Re: PATCH: 3.1.6-pws-4: floating point support

[Peter Stephenson <pws@ibmth.df.unipi.it>]

"Bart Schaefer" wrote:
> What's the most appropriate idiom for forcing a floating point number to be
> interpreted as an integer, e.g. to force integer division?
> 
> The thing that bothers me most -- and I don't know what if anything ksh does
> about this, so maybe it's moot -- is that
> 
> 	((x=y))
> 
> where y is float and x was not previously defined, creates x as float, which
> may be unexpected (if the assignment is in code that's held over from pre-
> float zsh).

Well, if y is a float but holds an integer you can get away with

(( x = $(( y )) ))

because the %g format used for outputting with $(( y )) doesn't use the
decimal point unless it's necessary.  But there's no mechanism for a cast
other than making sure your parameters are declared appropriately.

> While doing some random fooling around with this, I noticed:
> 
> zagzig<23> ((integer florp=9.2))
> zsh: bad math expression: unbalanced stack
> zagzig<24> typeset -F
> florp=9.2000000000
> 
> The variable got assigned in spite of the syntax error?  Ouch.

The math parser is rather a hack; there's never been any proper syntax
checking, which is why you always get that meaningless (to the user) error
message.  What happens here is that `integer' gets put on the stack as a
parameter, then so does florp.  Then when = is found, its right hand side
is evaluated and the operator called.  At that point, that operation is
finished, so the parser goes back and finds it's now got `parameter value'
on the stack with no operator.

I'll see if I can think of something.

-- 
Peter Stephenson <pws@ibmth.df.unipi.it>       Tel: +39 050 844536
WWW:  http://www.ifh.de/~pws/
Dipartimento di Fisica, Via Buonarroti 2, 56127 Pisa, Italy