Diff for /gforth/test/ttester.fs between versions 1.13 and 1.14

version 1.13, 2008/11/08 18:34:18 version 1.14, 2009/04/08 19:59:17
Line 1 Line 1
 \ for the original tester  \ This file contains the code for ttester, a utility for testing Forth words,
   \ as developed by several authors (see below), together with some explanations
   \ of its use.
   
   \ ttester is based on the original tester suite by Hayes:
 \ From: John Hayes S1I  \ From: John Hayes S1I
 \ Subject: tester.fr  \ Subject: tester.fr
 \ Date: Mon, 27 Nov 95 13:10:09 PST    \ Date: Mon, 27 Nov 95 13:10:09 PST  
 \ (C) 1995 JOHNS HOPKINS UNIVERSITY / APPLIED PHYSICS LABORATORY  \ (C) 1995 JOHNS HOPKINS UNIVERSITY / APPLIED PHYSICS LABORATORY
 \ MAY BE DISTRIBUTED FREELY AS LONG AS THIS COPYRIGHT NOTICE REMAINS.  \ MAY BE DISTRIBUTED FREELY AS LONG AS THIS COPYRIGHT NOTICE REMAINS.
 \ VERSION 1.1  \ VERSION 1.1
   \ All the subsequent changes have been placed in the public domain.
 \ for the FNEARLY= stuff:  \ The primary changes from the original are the replacement of "{" by "T{"
   \ and "}" by "}T" (to avoid conflicts with the uses of { for locals and }
   \ for FSL arrays), modifications so that the stack is allowed to be non-empty
   \ before T{, and extensions for the handling of floating point tests.
   \ Code for testing equality of floating point values comes
 \ from ftester.fs written by David N. Williams, based on the idea of  \ from ftester.fs written by David N. Williams, based on the idea of
 \ approximate equality in Dirk Zoller's float.4th  \ approximate equality in Dirk Zoller's float.4th.
 \ public domain  \ Further revisions were provided by Anton Ertl, including the ability
   \ to handle either integrated or separate floating point stacks.
 \ for the rest:  \ Revision history and possibly newer versions can be found at
 \ revised by Anton Ertl  
 \ public domain  
   
 \ The original has the following shortcomings:  
   
 \ - It does not work as expected if the stack is non-empty before the {.  
   
 \ - It does not check FP results if the system has a separate FP stack.  
   
 \ - There is a conflict with the use of } for FSL arrays and { for locals.  
   
 \ I have revised it to address these shortcomings.  You can find the  
 \ result at  
   
 \ http://www.forth200x.org/tests/tester.fs  
 \ http://www.forth200x.org/tests/ttester.fs  
   
 \ for history and possibly newer versions you can also look at  
 \ http://www.complang.tuwien.ac.at/cvsweb/cgi-bin/cvsweb/gforth/test/ttester.fs  \ http://www.complang.tuwien.ac.at/cvsweb/cgi-bin/cvsweb/gforth/test/ttester.fs
   \ Explanatory material and minor reformatting (no code changes) by
   \ C. G. Montgomery March 2009, with helpful comments from David Williams
   \ and Krishna Myneni.
   
   \ Usage:
   
   \ The basic usage takes the form  T{ <code> -> <expected stack> }T .
   \ This executes  <code>  and compares the resulting stack contents with
   \ the  <expected stack>  values, and reports any discrepancy between the
   \ two sets of values.
   \ For example:
   \ T{ 1 2 3 swap -> 1 3 2 }T  ok
   \ T{ 1 2 3 swap -> 1 2 2 }T INCORRECT RESULT: T{ 1 2 3 swap -> 1 2 2 }T ok
   \ T{ 1 2 3 swap -> 1 2 }T WRONG NUMBER OF RESULTS: T{ 1 2 3 swap -> 1 2 }T ok
   
   \ Floating point testing can involve further complications.  The code
   \ attempts to determine whether floating-point support is present, and
   \ if so, whether there is a separate floating-point stack, and behave
   \ accordingly.  The CONSTANTs HAS-FLOATING and HAS-FLOATING-STACK
   \ contain the results of its efforts, so the behavior of the code can
   \ be modified by the user if necessary.
   
   \ Then there are the perennial issues of floating point value
   \ comparisons.  Exact equality is specified by SET-EXACT (the
   \ default).  If approximate equality tests are desired, execute
   \ SET-NEAR .  Then the FVARIABLEs REL-NEAR (default 1E-12) and
   \ ABS-NEAR (default 0E) contain the values to be used in comparisons
   \ by the (internal) word FNEARLY= .
   
