Esteban Gutierrez

Hi everyone! I'm Esteban and I've been using the min devkit for a little while now, getting nice results building externals using the Eigen library, however, I've never had the need to load a big chunk of data at runtime. 

To be precisse, what I'd like to do is bassically to fix some data at runtime that is stored inside a (Eigen) matrix. Since, attributes can be used with this purpose when the data is a number or a string, I thought I was going to be able to put the database inside one of those, for example doing something like this:

attribute<MatrixXf> database{ this, "data base", database_maker()};

where MatrixXf is the database class and database_maker() is a function that returns a build of the database, nonetheless this does not work.

Hi everyone! I'm Esteban and I've been using the min devkit for a little while now, getting nice results building externals using the Eigen ... 

how-to-propperly-load-a-database-using-the-min-devkit

How to propperly load a database using the min devkit

Magnus Berendes

I'm currently experimenting with some things and noticed that I cannot get a bitexact copy of my input pcm samples.

When I'm reading in a 24 bit wavfile using sfplay~ and immediately write it back onto the disk using sfrecord~ (with @bitdepth 24 @dither 0 @quantization 0) I get off-by-ones at roughly 50% of the samples (+1 sample delay, which is not the problem).
As far as my understanding goes Max uses 64bit floats to represent audio internally which shouldn't have any problem representing a 24 bit int exactly, should it?

Obviously, this isn't an audible problem but for my application bitexact transport would be necessary.

Any ideas where I lose my precision and any idea how I could get it back?

Thanks!


Hi there,
I'm currently experimenting with some things and noticed that I cannot get a bitexact copy of my input pcm samples.
When I'm ... 

bit-depth-of-individual-modules

>> "uses 64bit floats to represent audio internally which  shouldn't have any problem representing a 24 bit int exactly, should it?"

or unless one of the objects in question still has that bug where the top half of the values are offset by 1.

(insiders know where we once had that bug... even matlab had that same bug for years)

i suspect that sfplay/sfrecord is the cause, you can try if the same happens with buffer releated objects then you´ll see.

My first experiment was literally sfplay into sfrecord, no gain involved anywhere.

Now I used a buffer~ object (buffer~ inbuf <filename> 1000 6 @dither 0 @filetype wave @format int24 @quantization 0 @samps 48000) and used importreplace to import a file, then using writewave I wrote the buffer to the disk and compared it with my input file - no bueno, same picture. Off by ones all over the place. 

To me it seems it is just happening with high amplitude samples, maybe see if I can find a threshold...

if it is what i mentioned above it, should be only the positive values.

Yes, it is exactly what you mentioned: all values in [0.5*MAX, MAX] (for 24bit int) are off by one. Okay, so it is this bug and apparently sfrecord~/sfplay~ AND buffer~ are affected. Do you know if this bug is also "in" the patch cords? I.e. if I implement file reading and writing myself, will I be fine?

Or any way to get bugfree buffer~/sfplay~ implementations?

(Version of Max is 8.5.3, currently updating to 8.5.5)

that is a good question if it is runtime wide, and i would have to speculate, and to be honest i assumed that was long fixed.

for music and for listening to the result it does not matter much, for other applications it can. and if you import and export 100 times, the offset will be 100 values.

how to fix that? no idea. personally i would use file/fileout in max 4.x as a post process to other things. :)

for the read/write scenario adding one bit will fix it. but as soon as you perform some algo in 64 bit before, you could still get some wrong values.

are you abusing signals for something? alternatively you can abuse jitter matrixes for data collections, too.

proposed testing scenario:

create an "original audio file" in text format containing only (signed) min, max, max-1, and 0.

open the text  with max, write it into buffer~

play it from the buffer~ using play~ and record with record~ into another buffer

check  the values in the buffer (as far as it is possible to get the full 64 value for max-1/24bit?)

save audio file from buffer and see what is in it.


Bit-exact representation of 24bit int

twhiston

Hi, I couldn't see any examples in the 8.2 sdk of how you might add a framework to a cmake generated XCode project and hoped someone could help. 
Ideally I want to be able to put something in the CMakeLists.txt which will add frameworks, such as cocoa, to my project. I also need to link them to the binary, and in some instance have a build stage which copies them into the Frameworks destination. Has anyone done this or have enough experience with CMake to explain how to do so?

Hi, I couldn't see any examples in the 8.2 sdk of how you might add a framework to a cmake generated XCode project and hoped someone could h

adding-frameworks-to-generated-xcode-projects-using-cmake

Hey I ran into a similar issue this week and had a heck of time figuring it out. I am not experienced enough in CMake to give a step by step. However, if you look at the reply to forum post I made, I laid out the links I used to resources I used to get there as well as what worked for my situation.

https://cycling74.com/forums/mac-keystrokes-using-the-mindevkit

Hi all, 
Not sure but I think this project might help out, I worked on it as well, and got finally a swift/c++ thing working. ( RNBO )
It might nudge you in the right direction
https://cycling74.com/forums/building-gui-without-juce-for-plugins?replyPage=1#reply-644a084176fed40015cf31e9

https://github.com/ceammc/SwiftRNBO

Adding frameworks to generated xcode projects using cmake

Trevor being Trevor

This post has a lot of information to give the necessary context, but essentially I can't get the project to build when I add the code to simulate a keystroke and I am not sure why. I was able to figure this out for Windows, but I am running into a problem on Mac that I can't get past. I followed the instructions of this video and github page which seem to be the most up to date.

https://www.youtube.com/watch?v=il5WblTBUgs 
https://github.com/Cycling74/max-sdk/blob/main/README-8.2-update.md 

I started by manipulating the min.hello-world project. I removed the greeting when the object loads, edited the bang function and created a function that responds to integers. Everything compiles and works as expected. However, I want to simulate a keystroke so I brought in the following libraries...

