At 2:34 PM -0400 6/6/02, Brian Tillman wrote:

>$ @[.vms]test .exe "" "-v" [.base]num.t

snip

>  Improperly handled condition, image exit forced.
>        Signal arguments              Stack contents
>        Number = 00000003                00000000
>        Name   = 000004B4                00000000
>                 00164DF9                27FC0000
>                 03C00000                7FE4C2A8
>                                         7FE4C26C
>                                         0016515D
>                                         00000033
>                                         00000000
>                                         00004080
>                                         00000000
>        Register dump
>        R0 = 00000020  R1 = 00000020  R2 = 00004308  R3 = 00000000
>        R4 = C5AD759D  R5 = 70B6A82B  R6 = 84DF78F6  R7 = E01C56C3
>        R8 = 00000000  R9 = 00000022  R10= 00004080  R11= 00000000
>        AP = 7FE4C1C0  FP = 7FE4C180  SP = 7FE4C1F4  PC = 00164DF9
>        PSL= 03C00000

Double ick.  If R0 is to be believed, i.e., is really a condition
value at this point, then it might be a clue:

$ exit %x20
%SYSTEM-W-NOPRIV, insufficient privilege or object protection violation

But I really don't see anything in the test that is likely to trigger
that.  It doesn't even load any libraries.  OK, let's try this:

$ set default [.t]
$ mcr [-]miniperl -c [.base]num.t
[.base]num.t syntax OK

That will probably do the same thing, but will also confirm that it
happens while Perl is compiling the script.  Let's also try

$ mcr [-]miniperl -e "$a = 1e34; print $a;"
1e+34

That's in the test, and while we *should* be able to handle 1e38 with
D_FLOAT, some gremlin might have stolen a few orders of magnitude.

Brian, your ongoing patience and help are much appreciated.  Isn't
there *anyone* else out there who wants to run 5.8.0 on VAX?
-- 
________________________________________
Craig A. Berry
mailto:[EMAIL PROTECTED]

"... getting out of a sonnet is much more
 difficult than getting in."
                 Brad Leithauser

Reply via email to