\tmp\ysfx\cmake-build-relwithdebinfo-mingw && CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe -DSOURCE_FILE=Projects/tmp/ysfx/thirdparty/WDL/source/WDL/eel2/asm-nseel-x64-sse.asm -DCHECKSUM_FILE=Projects/tmp/ysfx/sources/eel2-gas/sources/ref-hash-sha512.txt -P Projects/tmp/ysfx/cmake/ValidateSHA512.cmake && CLion\ch-0\213.5744.123\bin\cmake\win\bin\cmake.exe -E touch Projects/tmp/ysfx/cmake-build-relwithdebinfo-mingw/eel2-gas-validate-stamp.txt"
CMake Error at Projects/tmp/ysfx/cmake/ValidateSHA512.cmake:31 (message):
  Checksum invalid: asm-nseel-x64-sse.asm
  ('aa5896ad4860cba62ec3450b55221f4985db4f79d83fc14dadd2f2aceae87cdfd8645f35160bf47bb7700046e3ecdd892aa74639130b910598ddb70515787ed8'
  !=
  'd4a4d8bf4d906f6a6cdbe54bc8a2699e01a11e1109a716c696fc3df96046422a71aea29ff643db4b23cf2b29a3ccd8dbb309a14bab327f13d7f2be461463bc02')


ninja: build stopped: subcommand failed.

• tags
• author

[ 20%] Building CXX object CMakeFiles/lice.dir/thirdparty/WDL/source/WDL/swell/swell-gdi-lice.cpp.o
/image/ysfx/thirdparty/WDL/source/WDL/swell/swell-gdi-lice.cpp:47:10: fatal error: '../dirscan.h' file not found
#include "../dirscan.h"
         ^~~~~~~~~~~~~~

It seems to exist in upstream WDL: https://github.com/justinfrankel/WDL/blob/master/WDL/dirscan.h

Notes on how to process keyboard input.

ysfx_gfx_input(uint32_t mod, uint32_t key)

• mod is the modifiers bits ; it recognizes Ctrl and Alt
• key is unicode, and also can be a special key from an enumeration, whose values are in the unicode private use area (note: check pugl keys)

The values are pushed to a queue in the gfx state.
The internal mapping is like this:

• if it's a special key, it's converted to one of JSFX's multichar equivalents ('up', 'down' etc)
• if it's a regular key "a-z"/"A-Z" and it has mods, convert that value into another range
  (Alt: 321-346, Ctrl: 1-26, Ctrl+Alt: 257-282)
• if it's a regular key otherwise, and it's >=1 and <256, push the value as is

Other

• Ctrl/Alt presses aren't reported, they are modifier only

I just built the vst3 which worked fine (linux)
the only step i found missing was the submodules which i got with
git submodule update --init --recursive
if you want to add it to the readme.

As probably expected some jsfx load and work, some don't.
Are you interested in testing feedback already or is it too early in the process?
Anyway very exciting!

https://www.reaper.fm/sdk/js/gfx.php

Note (1): this appears to be the default font
https://int10h.org/oldschool-pc-fonts/fontlist/font?tandy1k-ii_200l

Note (2): gfx_showmenu is synchronous
this blocks @gfx from running, but not the entire host

I downloaded the VST3 development build but it is crashing inside Bitwig 4.2 as soon as it is loaded. Bitwig's log says this:

About to start the following process:  /opt/bitwig-studio/bin/BitwigPluginHost-X64-SSE41 host Jean-Pierre-Cimalando-ysfx-17 48000 256 2015549-17 16 not-dpi-aware 8 jack jack 8814603
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Successfully set thread realtime priority to 50
Creating plugin audio thread proxy 0
PluginHost: Starting new audio thread for processing plugin 0
Successfully set thread realtime priority to 50
X Error of failed request:  BadWindow (invalid Window parameter)
  Major opcode of failed request:  20 (X_GetProperty)
  Resource id in failed request:  0x0
  Serial number of failed request:  54
  Current serial number in output stream:  54
Could not read async reply: End of stream
Sending a shutdown request to the plugin host
Error sending request to plug-in host: Error writing to named pipe: Broken pipe
Waiting for plugin host process to exit
Child process with PID 2020446 exited with error code: 1
Plugin host process exited with code: 256
Waiting for plugin host process to exit
Plugin host process exited with code: 256

I'm using Arch linux.

More on "everything is optional in jsfx"

• If Input/Output pins are not mentioned at all, reaper will read and show this as js in/out,
  as long as there is @sample with something in it (and something can be anything)

example fx:
https://bitbucket.org/kawaCat/reascript-m2bpack/src/master/JSFX%20effects/kawa_Simple_Mid_Side.jsfx
https://github.com/JoepVanlier/JSFX/blob/master/SatanVerb/SatanVerb.jsfx
also a file with nothing in it but

@sample
hello

will have these I/O's

• If only input pins are mentioned reaper reads output as "unmapped" (which you would have to explicitly "zero out" in reaper)

example fx: almost all in /Reaper/Effects/analysis/

• If only output pins are mentioned reaper will display this correctly
  but the effect still works...
  (no example for this one, I just commented out the in pins of an EQ)

• And a completely empty file shows no I/O
  but still compiles and passes audio (as long as it has a filename and sits in the Effects folder)

All above examples currently don't output sound when loaded in the YSFX-plugin.