   \ When there is not a separate floating point stack and you want to
   \ use approximate equality for FP values, it is necessary to identify
   \ which stack items are floating point quantities.  This can be done
   \ by replacing the closing }T with a version that specifies this, such
   \ as RRXR}T which identifies the stack picture ( r r x r ).  The code
   \ provides such words for all combinations of R and X with up to four
   \ stack items.  They can be used with either an integrated or separate
   \ floating point stacks. Adding more if you need them is
   \ straightforward; see the examples in the source.  Here is an example
   \ which also illustrates controlling the precision of comparisons:
   
   \   SET-NEAR
   \   1E-6 REL-NEAR F!
   \   T{ S" 3.14159E" >FLOAT -> -1E FACOS TRUE RX}T
   
   \ The word ERROR is now vectored, so that its action can be changed by
   \ the user (for example, to add a counter for the number of errors).
   \ The default action ERROR1 can be used as a factor in the display of
   \ error reports.
   
 \ tester.fs is intended to be a drop-in replacement of the original.  \ Loading ttester.fs does not change BASE.  Remember that floating point input
   \ is ambiguous if the base is not decimal.
 \ ttester.fs is a version that uses T{ and }T instead of { and } and  
 \ keeps the BASE as it was before loading ttester.fs  
   
 \ In spirit of the original, I have strived to avoid any potential  \ The file defines some 70 words in all, but in most cases only the
 \ non-portabilities and stayed as much within the CORE words as  \ ones mentioned above will be needed for successful testing.
 \ possible; e.g., FLOATING words are used only if the FLOATING wordset  
 \ is present  
   
 \ There are a few things to be noted:  
   
 \ - Loading ttester.fs does not change BASE.  Loading tester.fs  
 \ changes BASE to HEX (like the original tester).  Floating-point  
 \ input is ambiguous when the base is not decimal, so you have to set  
 \ it to decimal yourself when you want to deal with decimal numbers.  
   
 \ - For FP it is often useful to use approximate equality for checking  
 \ the results.  You can turn on approximate matching with SET-NEAR  
 \ (and turn it off (default) with SET-EXACT, and you can tune it by  
 \ setting the variables REL-NEAR and ABS-NEAR.  If you want your tests  
 \ to work with a shared stack, you have to specify the types of the  
 \ elements on the stack by using one of the closing words that specify  
 \ types, e.g. RRRX}T for checking the stack picture ( r r r x ).  
 \ There are such words for all combination of R and X with up to 4  
 \ stack items, and defining more if you need them is straightforward  
 \ (see source).  If your tests are only intended for a separate-stack  
 \ system or if you need only exact matching, you can use the plain }T  
 \ instead.  
   
 BASE @  BASE @
 HEX  HEX
   
 \ SET THE FOLLOWING FLAG TO TRUE FOR MORE VERBOSE OUTPUT; THIS MAY  VARIABLE ACTUAL-DEPTH                   \ stack record
 \ ALLOW YOU TO TELL WHICH TEST CAUSED YOUR SYSTEM TO HANG.  
 VARIABLE VERBOSE  
    FALSE VERBOSE !  
   
 VARIABLE ACTUAL-DEPTH                   \ STACK RECORD  
 CREATE ACTUAL-RESULTS 20 CELLS ALLOT  CREATE ACTUAL-RESULTS 20 CELLS ALLOT
 VARIABLE START-DEPTH  VARIABLE START-DEPTH
 VARIABLE XCURSOR \ FOR ...}T  VARIABLE XCURSOR      \ for ...}T
 VARIABLE ERROR-XT  VARIABLE ERROR-XT
   
 : ERROR ERROR-XT @ EXECUTE ;  : ERROR ERROR-XT @ EXECUTE ;   \ for vectoring of error reporting
   
 : "FLOATING" S" FLOATING" ; \ ONLY COMPILED S" IN CORE  : "FLOATING" S" FLOATING" ;    \ only compiled S" in CORE
 : "FLOATING-STACK" S" FLOATING-STACK" ;  : "FLOATING-STACK" S" FLOATING-STACK" ;
 "FLOATING" ENVIRONMENT? [IF]  "FLOATING" ENVIRONMENT? [IF]
     [IF]      [IF]
Line 95  VARIABLE ERROR-XT Line 104  VARIABLE ERROR-XT
     [ELSE]      [ELSE]
         FALSE          FALSE
     [THEN]      [THEN]
 [ELSE] \ WE DON'T KNOW WHETHER THE FP STACK IS SEPARATE  [ELSE]            \ We don't know whether the FP stack is separate.
     HAS-FLOATING \ IF WE HAVE FLOATING, WE ASSUME IT IS      HAS-FLOATING  \ If we have FLOATING, we assume it is.
 [THEN] CONSTANT HAS-FLOATING-STACK  [THEN] CONSTANT HAS-FLOATING-STACK
   