#include <CoreGraphics/CoreGraphics.h>
#include <ApplicationServices/ApplicationServices.h>
#include <Carbon/Carbon.h>

As far as I understand, I only need Application Services, but the other ones were added as I tried to troubleshoot issues. Anyway, to use those libraries I updated my "respond to integer" function to the following code...

message<> number {this, "int", 
        MIN_FUNCTION {
            int key = 56;
            //EDIT: this code is close to correct, but definitely don't trust it.
            //there are better examples on Stack Overflow, if you are really stuck
            //comment and could share something better. Also, it appears
            //olsen put 11strokes code up on Github. It says it's cross platform, so it
            //may be a better and more comprehensive source than what I could offer.
            
            // Create an HID hardware event source
            CGEventSourceRef src = CGEventSourceCreate(kCGEventSourceStateHIDSystemState);
            
            // Create a new keyboard key press event
            CGEventRef evt = CGEventCreateKeyboardEvent(src, (CGKeyCode) key, true);
            
            // Post keyboard event and release
            CGEventPost (kCGHIDEventTap, evt);
            CFRelease (evt);
            CFRelease (src);
 
            //this line printed before I added the above
            cout<< "Sometimes I work, sometimes I don't" << endl;
            return {};
        }
    };

Instead of Xcode (I am not very familiar with it) I am using the terminal to build using the command "cmake --build ." inside the build folder which works great. However, the build fails once I add the keystroke specific code. This is the error given with crucial information in bold...

Undefined symbols for architecture x86_64:
"_CFRelease", referenced from:
      hello_world::number::'lambda'(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const&, int)::operator()(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const, int) const in min.hello-world.o
  "_CGEventCreateKeyboardEvent", referenced from:
      hello_world::number::'lambda'(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const&, int)::operator()(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const, int) const in min.hello-world.o
  "_CGEventPost", referenced from:
      hello_world::number::'lambda'(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const&, int)::operator()(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const, int) const in min.hello-world.o
  "_CGEventSourceCreate", referenced from:
      hello_world::number::'lambda'(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const&, int)::operator()(std::__1::vector<c74::min::atom, hello_world::number::allocator<c74::min>> const, int) const in min.hello-world.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

Even though it says the issue is for x86_64 I have seen it say arm64 instead as I've tried to get things working. For that reason, I don't think the issue is specific to the architecture, but I could be wrong. Also,  Xcode intellisense autocompletes these symbols as I type them and I am able to right click and "Jump to Definition" so there is a clearly a path to these symbols, but apparently the linker or compiler is unaware of that path. To combat this (and an earlier issue), I tried updating the CMakeLists.txt for min.helloworld to the following...

# Copyright 2018 The Min-DevKit Authors. All rights reserved.
# Use of this source code is governed by the MIT License found in the License.md file.
cmake_minimum_required(VERSION 3.0)
set(C74_MIN_API_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../../min-api)
include(${C74_MIN_API_DIR}/script/min-pretarget.cmake)
#############################################################
# MAX EXTERNAL
#############################################################
include_directories( 
	"${C74_INCLUDES}"
	"/System/Library/Frameworks/CoreFoundation.framework/Headers"
    	"/System/Library/Frameworks/ApplicationServices.framework/Headers"
    	"/System/Library/Frameworks/Carbon.framework/Headers" 
    	"/System/Library/Frameworks/CoreGraphics.framework/Headers"
)
set( SOURCE_FILES
	${PROJECT_NAME}.cpp
)

add_library( 
	${PROJECT_NAME} 
	MODULE
	${SOURCE_FILES}
)

target_link_libraries(  
	${PROJECT_NAME}
	"${C74_LIBS}"
	"-framework CoreFoundation" 
    	"-framework ApplicationServices"
    	"-framework Carbon"  
    	"-framework CoreGraphics"
)

include(${C74_MIN_API_DIR}/script/min-posttarget.cmake)
#############################################################
# UNIT TEST
#############################################################
include(${C74_MIN_API_DIR}/test/min-object-unittest.cmake)


Lastly, in trying to figure this out I was made aware of the 11olsen 11strokes project. That could potentially work for me, but I've already built something specific on Windows and replicating it on Mac would be much less future headache if possible. Also, learning is fun most of the time haha. 

Anyway I'm at the point where I don't know how to proceed. It's probably something simple, but this is a bit out of my element so I'm struggling to see it. Any advice or ideas are welcomed and greatly appreciated!

2021 Macbook Pro, M1 Pro, Ventura 13.3.1
cmake version 3.26.0
xcode version 14.3
max version 8.5.3
Let me know if you need anymore info!

This post has a lot of information to give the necessary context, but essentially I can't get the project to build when I add the code to ... 

mac-keystrokes-using-the-mindevkit

Following up on my own post, here's what I did to accomplish the task for those in the future. 

