metamax

I can't get any params within a GenExpr function to change. Even the example in the GenExpr guide doesn't work.

func(v)  {    
Param add(0);
return v+add;
}
out1 = func(in1);


Sending [jit.gen] a float prepended with "add" does nothing. Global params work fine.

Been at this for hours. Feedback is most appreciated. Cryptic hints, random jokes and non sequiturs are also welcome.

I can't get any params within a GenExpr function to change.  Even the example in the GenExpr guide doesn't work.   

func(v)  {	
Param add(0);
return v+add;
}
out1 = func(in1);


Sending [jit.gen] a float prepended with "add" does nothing.     Global params work fine.

Been at this for hours.  Feedback is most appreciated.  Cryptic hints, random jokes and non sequiturs are also welcome.

dynamic-parameters-within-genexpr-functions

I figured out awhile back by trial and error(i put 'Param's before functions and got an error message: "codebox: function definitions not allowed here (function definitions must precede all statements)"), that you can't put Params inside functions, instead you have to rewrite the function so that one of its arguments takes the place of the param, and then call the function later with the params name where the argument is... sorry, hard to explain... basically just rewrite your jit.gen like this:

func(v, a)  {
return v+a;
}
Param add(0);
out1 = func(in1, add);

Hey thanks. That helps. Although it still looks like a bug because the example given in the GenExpr manual simply does not work. Not to mention, I am building a library of operators to use with require(). ex:

...but I am dealing with matrix functions that have 9 or more param values. ex.

vec( c.r*swiz(m,0), c.g*swiz(m,1), c.b*swiz(m,2), c.r*swiz(m,3), c.g*swiz(m,4), c.b*swiz(m,5), c.r*swiz(m,6), c.g*swiz(m,7), c.b*swiz(m,8), c.a );

Yet, I can't use the following (outside of the function):

out1 = matrix(in1.r, in1.g, in1.b, in1.a, swiz(mX,0), swiz(mX,1), swiz(mX,2), swiz(mX,3), swiz(mX,4), swiz(mX,5), swiz(mX,6), swiz(mX,7), swiz(mX,8) );

Hey thanks.   That helps.   Although it still looks like a bug because the example given in the GenExpr manual simply does not work.   Not to mention, I am building a library of operators to use with require().  ex:

require("func");
Param name(0);
out1 = func(in1, name);

...but I am dealing with matrix functions that have 9 or more param values.  ex.

func(c, m)  {
vec( c.r*swiz(m,0), c.g*swiz(m,1), c.b*swiz(m,2), c.r*swiz(m,3), c.g*swiz(m,4), c.b*swiz(m,5), c.r*swiz(m,6), c.g*swiz(m,7), c.b*swiz(m,8), c.a );
}

Yet, I can't use the following (outside of the function):
Param mX(1, 0, 0, 0, 1, 0, 0, 0, 1);
out1 = func(in1, elements);

I have to use:
Param mX(1, 0, 0, 0, 1, 0, 0, 0, 1);
out1 = matrix(in1.r, in1.g, in1.b, in1.a, swiz(mX,0), swiz(mX,1), swiz(mX,2), swiz(mX,3), swiz(mX,4), swiz(mX,5), swiz(mX,6), swiz(mX,7), swiz(mX,8) );

Although it still looks like a bug because the example given in the GenExpr manual simply does not work.

ya, i hear ya, that wasn't quite sitting right with me either. so i went back and in order to make your code jive better with the GenExpr doc example, i realized the GenExpr example assigns something to the Param of the function within the function arguments when it is called later...

func(v, amp=1) {
   Param offset(0, min=-1, max=1);
   return amp*(v+offset);
}
out1 = func(in1, amp=0.5, offset=in1*2);


^notice the 'offset = in1*2' there, seems necessary for it to work. needs to be assigned('=') to something, doesn't just take input as an argument, nor does it seem to work like outer 'Param's where an outer message will register within the func

(edit: words fail me, but to put it another way, the documentation seems to state that a 'param' within a function works differently than a 'param' outside a function... the 'param' within a function is referred to in the doc as an 'attribute')

so it'd look more like this if you want it to work more like the GenExpr doc example:

ya, i hear ya, that wasn't quite sitting right with me either. so i went back and in order to make your code jive better with the GenExpr doc example, i realized the GenExpr example assigns something to the Param of the function within the function arguments when it is called later... 

^notice the 'offset = in1*2' there, seems necessary for it to work. needs to be assigned('=') to something, doesn't just take input as an argument, nor does it seem to work like outer 'Param's where an outer message will register within the func
(edit: words fail me, but to put it another way, the documentation seems to state that a 'param' within a function works differently than a 'param' outside a function... the 'param' within a function is referred to in the doc as an 'attribute')

so it'd look more like this if you want it to work more like the GenExpr doc example:


Raja, thanks again for the insight. Your solution is great... though there does appear to be a bug in Gen. Examples from the documentation generate errors. See patch below.

Also, regarding your workaround, it's not necessary to assign an attribute to access the function param. It's just a matter of creating a global param and equating it with the function param in the output arg.

Raja, thanks again for the insight.  Your solution is great... though there does appear to be a bug in Gen.    Examples from the documentation generate errors.   See patch below.

Also, regarding your workaround, it's not necessary to assign an attribute to access the function param.    It's just a matter of creating a global param and equating it with the function param in the output arg.   

Where.... where did raja go in this thread! Did our resident asshole delete his posts?!

Where.... where did raja go in this thread!  Did our resident asshole delete his posts?!

maybe his creative use of ascii and unicode symbols was too much for the backup script?

Maybe he used his alien ray to suck up all his old emails. Ever since they ported the email list to this forum it seems like random posts go missing.

Maybe he used his alien ray to suck up all his old emails.   Ever since they ported the email list to this forum it seems like random posts go missing.

(yes, i deleted almost everything from before 2017 and started over: cleaning up my act 😇

in this particular thread, i remember the last post by the OP had all the relevant answers needed for future viewers so i felt there was no harm💆‍♂️💆‍♀️🤗)

this post shall self-destruct at some random point in the future.

 👻
(yes, i deleted almost everything from before 2017 and started over: cleaning up my act 😇
in this particular thread, i remember the last post by the OP had all the relevant answers needed for future viewers so i felt there was no harm💆‍♂️💆‍♀️🤗)

this post shall self-destruct at some random point in the future.
👽

Dynamic parameters within GenExpr functions?