Hello,

To inform you the plugin is not accepted here in 2 Daw's : Bespoke (vst3) & Usine Hollyhock (vst3 & Au)
MacOs Mojave / macbookpro

The sliders which start with - are hidden by default.
This should not appear in the display name.
Also implement the slider_show API, and the change notification for slider visibility.

see here
https://github.com/JoepVanlier/JSFX/blob/4238c16ab81cd6590435c8219fd4d0a1483217fe/Squashman/Squashman.jsfx#L1-L7

In Gigperformer a project is called a gig. When I load a gig with ysfx included as a plugin Gigperformer is stuck telling me it is trying to load ysfx. Loading a gig without ysxf and then adding it works normal. (But then, I want to save the gig, that succeeds, but loading does fail).

It does not matter whether ysxf has a script loaded or not.

This is on Windows 11, no cpu load from Gigperformer. Maybe some event, semaphore or deadlock?

FYI: Reaper exhibits the same behavior

The sample rate of e.g. Reaper is not reflected in the system variable 'srate' in jsxf. When the sample rate is altered srate does also not reflect this change (not surprisingly).

Tested with GigPerformer 4.0 and Reaper 6.43. ysfx Plugin build was the developer build of Jan 5th, 2022.

BTW, amazing plugin. Is knew jsfx for use within Reaper and ReaJS for other VST host. The environment is Reaper is very nice, but proprietary, ReaJS is useful with other VST hosts, but rather limited (maximum of 2 audio channels, at least with GigPerformer) and kind of deprecated. All this is solved Ysfx. Thank you.

Is there any possibility in the future to load fx by (slot, rack or channelstrip), to save chain presets in a single instance?..thanks in advance..

Linux Mint 20.2: followed yor build commands step by step
On the step:"cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo .."
compilation terminates with:
ysfx_audio_wav.cpp:31:10: fatal error: dr_wav.h: File not found
i guess, it is the dr_libs; right?
https://github.com/mackron/dr_libs
So, what to do?

• gmem
• want_all_kb
• maxmem
• no_meter

As reported here:

https://aur.archlinux.org/packages/ysfx-git#comment-872572

/home/harv/.cache/yay/ysfx-git/src/ysfx-build/_deps/juce-src/modules/juce_gui_basics/windows/juce_ComponentPeer.h:504:18: error: ‘exchange’ is not a member of ‘std’

most likely this file lacks #include <utility>

sources/ysfx_utils_fts.cpp:43:16: error: no matching function for call to 'fts_open'
    FTS *fts = fts_open(argv, FTS_NOCHDIR|FTS_PHYSICAL, +compar);
               ^~~~~~~~
/usr/include/fts.h:131:6: note: candidate function not viable: no known conversion from 'auto (*)(const FTSENT **, const FTSENT **) -> int' (aka 'auto (*)(const _ftsent **, const _ftsent **) -> int') to 'int (*)(const FTSENT *const *, const FTSENT *const *)' (aka 'int (*)(const _ftsent *const *, const _ftsent *const *)') for 3rd argument
FTS     *fts_open(char * const *, int,
         ^
1 error generated.

From the fts_open manpage:

FTS(3)                 FreeBSD Library Functions Manual                 FTS(3)

NAME
     fts – traverse a file hierarchy

LIBRARY
     Standard C Library (libc, -lc)

SYNOPSIS
     #include <fts.h>

     FTS *
     fts_open(char * const *path_argv, int options,
         int (*compar)(const FTSENT * const *, const FTSENT * const *));

...

When I open and display two instance of ysfx simultaneously the screen gets garbled big time, where it comes to texts.
In Reaper you can detach the plugin-interfaces by double-clicking the plugin.

Problem does not occur when the plugin-interfaces are not being shown at the same time.

Might be related to ui-lib.jsfx-inc https://geraintluff.github.io/jsfx-ui-lib

This conflicts with the CMAKE_DL_LIBS variable: https://cmake.org/cmake/help/latest/variable/CMAKE_DL_LIBS.html

if another project uses CMAKE_DL_LIBS and includes ysfx, things may go awry due to dl now being a target:

CMake Error in 3rdparty/libossia/3rdparty/Servus/servus/CMakeLists.txt:
  export called with target "Servus_iface" which requires target "dl" that is
  not in any export set.


-- Generating done
CMake Error in 3rdparty/libossia/3rdparty/wiiuse/src/CMakeLists.txt:
  export called with target "wiiuse" which requires target "dl" that is not
  in any export set.

This would allow to test "if it works or not" straight after building.

A solution is to use different function tables for compiling different sections, where forbidden functions are replaced with dummies.
This will make the @gfx thread check unnecessary.

See pdc_delay variable and others.

This is probably not a big issue and should not keep you from going to sleep sometimes (as far as I'm concerned):

When you have for example the following code in the @gfx section, processing of all code in that section stops. In this example gfx.debug will not increase:

@sample
gfx.debug_sample += 1;

@gfx
gfx.debug += 1;
 
// Uncomment this for letting things run 'normal'
/*
gfx_x=10;
gfx_y=10;
tmp=#;
sprintf(tmp,"Just to do something %d", gfx.debug);
gfx_drawstr(tmp);
*/

You can see in the editor whether gfx.debug and gfx.debug_sample (the last one will keep increasing) are increasing or not:

BTW: I'm not trying to find as many bugs as I can find and I'm also not implying they al must be fixed at once, but you people seem really determined to make this plugin a good plugin and therefore I will report what I run into. Maybe at some time I find the strength to set up an environment that I can compile this plugin myself and be a little more constructive. One can always hope...