First, I decided to eliminate CMake and the min-devkit as variables completely and just write a basic "Command Line Tool" in Xcode to simulate a keystroke. This was a struggle and I went at the problem from many angles over a couple of days. As far I can tell, this isn't possible via C++ or Apple doesn't let any old app simulate keystrokes anymore . In hindsight, I'm thinking it's the latter. I would run my code, it would trigger Security and Privacy settings and I would allow it, but then it still not give the expected output. As far as I can tell you need to submit some kind of enhancement request now. (I tested this on Apple Silicon and Intel Based Macs by the way)
https://developer.apple.com/forums/thread/28605

Second, from my post above I was now aware of 11strokes by olsen. I downloaded it and tested it in Max, works great! That told me it was definitely possible to simulate keystrokes. In addition, I'm not sure now, but at the time I was suspicious that C++ may be part of the issue. Also I noticed 11olsen used the max-sdk instead of the min-devkit so I decided to switch to that. I am not as familiar with C so it took a little while to understand and get a customized basic object up and running. 

Once that was ready, I brought in the code to simulate a single keystroke. Immediately, I got undefined symbols error I described in my first post. At this point, I knew the problem was with CMake. I needed to manipulate the CMakeLists.txt file located under max-sdk/source/basics/myCustomExternal to make it aware of the Carbon framework I was using. This is what ended up working for me...

include(${CMAKE_CURRENT_SOURCE_DIR}/../../max-sdk-base/script/max-pretarget.cmake)
#############################################################
# MAX EXTERNAL
#############################################################
include_directories( 
	"${MAX_SDK_INCLUDES}"
	"${MAX_SDK_MSP_INCLUDES}"
	"${MAX_SDK_JIT_INCLUDES}"
	/Developer/Headers/FlatCarbon
)
FIND_LIBRARY(CARBON_LIBRARY Carbon /System/Library/Frameworks)
MARK_AS_ADVANCED (CARBON_LIBRARY)
SET(EXTRA_LIBS ${CARBON_LIBRARY})
file(GLOB PROJECT_SRC
     "*.h"
	 "*.c"
     "*.cpp"
)
add_library( 
	${PROJECT_NAME} 
	MODULE
	${PROJECT_SRC}
)
target_link_libraries(myCustomExternal(<-your project name) PRIVATE ${EXTRA_LIBS})
include(${CMAKE_CURRENT_SOURCE_DIR}/../../max-sdk-base/script/max-posttarget.cmake)


The way that CMake is implemented for the sdk would take me, a beginner, at least a few days to wrap my head around. I mostly got there by trail and error and using these sources...
https://gitlab.kitware.com/cmake/community/-/wikis/doc/cmake/platform_dependent_issues/HowToUseExistingOSXFrameworks

https://stackoverflow.com/questions/2908640/how-to-add-a-framework-to-cmake

Mainly the second video in this youtube playlist
https://www.youtube.com/playlist?list=PLalVdRk2RC6o5GHu618ARWh0VO0bFlif4 

Also, the official CMake documentation and ChatGPT gave me a clue at some point. 

Lastly, something that helped me troubleshoot my CMakeLists.txt file was deleting the max-sdk/build folder's contents as I made educated guesses. Then running "cmake -G Xcode .." from the build folder in the terminal and then follow that with "cmake --build ." The first command would usually throw errors if there was something wrong with my CMakeLists.txt.

Overall, I banged my head on this problem for about a week and some of that was definitely my lack of knowledge. However, I think some extra guidance like some common basic CMake manipulations (like adding a framework in my case) would go a long way. Anyway, if this post helps you out with something like this, comment! It would be encouraging to know this helped.

Mac keystrokes using the minDevkit

b woods

Can anyone help me understand how to most efficiently share float data across externals and other Max-native objects without using patch cords?     I am trying to write an external to access/receive/retrieve data (viz., a float array) from other objects without connecting with patch cords. I have been reading various versions of the API documentation and different forum posts and SDK examples and believe there may be several ways to do this (using, for instance, tables or attributes or dictionaries or the send/receive or pattr or coll systems) but as a beginner (and more a C than C++ writer) I find the documentation to be more of a reference than a how-to and thus the nuts-and-bolts of implementing this elude me.  For instance, I see info on "getter and setter" functions and "getvalueof and setvalueof" functions but how would my external object find the pointer or name or handle or symbol to allow it to access data in another object? Do I need to write my own "MakeDictionary" object so that I can a priori set the name and "register" it, then write another external to access that object's registered dictionary? Any help with this would be much appreciated.  Cheers.

Can anyone help me understand how to most efficiently share float data across externals and other Max-native objects without using patch ... 

write-external-to-access-data-without-patch-cords

that kind of general question is harder to answer than a specific problem you are facing. Dicts always have a global name no matter if it's created in a patcher or inside your external. There's a dict section in the sdk doc to find out which methods to use. Once created (registered) any other external or [dict] in a patcher can also access the same dict. 
The easiest way to get a handle to any other object (not a dict) in a patcher is to assign a scripting name to it and:

t_object* jp;

// get patcher containing your external
object_obex_lookup(x, gensym("#P"), (t_object**)&jp);
//find obj by scrpt name
t_object* yourTarget = (t_object*)object_method(jp, gensym("getnamedbox"), gensym("theName"));
//now you have a handle for the box, to get the contained object:
t_object* targetObj = jbox_get_object(yourTarget);

with the handle you can then use getvalueof method (UI objects) or access attributes. But again you can't expect someone answering with an essay about all topics that you touch on in your question. In general dicts are a good way to share data between externals, I think you don't even have to worry about thread safety with them.  


