function graphics curiosity? bug?

[Jonathan]

Have a look at the following code:

x = findgen(128)
y = 1.0 + 0.1*randomn(seed,128)
y2 = 1.0 + 0.2 randomn(seed,128)
b = widget_base( xsize=480, xoffset=940, ysize=360, yoffset=0 )
w = widget_window( b, x_scroll_size=470, y_scroll_size=350 )
widget_control, b, /realize
widget_control, w, get_value=d
p = plot( x, y, current=d, xstyle=2, ystyle=2 )
p2 = plot( x, y2, color='red', linestyle='', symbol='+', current=d, /overplot )
d.uvalue = { x:x, y:y, y2:y2, b:b, p:p, p2:p2 }
end

This creates a widget window, w, under a base, b, and then places two overlapping plots within that window. Now, if the data changes for the second plot, I would like to do the following steps:

y3 = 1.0 + 0.2 randomn(seed,128) ; new data
p2.delete ; erases the p2 data in the plot
p2 = plot( x, y3, color='red, , linestyle='', symbol='+', current=d, /overplot )
d.uvalue.p2 = p2 ; store the plot identifier in the window's uvalue structure

The last line generates the following error message:
% Attempt to store into an expression: Structure reference.
% Execution halted at: $MAIN$
What has happened is that IDL forgot the type of d.uvalue.p2, so when I try to put a new (identical) p2 there, it rejects the attempt.

This turns out to be a huge hassle for me.
My solution is a kludge, which is to create a new structure for d.uvalue and replace the whole thing, rather than just one element.

Is there a better, simpler way?

[Markus Schmassmann]

Post by Jonathan
x = findgen(128)
y = 1.0 + 0.1*randomn(seed,128)
y2 = 1.0 + 0.2 randomn(seed,128)
b = widget_base( xsize=480, xoffset=940, ysize=360, yoffset=0 )
w = widget_window( b, x_scroll_size=470, y_scroll_size=350 )
widget_control, b, /realize
widget_control, w, get_value=d
p = plot( x, y, current=d, xstyle=2, ystyle=2 )
p2 = plot( x, y2, color='red', linestyle='', symbol='+', current=d, /overplot )
d.uvalue = { x:x, y:y, y2:y2, b:b, p:p, p2:p2 }
end
This creates a widget window, w, under a base, b, and then places
two overlapping plots within that window. Now, if the data changes for the
y3 = 1.0 + 0.2 randomn(seed,128) ; new data
p2.delete ; erases the p2 data in the plot
p2 = plot( x, y3, color='red, , linestyle='', symbol='+', current=d, /overplot )
d.uvalue.p2 = p2 ; store the plot identifier in the window's uvalue structure
% Attempt to store into an expression: Structure reference.
% Execution halted at: $MAIN$
What has happened is that IDL forgot the type of d.uvalue.p2, so
when I try to put a new (identical) p2 there, it rejects the attempt.
This turns out to be a huge hassle for me.
My solution is a kludge, which is to create a new structure for
d.uvalue and replace the whole thing, rather than just one element.
Is there a better, simpler way?

; a simpler way:
p2.putData, y3

; ps: a '*' is missing in the definition of y2 & y3
y2 = 1.0 + 0.2 * randomn(seed,128)
y3 = 1.0 + 0.2 * randomn(seed,128)
; |

Directly updating the values is also much faster than creating a new
plot. I hope this help,

Markus

[Jonathan]

Post by Markus Schmassmann

Post by Jonathan
x = findgen(128)
y = 1.0 + 0.1*randomn(seed,128)
y2 = 1.0 + 0.2 randomn(seed,128)
b = widget_base( xsize=480, xoffset=940, ysize=360, yoffset=0 )
w = widget_window( b, x_scroll_size=470, y_scroll_size=350 )
widget_control, b, /realize
widget_control, w, get_value=d
p = plot( x, y, current=d, xstyle=2, ystyle=2 )
p2 = plot( x, y2, color='red', linestyle='', symbol='+', current=d, /overplot )
d.uvalue = { x:x, y:y, y2:y2, b:b, p:p, p2:p2 }
end
This creates a widget window, w, under a base, b, and then places
two overlapping plots within that window. Now, if the data changes for the
y3 = 1.0 + 0.2 randomn(seed,128) ; new data
p2.delete ; erases the p2 data in the plot
p2 = plot( x, y3, color='red, , linestyle='', symbol='+', current=d, /overplot )
d.uvalue.p2 = p2 ; store the plot identifier in the window's uvalue structure
% Attempt to store into an expression: Structure reference.
% Execution halted at: $MAIN$
What has happened is that IDL forgot the type of d.uvalue.p2, so
when I try to put a new (identical) p2 there, it rejects the attempt.
This turns out to be a huge hassle for me.
My solution is a kludge, which is to create a new structure for
d.uvalue and replace the whole thing, rather than just one element.
Is there a better, simpler way?

p2.putData, y3
; ps: a '*' is missing in the definition of y2 & y3
y2 = 1.0 + 0.2 * randomn(seed,128)
y3 = 1.0 + 0.2 * randomn(seed,128)
; |
Directly updating the values is also much faster than creating a new
plot. I hope this help,
Markus

That answers my question. Thank you.
I don't know why this is not made clear in the documentation, nor are there examples.