Re: [Gcl-devel] Re: ANSI compliance in gcl

[Paul F. Dietz]

Sam Steingold wrote:

> so how many tests does GCL fail at the moment?

In gcl:

112 out of 6112 total tests failed: SYMBOL-=, SYMBOL-ARITHMETIC-ERROR, SYMBOL-CELL-ERROR, 
SYMBOL-CONDITION,
   SYMBOL-CONTROL-ERROR, SYMBOL-COUNT, SYMBOL-DESCRIBE,
   SYMBOL-DIVISION-BY-ZERO, SYMBOL-END-OF-FILE, SYMBOL-ERROR,
   SYMBOL-FILE-ERROR, SYMBOL-FIRST, SYMBOL-FLOATING-POINT-OVERFLOW,
   SYMBOL-FLOATING-POINT-UNDERFLOW, SYMBOL-IN-PACKAGE, SYMBOL-LOGXOR,
   SYMBOL-PACKAGE-ERROR, SYMBOL-PROGRAM-ERROR, SYMBOL-RETURN,
   SYMBOL-SERIOUS-CONDITION, SYMBOL-SIMPLE-CONDITION,
   SYMBOL-SIMPLE-ERROR, SYMBOL-SIMPLE-TYPE-ERROR,
   SYMBOL-SIMPLE-WARNING, SYMBOL-STORAGE-CONDITION,
   SYMBOL-STREAM-ERROR, SYMBOL-TYPE-ERROR, SYMBOL-UNBOUND-VARIABLE,
   SYMBOL-UNDEFINED-FUNCTION, SYMBOL-WARNING, SPECIAL-OPERATOR-P.1,
   SPECIAL-OPERATOR-P.2, SPECIAL-OPERATOR-P.3, CL-FUNCTION-SYMBOLS.1,
   CL-MACRO-SYMBOLS.1, CL-VARIABLE-SYMBOLS.1,
   CL-TYPES-THAT-ARE-CLASSES-SYMBOLS.1, EVAL-AND-COMPILE-MACROS,
   COMPILE.1, COMPILE.2, COMPILE.5, COMPILE.6, COMPILE.7, COMPILE.8,
   DCF-FUNS, DCF-MACROS, SETF-VALUES.5, PSETQ.7, PSETF.7, THROW-ERROR,
   FBOUNDP.6, FBOUNDP.7, FBOUNDP.8, FBOUNDP.9, FBOUNDP.10, FLET.17,
   FLET.27, FLET.30, FMAKUNBOUND.4, FMAKUNBOUND.5, FMAKUNBOUND.6,
   FMAKUNBOUND.7, FUNCALL.10, FUNCTION.2, FUNCTION.6, FUNCTION.7,
   FUNCTION.10, FUNCTIONP.2, FUNCTIONP.6, FUNCTIONP.7, FUNCTIONP.10,
   LABELS.17, MACROLET.5, MACROLET.13, MULTIPLE-VALUE-SETQ.5,
   MULTIPLE-VALUE-SETQ.8, LOOP.6.19, LOOP.6.20, LOOP.6.37, LOOP.6.38,
   LOOP.7.15, LOOP.7.16, LOOP.7.17, LOOP.7.18, LOOP.7.19, LOOP.7.20,
   LOOP.7.21, LOOP.7.22, LOOP.7.23, LOOP.8.19, LOOP.8.20, LOOP.8.21,
   LOOP.8.22, LOOP.8.ERROR.2, LOOP.9.10, LOOP.9.11, LOOP.9.27,
   LOOP.9.28, LOOP.9.37, LOOP.9.38, LOOP.10.9, LOOP.10.10, LOOP.10.37,
   LOOP.10.38, LOOP.10.57, LOOP.10.58, LOOP.10.85, LOOP.10.86,
   LOOP.11.9, LOOP.14.29, NCONC-7, IN-PACKAGE-2.

> > It is very useful to run the tests through other implementations to
> > check for this -- I use Allegro CL and CMUCL; others run the tests
> > through CLISP and SBCL (and I think Lispworks).
>
>
> so how many tests do CMUCL, SBCL, ACL and Lispworks fail?

I do not have an up to date CMUCL; in a release called 18d+ from
September there are 243 test failures (but most of those have been
fixed recently.)

Using Allegro's trial ACL on x86 Linux I currently get 677 failures.

I don't have Lispworks or SBCL installed but welcome feedback from those
who do.

        Paul