Thanks, 110L532.   I kept on searching through the forums and found the answer in this one.  Not sure how I missed it, but there is code for this in the examples that come with the SDK. It is under the "misc" folder and comprises the "master" and "servant" routines. I was able to modify their _anything and _notify routines to send a float array pointer from the master to the servant. 

write external to access data without patch cords

11OLSEN

Hi!
We have [dropfile] to be a drag n' drop target but we have no source for such operation in Max. Can anyone provide some code snippet or examples in C that deal with drag n drop of files (being the source)  or has already done some investigation into that direction? Win or Mac doesn't matter, looking for a solution on both.


Hi!
We have [dropfile] to be a drag n' drop target but we have no source for such operation in Max. Can anyone provide some code snippet or 

drag-files-from-max

max doesnt know anything about the path which the mouse is theoretically pointing on when it is in the finder/explorer i am afraid.

but as a workaround why not use dropfile to drop a target folder onto it and then have an "export" or "save" button beside it. (maybe with "desktop" and "user/.../max 7/..." as default locations)

What I want is sending a list of filenames to an external which is then telling the OS to drag these files until mouse is released. To any target. I don't think the drag source has to know much about the target..
i'm tired of doing workarounds ;)
lg
O.


you (the external) somehow must link the mouse position to the OS browser. well, maybe that is possible (just as screen recording once was possible with nato)

but what do i know about externals. the only thing i know is that it can be wise to remain sceptical. about everything. :)

i remember well how baffled (or better amazed) i was when i found out that the drag and rop manager in mac os 7 allowed to drag and drop layers of pictures files directly between photoshop 3 and graphic converter.

I have promising results. For Windows  Anders Karlsson  created this library https://sourceforge.net/projects/win32cdnd/ that allows handling some stuff on a slightly higher level. He tagged it as GPL 2.0.
If I use this I have to go with GPL 2.0, too. Does this work for an external? 
As I understand it would force everyone using it to distribute the source with it. Can we just link for example to a github repo instead? 
Everyone using it must release under GPL 2. I understand this when the sourcecode is reused for another external but how far does it go up the chain. 
It doesn't mean the patcher or standalone or m4live device using the external has to go with GPL, right? 



http://11olsen.de/code/download/6-max-msp-externals/22-11dragfiles 

11Olsen, it works great.  This is fantastic news. Thank you so much.
Hugo


11Olsen, how can i use this in to my max patch?

The help file for the object is pretty self-explanatory, isn't it? Or do you mean how to use 3rd party externals in general?

drag files FROM Max, solved

Hi, the [thispatcher] object is only useful to style the window of the "first view" of a patcher. So i'm trying to use an external to modify the window of a second view of the same patcher. Many things already work by accessing attrs and methods of the correct patcherview. But how do i change window flags like grow/nogrow , float/nofloat?  That would save me a lot of fiddling with the native window on both platforms.
 

Hi, the [thispatcher] object is only useful to style the window of the "first view" of a patcher. So i'm trying to use an external to modify

2nd-patcherview-window-flags

I found out window flags are set with a method of t_patcher: (https://cycling74.com/forums/objects-with-editing-windows)

t_atom rv; // return value
object_method_parse(patcher, gensym("window"), "flags float", &rv);
object_method_parse(patcher, gensym("window"), "flags nominimize", &rv);
object_method_parse(patcher, gensym("window"), "flags nogrow", &rv);
object_method_parse(patcher, gensym("window"), "exec", &rv);

void *result;
result = object_method(patcher, _sym_makeview);
patcherview_set_visible((t_object *)result, 1);

the new window does not take the flags into account. and there's no way to set flags for the new window.

What i need in the end is a "new view" that opens as floating window above a fullscreen window.
Atm the new view always opens in fullscreen if the app is in fullscreen mode.(macOSX)
  


Hi there,

so if you send the "window" message to the patcher it will only affect the current view.
To affect additional views you need to send the "window" message to the specific view you want to modify. The message arguments remain exactly the same.

A few lines of code are worth a thousand words...


t_object *patcher = NULL;
object_obex_lookup(x, _sym_pound_P, &patcher);

{
    t_atom av[2];
    t_object *secondview = (t_object *)object_method(patcher, gensym("makeview"));
    patcherview_set_visible(secondview, 1);
    
    atom_setsym(av, gensym("flags"));
    
    atom_setsym(av+1, gensym("nominimize"));
    object_method(secondview, gensym("window"), patcher, 2, av);
    
    atom_setsym(av, gensym("exec"));
    object_method(secondview, gensym("window"), patcher, 1, av);
}
    
{
    t_atom av[2];
    t_object *thirdview = (t_object *)object_method(patcher, gensym("makeview"));
    patcherview_set_visible(thirdview, 1);
    
    atom_setsym(av, gensym("flags"));

    atom_setsym(av+1, gensym("float"));
    object_method(thirdview, gensym("window"), patcher, 2, av);
    
    atom_setsym(av+1, gensym("nogrow"));
    object_method(thirdview, gensym("window"), patcher, 2, av);
    
    atom_setsym(av, gensym("exec"));
    object_method(thirdview, gensym("window"), patcher, 1, av);
}


If we think about the current view as the first view, the code above creates two additional views. 
The second view will have the "minimize" flag disabled, whereas the third view will be a floating window with the "grow" flag disabled.


it is great that you are building something better than thispatcher,..

