| Re: endless loops suck [message #33820] |
Mon, 03 February 2003 17:57  |
Thomas Gutzler
Messages: 44
Registered: November 2002
|
Member |
|
|
Pete wrote:
> Hi,
> I've come across this problem before when updating plots.
>
> This will not not break:
>
> PRO doomed_to_continue
> a=0L
> While (1) DO BEGIN
> a=a+1
> ENDWHILE
> END
>
> This will break:
>
> PRO happy_to_break
> a=0L
> While (1) DO BEGIN
> a=a+1
> wait,0.0001
> ENDWHILE
> END
>
> as will this:
>
> PRO happy_to_break_2
> a=0L
> While (1) DO BEGIN
> a=a+1
> print,''
> ENDWHILE
> END
Weird!
I'm starting to put some waits meta-info-prints in my loops. This spawns
a new question:
How is this possible ?
FOR i=0,4 DO PRINT, i, ' ', /NONEWLINE
^^^^^^^^^^ this may be a problem
Result:
0 1 2 3 4
I couldn't find anything usefull in the help. If this isn't possible I
have to use the wait :/
thx,
Tom
|
|
|
|
| Re: endless loops suck [message #33823 is a reply to message #33820] |
Mon, 03 February 2003 15:36  |
Pete[2]
Messages: 3
Registered: February 2003
|
Junior Member |
|
|
Hi,
I've come across this problem before when updating plots.
This will not not break:
PRO doomed_to_continue
a=0L
While (1) DO BEGIN
a=a+1
ENDWHILE
END
This will break:
PRO happy_to_break
a=0L
While (1) DO BEGIN
a=a+1
wait,0.0001
ENDWHILE
END
as will this:
PRO happy_to_break_2
a=0L
While (1) DO BEGIN
a=a+1
print,''
ENDWHILE
END
At times, I've had to put in a wait statement after a plot statement to make
it display, if the plot statement is in a loop and if there is no other way
to interrupt the excecution of the loop. Weird!
Peter
"Thomas Gutzler" <tgutzler@ee.uwa.edu.au> wrote in message
news:3E3DDF77.3040908@ee.uwa.edu.au...
> good morning,
>
> David Fanning wrote:
>>
>> So I admit, "infinite" WHILE loops *can* be interrupted, as long as
>> they use BEGIN ... END statement blocks. Try this:
>>
>> PRO Test
>> While 1 DO BEGIN
>> Print, 'Test before you Post!'
>> ENDWHILE
>> END
>
> I thought about this and figured out that I have a loop with begin
> and end. It looks like that:
>
> PRO SnakeInterp,xi,yi,dmax,dmin
> ; x,y are arrays of coordinates
> ; dmin, dmax are preferred min and max distance between coordinates
> [...]
> ; WHILE (MAX(d) GT dmax) DO BEGIN ; this would be correct
> WHILE (1) DO BEGIN ; this obviously loops to dead
> idx0 = (d GT dmax)
> => z = SnakeIndex(idx0) ; debugger is here
> p = INDGEN(N+1)+1
> xi = InterPol([xi,xi[0]],p,z)
> yi = InterPol([yi,yi[0]],p,z)
> N = N_ELEMENTS(xi)
> d = abs( [xi[1:N-1],xi[0]] - xi) + abs( [yi[1:N-1],yi[0]] - yi)
> ENDWHILE
> END
>
> FUNCTION snakeindex, idx
> ; SNAKEINDEX Create index for adpative interpolating the snake
> arr = idx
> N = N_ELEMENTS(xi)
> arr = (TRANSPOSE([[arr],[INTARR(N)+1]]))[0:2*N-2]
> y = DOUBLE(WHERE([1,arr] EQ 1))/2+1
> RETURN, y
> END
>
> I pressed "Step out", then CTRL+BREAK (10 times) then CTRL+c (another 10
> times) then I made myself a tea and read some news. After half an hour
> taskmanager was my friend again :(
> I thought maybe the compiler needs a little bit of time to finish
> currently running processes before it polls the keyboard again and breaks.
>
>> On my Windows machine I can interrupt this program (after I've
>> learned my lesson) by doing a CNTL-Break.
>
> I could interrupt the Print, 'Test before you Post!' loop with this, yes.
>
> Perhaps I just should pay more attention to my sourcecode.
>
> Tom
>
|
|
|
|
| Re: endless loops suck [message #33827 is a reply to message #33823] |
Sun, 02 February 2003 19:18 |
Thomas Gutzler
Messages: 44
Registered: November 2002
|
Member |
|
|
good morning,
David Fanning wrote:
>
> So I admit, "infinite" WHILE loops *can* be interrupted, as long as
> they use BEGIN ... END statement blocks. Try this:
>
> PRO Test
> While 1 DO BEGIN
> Print, 'Test before you Post!'
