Re: number->string of float with radix

[Marius Vollmer]

Bill Schottstaedt <address@hidden> writes:

> standard input:1:1: Wrong type argument in position 2 (expecting INUMP): 2.0
> 
> "INUMP" is internal jargon -- "integer" or "exact integer" would be better.

Yes, I agree.  Our error messages suck generally, I'd say.  If you
want to work on this, I'd say we should define a new macro

    #define SCM_MAKE_VALIDATE2(pos, val, test, msg) \
      do { \
        SCM_ASSERT_TYPE (test, val, pos, FUNC_NAME, msg); \
      } while (0)

and then incrementally replace uses of SCM_MAKE_VALIDATE with improved
versions.  Or use SCM_ASSERT_TYPE directly.  Hmm.

> Also, I wonder about these kinds of calls:
> 
> guile> (number->string 10.5 2)
> "10.5"
> 
> It's actually not senseless to have base-2 float notation ("1010.1",
> "a.8" hex), but if Guile just returns base 10, maybe it should warn
> the caller?  Otherwise, for example, string->number does not invert
> number->string:
> 
> guile> (string->number "10.5" 2)
> #f

Yes.  We are in clear violation of R5RS here.  I'll record a bug for
this in the workbook.

-- 
GPG: D5D4E405 - 2F9B BCCC 8527 692A 04E3  331E FAF8 226A D5D4 E405