"Hi, the [thispatcher] object is only useful to style the window of the "first view" of a patcher."

...but why? i use thispatcher all the time to change between panel, palette, and regular view, move windows around or hide them, all of the commands work after initialisation too.
(?)


Hi Roman, I can explain the simple idea behind this. I have m4l devices which i want to be able to be displayed in a window floating above Live.  And Max already has this concept of multiple windows displaying the same patcher. 
Send "front" to a [thispatcher] in a m4l device. You'll see a perfect duplication of the device in a window. This is a second view of the same patcher but it doesn't float above Live. And all window related messages to thispatcher are applied to the first view which is the fixed frame that is diplayed in the Live UI.

But now, I can have completely different parameters for both views.

The window title says "sub patch" but this is actually the top level patcher of the m4l device..

Luigi, for the future, is there a way to find out the signature of object methods? How could I have known that the "window" method of the patcherview needs the additional patcher argument?


Yes, good question... in an ideal world the answer should have come from those with access to the source code: i.e. C74 people. That's why I didn't answer you immediately. I wanted to see if someone from C74 would answer. But it didn't happen. They don't seem to be too responsive to this kind of dev questions, and I didn't want your question to go unanswered because it is a very pertinent and legitimate one.

So the way to find out the signature of object methods (and other things) when you don't have access to the source code is to disassemble the Max executable.
Not the most enjoyable thing to do, but if you don't have access to the source code and those who do are not available to help, this kind of reverse engineering is the only way to find out information about the code... and sometimes even this is not enough.

OK, i understand.  The forums is all we have and good guessing and try n' error skills.
Best,
O. 

..in your gif , you change the track selection in Live just after you floated the windows, ..hard to see what happens in the original track. Is the device still 'docked' and reflects the changes of the floating one ? and vice et versa ?
Are both the same 'visualization" of the same patch ?


ya, i'm changing track to show the window is floating. Yes, just 2 views of the same patch. If you turn a dial in new view you'll see it moving in the "docked view".

Luigi, I'm really close to having the external finished for both platforms. One last thing I would like to add is highlighting an object in a view for Max educational tasks. I think, hence the name, it's the "tinge" and "fadetinge" methods. But the only thing i found out with "tinge" is that it seems to expect an int as first arg and a pointer as second. but any expected combination doesn't work:
//object_method(x->w_patcherview, _sym_tinge, 1,   &rect);
//object_method(x->w_patcherview, _sym_tinge, 1, obj_box );
 //object_method(x->w_patcherview, _sym_tinge, 1,  box_id_sym); 
//object_method(x->w_patcherview, _sym_tinge, 1, patcher, obj_box );
//object_method(x->w_patcherview, _sym_tinge, 1,  patcher, box_id_sym);
and i tryed many more.
You know the signature? 
Is this the method that shows this yellow transparent panel above an object's box?


Hi there,
congrats for your external. It seems really cool.
Yes, the tinge method is something that would be super-handy to have as part of the Max API as well.
Unfortunately it is not so...

Here is the signature that I found, which I am pretty sure is correct.
However I am not sure what each argument is supposed to represent.

void patcherview_tinge(t_object *pv, long count, t_atom *objects, t_atom *retv, long flag1, long flag2, long flag3)
{
object_method(pv, _sym_tinge, count, objects, retv, flag1, flag2, flag3);
}

So let me tell you what I am guessing the arguments are:
pv: pointer to the patcherview (no brainer)
count: the number of objects in the objects array
objects: an array of atoms of A_OBJ pointers representing the objects in the patcher/patcherview.
retv: no freakin' idea... I interpreted it as some kind of return value, but I am probably wrong.
flag1, flag2, flag3: these are 3 flags which are supposed to affect the resulting tinge operation.
They can be zero or non-zero (off/on), They probably set some kind of options, but I am afraid that's all I could figure out.

You may have to mess with it a little bit to see if you can get something useful out of it.
If you do, please post back your findings.
I would be curious to know too since I have been wanting to use this method as well, but I was never able to.

the tinge method is something that would be super-handy to have as part of the Max API

thanks Luigi (object_method(pv, _sym_tinge, count, objects, retv, flag1, flag2, flag3)) 
The count acts very strangly. If i use a single atom like this it works:
t_atom a;
t_atom rv;
t_object *obj = (t_object *)object_method(x->jp, _sym_getnamedbox, gensym("me"));
atom_setobj(&a, obj);
object_method(x->w_patcherview, _sym_tinge, 1, &a, &rv, 0, 0, 0 );
the tinge becomes visible and fades out after a sec.

Creating an atom array with 2 object, I have to set the count to 3 or only the first object is tinged.
Using 3 objects I have to set the count to 7.
Using 4 objects I have to set the count to 13 ! 
Without understanding this I can't calc the count for a number of objects.

retv returns A_NOTHING or undefined type numbers.

For flag 1 and 2 I couldn't notice any change in any combination. Flag 3 set to 1 keeps the tinge alive and doesn't fade it to transparent. The big question is how to clear this static tinge for an object. Called multiple times, it's layering tinges on top until there's no transparency left. Maybe "fadetinge" can clear them?

