From bug-request at octave dot org Fri Feb  4 11:03:43 2005
Subject: Re: Should be able to say 'for (k=1:N)'
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: Paul Kienzle <pkienzle at users dot sourceforge dot net>
Cc: Julius Smith <jos at ccrma dot stanford dot edu>, bug@octave.org
Date: Fri, 4 Feb 2005 11:07:56 -0600

On  4-Feb-2005, Paul Kienzle <pkienzle at users dot sourceforge dot net> wrote:

| This is not a parser inconsistency.
| 
| "for k = ..." is special syntax in the language.
| 
| The consistent thing is to consider (k=1:3) to be a value.
| 
| Whatever you expect from the statement:
| 
| 	for 1:3, ... end
| 
| you should get the same thing from
| 
| 	for (k=1:3), ... end
| 
| The following does look weird, but it is consistent:
| 
| 	for k=x=1:3, k, x, end
| 
| It assigns x to [1,2,3], then lets k cycle over x.

Yes, this is all correct.  However, Matlab does seem to allow

  for VAR = VAL ... end

and

  for ( VAR = VAL ) ... end

Both are apparently special forms in the language because it does not
allow

  for ( ( VAR = VAL ) ) ... end

I think this is consistent behavior in Matlab.  It does not allow

  a = b = c;

because an assignment is a statement, not an expression that produces
a value.

I admit that I have never before seen anyone use

  for (k = 1:N) ...

in Matlab code, so I would guess it is relatively rare.

The change I made to handle this form only allows the additional
special case

  for ( VAR = VAL ) ... end

to work.

Note that currently

  for VAL ... end

is not valid, and

  for VAR = VAL ... end

with VAL being an assignment expression still works after my change,
so your

  for k=x=1:3 k, x, end

code will still work.

I'm in favor of keeping the change for compatibility, but am open to
other suggestions.

jwe



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------