Re: Default arguments

[John W. Eaton]

On 18-Dec-2006, Søren Hauberg wrote:

| Søren Hauberg skrev:
| >> Although I've checked this patch in to the CVS archive, I consider it
| >> to be quite experimental.  If any nontrivial problems show up that are
| >> related to this change, I'll simply remove it.
| > I can't really test this until sometime next week (I have a hard 
| > deadline on monday), but then I'll report back...
| Okay, I've done a CVS checkout to play around with your changes. I 
| haven't found any real problems, but there is one thing, that _might_ 
| cause problems.
| 
| The following function works:
| 
| function out = myfun(a = 2, b = sin(a))
|    out = a + b;
| endfunction
| 
| but this doesn't:
| 
| function out = myfun(a = sin(b), b = 2)
|    out = a + b;
| endfunction
| 
| Now, I can understand why this is the case, but I'm affraid that it 
| could cause confusion, i.e. "why doesn't my function work anymore? I've 
| just switched the order of the input arguments."
| 
| Otherwise, it appears to be working quite nicely,

I suppose we should

  * document that function arguments are evaluated left to right,
    which means that the first case is OK but second is not

or

  * document that the order of evaluation of function arguments is
    undefined (more freedom for the interpreter/compiler) in which
    case neither can be expected to work

If we choose the second case, it might be nice to have a warning or
error but I don't see an easy way to make that happen.

jwe