BTW: I'm using "showtarget" which is already better than nothing. It places an outlined red dot marker somewhere in the view.
//on
object_method(x->w_patcherview, _sym_showtarget, 1, int X, int Y);
//off
object_method(x->w_patcherview, _sym_showtarget, 0);

Luigi, maybe you can also confirm this signature. thanks

Hi Olsen,

 as far as the signature for patcherview_showtarget() goes, what you have is correct.
Here is the signature:
patcherview_showtarget(t_object *pv, long state, int x, int y);

As far as the tinge method goes, I am afraid that's as far as we'll get without C74's help.
You already got it somewhat working, which is something I was never able to do.
It seems to make it work properly you need a combination of tinge and fadetinge, but it is hard to know how it is supposed to be used.

BTW, I double checked and I noticed that the rv argument is not a return value but an input value to the function.
That's why you weren't seeing anything meaningful being returned in there.
I believe it somehow affects the type of tinge action, so you need to define the atom in your code and pass it.
Maybe that in combination to the objects arg will give new meaning to the count arg???
Sorry, I know it's a long shot but I am not able to be more precise.

As I said, please take everything I say with a grain of salt as lots of it is just guesswork.
As you have seen, the arguments I guessed for patcherview_tinge() may not even be the correct ones in the end.
It seems the count arg is somehow related to the number of objects to tinge but it is not the real object count.

Again, if C74 could spare 5 minutes to point us in the right direction, that would be highly appreciated. Maybe we could start a new dev thread on how to tinge objects in a patcherview tied to a feature request to include the tinge method in a future version of the SDK?

I can add that if flag 3 is set, the "tinge" method is returning a reference to the tinge which can be cleared  later with "fadetinge". 
void *result = NULL;
result = object_method(x->w_patcherview, _sym_tinge, 1, &a, &a2, 0, 0, 1);
object_method(x->w_patcherview, _sym_fadetinge, result);

But the 2 pointers are still confusing. It's interesting that you can exchange them (first empty, second holds objects) and it behaves exactly the same. Objects are tinged but the count is off. 

the signature that I get from fadetinge is:patcherview_fadetinge(patcherview *pv, void *result, long n);

So I bet that the extra long argument allows you to set the fade time.

Ok, we can't give up now then... we almost have it.
Email me off-list at: superbigio[at]yahoo[dot]com


happy new year !
...one year after I am still veryyyy entousiast with this external..

Did you managed to make it working ? maybe a beta version is publishabled ?

Hi Julien, happy new year. Wow how fast this year went over. I only have a Windows prototype. Other stuff prevented me from porting this to Mac. And a motivation killer is a bug with jit.pwindow which doesn't display properly in the second view of the device. I reported this as a bug to c74 and they promised to look into it but never got back at me.

happy new year ! ...one year after I am still veryyyy entousiast with this external..

EDIT : I am only on windows in case of Mac specific problem...

Hi, yes, I have a beta version working on both platforms. Contact me so I can send via email!
best,
O. 

11Olsen.
your external seems soooo promizing !
I have to say that without the help file, some stuff stay a bit obscur to me..*
But your amxd is very useful !


One thing in particular is strange for me  :

when the device is in floating mode and if i select another track (where my patch with your external is not), the float window doesn't report GUI changes, I can click for example on a slider on the floating window, if this slider moves a track volume in Live, i will see the track slider move in Live, but in the floating window, the slider doesn't move...but it works anyway. And if my floating window for example report a Live parameter, for example tempo : if I move the live tempo in Live while the track where my device is is selected, I can see in my floating window the report of the tempo change, but there is no report if I select another track...

I hope i am clear enough, sorry for my bad english...


@SCHLAM Do you have an example so I can reproduce the bug. I have not noticed something like frozen UI. 

In the little bpatcher with 2 live.dial and two sliders, add an observer of any live parameter.

Then if you open this bpatcher with your (wonderful=) external, 
the bpatcher in floating window only reports  the value if the track containing the device is selected.

Also while the track containing the device is not selected, if you move the dial controlling this parameter,
the corresponding value will move in Live but not in the floating window...
...until you select the track where the device is AND select the floating window..
EDIT: selecting the floating window (while another track is selected) doesn't refresh it until you select the track where 11.newview is..


Here is a picture of the bpatcher I tried, it only get/set the master volume.
can you reproduce ?

here window10 / Live 10.2.10b2 / max 8.5.2

Also, 
I didn"t manage too to move left or right the position of the scrollbar in the docked device.
It's surely because I didn't understand well how to set the rectangle I want to be shown in the  floating window..

I love too your trick with the 2 live.dials bigger but :
It should be great to do not have to click on a "colored pixel" of the dial to be able to move it.
It should be great too to choose if you want these big dials to be displayed or not when the Live app is focused or not


We know that times today can be tough, but with the right tools, you can grow your money and manage your wealth the same way that the wealthy have been doing for centuries. The only thing that separates the rich from the poor is the wisdom to know that you have the power to change your financial situation, anytime, anywhere. One investment at a time.
https://manifestmoneytips.com/

There are ways to prevent the click-through background but let me check the mapping thing first.