> ENDWHILE
> END
I thought about this and figured out that I have a loop with begin
and end. It looks like that:
PRO SnakeInterp,xi,yi,dmax,dmin
; x,y are arrays of coordinates
; dmin, dmax are preferred min and max distance between coordinates
[...]
; WHILE (MAX(d) GT dmax) DO BEGIN ; this would be correct
WHILE (1) DO BEGIN ; this obviously loops to dead
idx0 = (d GT dmax)
=> z = SnakeIndex(idx0) ; debugger is here
p = INDGEN(N+1)+1
xi = InterPol([xi,xi[0]],p,z)
yi = InterPol([yi,yi[0]],p,z)
N = N_ELEMENTS(xi)
d = abs( [xi[1:N-1],xi[0]] - xi) + abs( [yi[1:N-1],yi[0]] - yi)
ENDWHILE
END
FUNCTION snakeindex, idx
; SNAKEINDEX Create index for adpative interpolating the snake
arr = idx
N = N_ELEMENTS(xi)
arr = (TRANSPOSE([[arr],[INTARR(N)+1]]))[0:2*N-2]
y = DOUBLE(WHERE([1,arr] EQ 1))/2+1
RETURN, y
END
I pressed "Step out", then CTRL+BREAK (10 times) then CTRL+c (another 10
times) then I made myself a tea and read some news. After half an hour
taskmanager was my friend again :(
I thought maybe the compiler needs a little bit of time to finish
currently running processes before it polls the keyboard again and breaks.
> On my Windows machine I can interrupt this program (after I've
> learned my lesson) by doing a CNTL-Break.
I could interrupt the Print, 'Test before you Post!' loop with this, yes.
Perhaps I just should pay more attention to my sourcecode.
Tom
|
|
|
|
| Re: endless loops suck [message #33831 is a reply to message #33827] |
Sat, 01 February 2003 08:14 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
I wrote yesterday:
> I don't think so. The whole point of the BEGIN ... END syntax
> is to make the statements inside appear to be a single statement
> to the compiler. The only recourse is to let the loop go to
> completion. With an infinite loop, well, they don't call
> them "infinite" for no reason. :-)
Apparently I owe an apology to Pavel. It has been
pointed out to me by a long-time IDL newsgroup reader,
who wishes not to embarrass me, that I don't know what
the hell I'm talking about when it comes to this topic.
(Have you noticed that IDL newsgroup readers are exceptionally
kind to one another? It's one of the things that makes
this newsgroup so much fun to hang out in. :-)
So I admit, "infinite" WHILE loops *can* be interrupted, as
long as they use BEGIN ... END statement blocks.
Try this:
PRO Test
While 1 DO BEGIN
Print, 'Test before you Post!'
ENDWHILE
END
On my Windows machine I can interrupt this program (after
I've learned my lesson) by doing a CNTL-Break.
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: endless loops suck [message #33835 is a reply to message #33831] |
Fri, 31 January 2003 11:38 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Pavel A. Romashkin (pavel_romashkin@hotmail.com) writes:
> I think you can stop a loop if you use BEGIN ... END syntax. Without it,
> the loop is compiled into one command and is not interrupted until
> finished; with it, each item inside is a separate command, so IDL will
> check for breaks every time. I am not sure that it affects the speed at all.
I don't think so. The whole point of the BEGIN ... END syntax
is to make the statements inside appear to be a single statement
to the compiler. The only recourse is to let the loop go to
completion. With an infinite loop, well, they don't call
them "infinite" for no reason. :-)
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
|
|
|
|
| Re: endless loops suck [message #33899 is a reply to message #33823] |
Tue, 04 February 2003 09:11 |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
It must be platform-specific. On my Alpha workstation, running IDL/v5.4, I can
break out of all of these loops with control-C. On the other hand, with the
same version of IDL under Windows 95, I could break out of happy_to_break with
Control-Break, but with doomed_to_continue, I had to use Control-Alt-Delete to
end the IDL task!
On a somewhat related note, I was reminded of another program where putting in
a short call to wait turned out to be beneficial. A telemetry processing
program I had written was eating up all the CPU on the machine, so that other
processes were not getting their fair share. I'm not sure why this was;
perhaps because it was so I/O intensive the computer was boosting its priority.
