From owner-help-octave at bevo dot che dot wisc dot edu  Fri Nov 15 18:30:55 1996
Subject: "plot" within function def
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: Ted dot Harding at nessie dot mcc dot ac dot uk (Ted Harding)
Cc: help-octave at bevo dot che dot wisc dot edu
Date: Fri, 15 Nov 1996 18:30:25 -0600

On 15-Nov-1996, Ted Harding <Ted dot Harding at nessie dot mcc dot ac dot uk> wrote:

: I don't know if this is a bug: I can only describe the behaviour.
: The following is a file: call it "test.m"; its purpose is to define a
: function "config".

It is probably a bug in something, but I'm not sure what.

: Now for the fun bit.
: 
: 1. Start octave from scratch.
: 2. Do "test" to read in the above file.
: 3. Do "config".
:       You will see ONLY ONE TRIANGLE (the first).
: 4. Do "config" again.
:       You will now see TWO TRIANGLES, as intended. And the same for every
:       repetition.
: 5. Quit octave.
: 
: Next, edit "test.m" so as to uncomment the line "plot([0 1],[0 1])".
: Then repeat the above. NOW YOU WILL SEE TWO TRIANGLES FIRST TIME, as
: intended.
: 
: It therefore seems that the "plot" command, when it occurs inside a
: function definition as above, must first be "primed" by being executed
: before the the function is called.
: 
: There it is: any comments?

I was able to observe this behavior, but I got inconsistent results.
After poking around for a while, it seems that the strings that are
sent to the plotter stream are ok, and that they are being flushed, so
I don't know why the display is not always updated.

Anyway, here is a patch that seems to work around the problem,
whatever it really is.

*** scripts/plot/plot_int.m~	Sun Oct  8 23:26:11 1995
--- scripts/plot/plot_int.m	Fri Nov 15 18:24:40 1996
***************
*** 81,86 ****
--- 81,88 ----
        endif
        endif
  
+       replot;
+ 
      unwind_protect_cleanup
  
        if (! hold_state)


Hmm,

jwe