I can reproduce, it seems to be a special case with bpatcher. Thanks for spotting this! When the last view of a patcher containing the bpatcher is closed, the GUI in an additional opened view of the bpatcher isn't updated anymore.  
Hope add a fix/workaround to the external soon.
to move the visible patcher area of a view you have to use the visiblecanvasrect messages after the view is visible. If you mean scrolling the docked view you'll have to use the abstraction in the test device that makes this possible, otherwise Live resets the canvasrect everytime.
   
 

I noticed too that  [panel] 's property "drag window" is greyed in the inspector 

if for example you click several times on the first button of your test device, several floating views will be created, then it's only on the last created one that the red close button is working. For the others one, how do you close them if they have been opened with a "window flags notitle" argument, and the "windows10" close button is not available ?



Best is to make sure the user is not allowed to open multiple additional views. eg check the "isopen" output. The external can only be attached to one view at a time. But it is also possible to attach the object to already opened views and even other patchers. see "open_specials" subpatcher. You can use multiple instances of the external if you really need to handle 3 or 4 views of the same patcher at the same time (rare case, i think).

"notitle" windows are closed with a "close" msg to the object (provided that it is attached to the contained view)
   

thank you again!
let me explain what I am trying to achieve.
Maybe you can tell me some tricks...
I scratch my head so much..


I need 3 max patchers in my setup:

-Patcher1 is a maxpatch that is not linked with Live (I need it to exchange stuffs with Patcher2 in Live) it runs on standalone max8.5.2. it doesn't have to be concerned by 11newview ...
-Patcher2 is an audio device in Live 11.2.10b2
-Patcher3 is in Patcher2 (but maybe you will advice me to put it elsewhere..)


When the Live set is opened : 
-Patcher3 auto opens. It is a very little patcher with transparent bgcolor and two toggles (and a draggable panel) that sends "11newview" message to control Patcher2's view (the main device) . Patcher3 appears at the top of Live like 2 stock toggles in an empty space of the Live's toolbar.

-----toggle1 : open/close Patcher2 in floating view (and always patcher2 even if another patcher or window is focused..) .Patcher2 is opened by Patcher3 with some characteristics ; "Preset_A".... (zoomfactor X, size Y Z...etc...). or Preset_B etc..

this "open/close" toggle, recalls the last position of Patcher2 if it has been move, and recalls the characteristics of the last "Preset_X" chosen.

-----toggle2 : change the characteristics of Patcher2'view presets( _A, _B, _C)((zoomfactor, size...etc...) on the fly. (toggle1 always open/close patcher 2 whatever the preset and remember it.

I use since a longtime a working version of this mechanism but i don't manage to implement that with 11newview.. Each time I think I am close : little Patcher3 becomes the patcher controlled by 11newview. I don't manage that "Patcher2" allways stays the only patcher controlled by 11newview whatever I focus another window. There is only one 11newview in Patcher2. In patcher3 there is just toggles..

I tried some stuff in your special patcher but really don't understand well how to achieve that...

I cannot copy a patch for now, but maybe you have an example of this kind of routine ?

Too bad that you can't send me something as it is time consuming to build it from your description. If there's sensitive code just take it out.? But it sounds like an easy task. Simply use a separate 11newview for patcher 2 and 3 like you would do with [thispatcher].   

I will try to send you something today..
The main thing I don't understand is how to allways "bind" 11newview to a dedicated patcher whatever happens. I thought that banging "open Patcher2.maxpatch, open invisible etc....." will allways open Patcher2 when "Patcher2.maxpatch" is one of the outputs of the message "toppatcherlist". 
But while  patcher3 is opened once, the toggle open/close on it banging to "open Patcher2.maxpatch, open invisible etc.....", close/open itself and not Patcher2 anymore...

well...hard to talk without my computer..
thanks for your help !


changing the patcher focus is only for special cases. don't mess with it if you don't need to. The default patcher focus of the object is on the patcher containing it, put a newview object into each of the patchers. You really have to think of it as a [thispatcher] with some extra features, The second thing is the view focus/binding which is normally on the new created view. any view that was in focus before is released but not closed. 
But we could save many words if you send something to work with. Also I think the conversation is maybe not so interessting for forum members and we know each others email adr. ;) 


2nd patcherview window flags

Brad Garton

I haven't been keeping up with the forum lately, so perhaps this was answered (quick search didn't yield anything definitive).  Does using the 8.2.0 SDK require CMake, or can it also be used with XCode to build M1 objects?  Alternatively, can the 8.0.3 SDK be used to build M1 objects if it is set with the appropriate arch flags?

I haven't been keeping up with the forum lately, so perhaps this was answered (quick search didn't yield anything definitive).  Does using t

m1-externals-creation-8-2-08-0-3

Ha, never mind.  I think I see what to do now.  Silly me.

I am also silly, but an answer to your own question could help me to know why I am silly...
There are so many open source Max externals around, that I would love to see a step by step tutorial for silly people like me, how to build an external from a repository lets say from 2016 (Jamoma for example) for M1 and Max 8... XCode is installed...
That would be a good starting point...


M1 externals creation, 8.2.0/8.0.3

PETE KELLOCK

I’m trying write an external that has multiple voices, each voice driven by an ITM time object. I want each voice, 0 to NVOICES-1, to run independently of the others. I’m working from the info on ITM in Chapter 18 of the Max API doc, and adapting the delay2 code example that’s shipped with the Max sdk.

