Re: [Oorexx-devel] (Adverse ?) Side effects of commit 4296

[Rick M. <obj...@gm...>]

Mark,

My previous fix "over shot" a little, going from on trapping to too
much.  This is fixed now.

Rick

On Mon, Mar 23, 2009 at 4:00 PM, Mark Miesfeld <mie...@gm...> wrote:
> Commit 4296 causes a large number tests in the test suite to fail.  It
> seems like two basic categories of failures.  As usual, I'm having
> trouble sorting out what should be what. <grin>
>
> The first category, and most of the failures, are from rxmath.  Before
> 4296, these two lines:
>
>  answ = rxCalcSin(45)
>  say 'rxCalcSin(45) is:' answ
>
> produce:
>
> rxCalcSin(45) is: 0.707106781
>
> A 4296 or later build produces:
>
>    16 *-* answ = rxCalcSin(45)
> Error 44 running E:\work.ooRexx\qRxmath.rex line 16:  Function or
> message did not return data
> Error 44.1:  No data returned from function "RXCALCSIN"
>
> I see why, but am not sure what the proper thing to do is.  After the
> rxmath calculates the number, it goes to format the return using:
>
> context->DoubleToObjectWithPrecision(x, precision);
>
> which goes into RexxNumberString::newInstanceFromDouble()
>
> and has:
>
>  sprintf(doubleStr, "%.*g", precision + 2, number);
>
> then when it goes into RexxNumberString::prepareNumber() it hits this:
>
>  RexxNumberString *newObj = this->clone();
>    if (newObj->length > NumberDigits)
>    {  /* is the length larger than digits()*/
>       /* raise a numeric condition, may    */
>       /*  not return from this.            */
>        reportCondition(OREF_LOSTDIGITS, (RexxString *)newObj);
>
> Since the number was formatted in sprintf() as 2 digits longer than
> the precision, it never returns from reportCondition() and the
> external function rxCalcSin() ends up not returning anything.
>
> Rick, here's a quick cut and paste program to demonstrate the error:
>
> /* Rexx */
>  parse version v
>  say
>  say "Rexx 'version' information:"
>  say " " v
>  say
>
>  ret = rxFuncAdd("mathLoadFuncs", "rxmath", "mathLoadFuncs")
>  say 'func add ret:' ret
>  ret = mathLoadFuncs()
>  say 'math load ret:' ret
>
>  sq = rxCalcSqrt(16)
>  say 'square root of 16 is:' sq
>
>  answ = rxCalcSin(45)
>  say 'rxCalcSin(45) is:' answ
>
> return
>
> So, once again, I think Rick needs to figure out how to handle this.
>
> I'll put the second category of failures in a second post.
>
> --
> Mark Miesfeld
>
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> Oorexx-devel mailing list
> Oor...@li...
> https://lists.sourceforge.net/lists/listinfo/oorexx-devel
>