In any case, simply adding a "WAIT, 0.005" statement for each packet solved the
problem, without significantly impacting the speed of the program itself. The
program ran just as fast as before when it had the machine to itself, but
allowed other processes to also get their fair share of the clock cycles.
Bill Thompson
"Pete" <peterscarth@despammed.com> writes:
> Hi,
> I've come across this problem before when updating plots.
> This will not not break:
> PRO doomed_to_continue
> a=0L
> While (1) DO BEGIN
> a=a+1
> ENDWHILE
> END
> This will break:
> PRO happy_to_break
> a=0L
> While (1) DO BEGIN
> a=a+1
> wait,0.0001
> ENDWHILE
> END
> as will this:
> PRO happy_to_break_2
> a=0L
> While (1) DO BEGIN
> a=a+1
> print,''
> ENDWHILE
> END
> At times, I've had to put in a wait statement after a plot statement to make
> it display, if the plot statement is in a loop and if there is no other way
> to interrupt the excecution of the loop. Weird!
> Peter
> "Thomas Gutzler" <tgutzler@ee.uwa.edu.au> wrote in message
> news:3E3DDF77.3040908@ee.uwa.edu.au...
>> good morning,
>>
>> David Fanning wrote:
>>>
>>> So I admit, "infinite" WHILE loops *can* be interrupted, as long as
>>> they use BEGIN ... END statement blocks. Try this:
>>>
>>> PRO Test
>>> While 1 DO BEGIN
>>> Print, 'Test before you Post!'
>>> ENDWHILE
>>> END
>>
>> I thought about this and figured out that I have a loop with begin
>> and end. It looks like that:
>>
>> PRO SnakeInterp,xi,yi,dmax,dmin
>> ; x,y are arrays of coordinates
>> ; dmin, dmax are preferred min and max distance between coordinates
>> [...]
>> ; WHILE (MAX(d) GT dmax) DO BEGIN ; this would be correct
>> WHILE (1) DO BEGIN ; this obviously loops to dead
>> idx0 = (d GT dmax)
>> => z = SnakeIndex(idx0) ; debugger is here
>> p = INDGEN(N+1)+1
>> xi = InterPol([xi,xi[0]],p,z)
>> yi = InterPol([yi,yi[0]],p,z)
>> N = N_ELEMENTS(xi)
>> d = abs( [xi[1:N-1],xi[0]] - xi) + abs( [yi[1:N-1],yi[0]] - yi)
>> ENDWHILE
>> END
>>
>> FUNCTION snakeindex, idx
>> ; SNAKEINDEX Create index for adpative interpolating the snake
>> arr = idx
>> N = N_ELEMENTS(xi)
>> arr = (TRANSPOSE([[arr],[INTARR(N)+1]]))[0:2*N-2]
>> y = DOUBLE(WHERE([1,arr] EQ 1))/2+1
>> RETURN, y
>> END
>>
>> I pressed "Step out", then CTRL+BREAK (10 times) then CTRL+c (another 10
>> times) then I made myself a tea and read some news. After half an hour
>> taskmanager was my friend again :(
>> I thought maybe the compiler needs a little bit of time to finish
>> currently running processes before it polls the keyboard again and breaks.
>>
>>> On my Windows machine I can interrupt this program (after I've
>>> learned my lesson) by doing a CNTL-Break.
>>
>> I could interrupt the Print, 'Test before you Post!' loop with this, yes.
>>
>> Perhaps I just should pay more attention to my sourcecode.
>>
>> Tom
>>
|
|
|
|
| Re: endless loops suck [message #33900 is a reply to message #33823] |
Tue, 04 February 2003 09:13 |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
"Pete" <peterscarth@despammed.com> writes:
> At times, I've had to put in a wait statement after a plot statement to make
> it display, if the plot statement is in a loop and if there is no other way
> to interrupt the excecution of the loop. Weird!
Have you tried using EMPTY instead?
Bill Thompson
|
|
|
|
| Re: endless loops suck [message #33915 is a reply to message #33820] |
Tue, 04 February 2003 01:14 |
Timm Weitkamp
Messages: 66
Registered: August 2002
|
Member |
|
|
Today at 09:57 +0800, Thomas Gutzler wrote:
> How is this possible ?
>
> FOR i=0,4 DO PRINT, i, ' ', /NONEWLINE
> ^^^^^^^^^^ this may be a problem
> Result:
> 0 1 2 3 4
Use a format string (see also "Format codes" in the online help), like
this:
FOR i=0,4 DO PRINT, i, FORMAT='($,I1," ")'
PRINT
Cheers,
Timm
--
Timm Weitkamp <http://people.web.psi.ch/weitkamp>
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