I’m able to create an array of time objects, one per voice, declaring it in the master structure of the external (called reA) as:
t_object *d_timeObj[NVOICES];

…and I can declare the callback function:
void reA_doOne(t_reA *x);

I’m then able to instantiate these giving a single callback function reA_doOne() as follows:
for (voiceNo=NVOICES-1; voiceNo>=0; voiceNo--)  {
x->d_timeObj[voiceNo] = (t_object *) time_new((t_object *)x, gensym("delaytime"), (method)reA_doOne, TIME_FLAGS_TICKSONLY | TIME_FLAGS_USECLOCK);
}

That all compiles and runs, but it’s not quite enough. When reA_doOne() gets called, it doesn’t know which time object in the array (ie which voice ) is calling it.  Hmm.

So I tried the following...
Declaration:  
void reA_doOne(t_reA *x, long voiceNo);

Instantiation of the time object:
for (voiceNo=NVOICES-1; voiceNo>=0; voiceNo--)  {
x->d_timeObj[voiceNo] = (t_object *) time_new((t_object *)x, gensym("delaytime"), (method)reA_doOne(*x, voiceNo), TIME_FLAGS_TICKSONLY | TIME_FLAGS_USECLOCK);
}

But that fails.  It seems, not too surprisingly, that the time_new() function will only accept a method name reA-doOne; it won’t accept adding arguments to that method in the call. The compiler error I get is:
Semantic Issue Group
Passing 't_reA' (aka 'struct reA') to parameter of incompatible type 't_reA *' (aka 'struct reA *'); remove *
Passing argument to parameter 'x' here

I’ve tried lots of variants of this, but haven’t found a way yet. 
Again, all ll I need is that when the reA_doOne() function gets called, I know which timeObject in the array of time objects (ie which voice) called it. 

My fallback is to create a slew of explicitly-named functions, one for each voice, reA_doOne00(), reA_doOne01(), reA_doOne02() etc, each set up as the callback function for one timeObject. But that's ugly and doesn't allow the number of voices to be changed easily.

I’m trying write an external that has multiple voices, each voice driven by an ITM time object. I want each voice, 0 to NVOICES-1, to run in

max-sdk-t-timeobject-–-creating-a-callback-function-that-includes-a-voice-number

Max SDK t-timeobject – creating a callback function that includes a voice number

Iain Duncan

I'm  trying to figure out how to create and object and have it callback into  my external's instance. I have successfully created some objects and  connected them as detailed here https://cycling74.com/sdk/max-sdk-8.2.0/chapter_scripting.html#chapter_scripting_objects 

But I can't figure out how  to connect one of them to my external itself. I guess I need the  t_object for my object. However, passing in x->obj (which is what I call my obj in its struct) doesn't do it. Any  suggestions or examples would be lovely. Below is what is not working, I get an error msg:

"zgetfn (patchlineupdate): corrupt object"


// create the prepend id object, which needs to be connected to this s4m instances inlet 0
    t_object *prepend_id;
    prepend_id = newobject_sprintf(x->patcher, "@maxclass newobj @text \"prepend live-api-id\" @hidden 0 @patching_position %.2f %.2f", 0, 0);
    // connect to self...
    t_atom msg[5], rv;
    atom_setobj(msg, prepend_id);         // source
    atom_setlong(msg + 1, 0);             // outlet number (0 is leftmost)
 

   // the below is not working
    t_object *this_box;
    this_box = (t_object *)&x->obj;
    atom_setobj(msg + 2, this_box);         // destination is self

   atom_setlong(msg + 3, 0);             // inlet number (0 is leftmost)
    atom_setlong(msg + 4, 1);             // optional 5th arg, if neg cable is hidden
    object_method_typed(x->patcher, gensym("connect"), 5, msg, &rv);

I'm  trying to figure out how to create and object and have it callback into  my external's instance. I have successfully created some objec

stuck-trying-to-make-an-object-connected-to-this-external

I think as obj is the first member simply using x points to your object. But maybe you need the box in your case? Just guessing..

It also doesn't work for me. Forget my comment about the box. We get the box by default, and can get the object from it when needed (jbox_get_object). But we need the box for the connect message. 
and this works:
t_object* this_box = (t_object*)object_method(x->patcher, gensym("getnamedbox"), gensym("myname")); 

so using (t_object*)x or (t_object*)&x->obj seems wrong but there has to be a way 

Solved! I needed to get the box with object_obex_lookup.

Cycling '74 Forums

Hi everyone! I'm Esteban and I've been using the min devkit for a little while now, getting nice results building externals using the Eigen ...

that is a good question if it is runtime wide, and i would have to speculate, and to be honest i assumed that was long fixed. for music and ...

Hi all, Not sure but I think this project might help out, I worked on it as well, and got finally a swift/c++ thing working.

Following up on my own post, here's what I did to accomplish the task for those in the future.

Thanks, 110L532. I kept on searching through the forums and found the answer in this one.

The help file for the object is pretty self-explanatory, isn't it? Or do you mean how to use 3rd party externals in general?

changing the patcher focus is only for special cases. don't mess with it if you don't need to.

I am also silly, but an answer to your own question could help me to know why I am silly... There are so many open source Max externals ...

I’m trying write an external that has multiple voices, each voice driven by an ITM time object.

oh well, I hope i can remember this one if I ever need the box. cheers