 HAS-FLOATING [IF]  HAS-FLOATING [IF]
     \ SET THE FOLLOWING TO THE RELATIVE AND ABSOLUTE TOLERANCES YOU      \ Set the following to the relative and absolute tolerances you
     \ WANT FOR APPROXIMATE FLOAT EQUALITY, TO BE USED WITH F~ IN      \ want for approximate float equality, to be used with F~ in
     \ FNEARLY=.  KEEP THE SIGNS, BECAUSE F~ NEEDS THEM.      \ FNEARLY=.  Keep the signs, because F~ needs them.
     FVARIABLE REL-NEAR DECIMAL 1E-12 HEX REL-NEAR F!      FVARIABLE REL-NEAR DECIMAL 1E-12 HEX REL-NEAR F!
     FVARIABLE ABS-NEAR    DECIMAL 0E HEX ABS-NEAR F!      FVARIABLE ABS-NEAR    DECIMAL 0E HEX ABS-NEAR F!
   
     \ WHEN EXACT? IS TRUE, }F USES FEXACTLY=, OTHERWISE FNEARLY=.      \ When EXACT? is TRUE, }F uses FEXACTLY=, otherwise FNEARLY=.
           
     TRUE VALUE EXACT?      TRUE VALUE EXACT?
     : SET-EXACT  ( -- )   TRUE TO EXACT? ;      : SET-EXACT  ( -- )   TRUE TO EXACT? ;
Line 115  HAS-FLOATING [IF] Line 124  HAS-FLOATING [IF]
     DECIMAL      DECIMAL
     : FEXACTLY=  ( F: X Y -- S: FLAG )      : FEXACTLY=  ( F: X Y -- S: FLAG )
         (          (
         LEAVE TRUE IF THE TWO FLOATS ARE IDENTICAL.          Leave TRUE if the two floats are identical.
         )          )
         0E F~ ;          0E F~ ;
     HEX      HEX
           
     : FABS=  ( F: X Y -- S: FLAG )      : FABS=  ( F: X Y -- S: FLAG )
         (          (
         LEAVE TRUE IF THE TWO FLOATS ARE EQUAL WITHIN THE TOLERANCE          Leave TRUE if the two floats are equal within the tolerance
         STORED IN ABS-NEAR.          stored in ABS-NEAR.
         )          )
         ABS-NEAR F@ F~ ;          ABS-NEAR F@ F~ ;
           
     : FREL=  ( F: X Y -- S: FLAG )      : FREL=  ( F: X Y -- S: FLAG )
         (          (
         LEAVE TRUE IF THE TWO FLOATS ARE RELATIVELY EQUAL BASED ON THE          Leave TRUE if the two floats are relatively equal based on the
         TOLERANCE STORED IN ABS-NEAR.          tolerance stored in ABS-NEAR.
         )          )
         REL-NEAR F@ FNEGATE F~ ;          REL-NEAR F@ FNEGATE F~ ;
   
Line 139  HAS-FLOATING [IF] Line 148  HAS-FLOATING [IF]
           
     : FNEARLY=  ( F: X Y -- S: FLAG )      : FNEARLY=  ( F: X Y -- S: FLAG )
         (          (
         LEAVE TRUE IF THE TWO FLOATS ARE NEARLY EQUAL.  THIS IS A          Leave TRUE if the two floats are nearly equal.  This is a 
         REFINEMENT OF DIRK ZOLLER'S FEQ TO ALSO ALLOW X = Y, INCLUDING          refinement of Dirk Zoller's FEQ to also allow X = Y, including
         BOTH ZERO, OR TO ALLOW APPROXIMATE EQUALITY WHEN X AND Y ARE TOO          both zero, or to allow approximately equality when X and Y are too
         SMALL TO SATISFY THE RELATIVE APPROXIMATION MODE IN THE F~          small to satisfy the relative approximation mode in the F~ 
         SPECIFICATION.          specification.
         )          )
         F2DUP FEXACTLY= IF F2DROP TRUE EXIT THEN          F2DUP FEXACTLY= IF F2DROP TRUE EXIT THEN
         F2DUP FREL=     IF F2DROP TRUE EXIT THEN          F2DUP FREL=     IF F2DROP TRUE EXIT THEN
Line 234  HAS-FLOATING-STACK [IF] Line 243  HAS-FLOATING-STACK [IF]
     [THEN]      [THEN]
 [THEN]      [THEN]    
   
 : EMPTY-STACK   \ ( ... -- ) EMPTY STACK: HANDLES UNDERFLOWED STACK TOO.  : EMPTY-STACK   \ ( ... -- ) empty stack; handles underflowed stack too.
     DEPTH START-DEPTH @ < IF      DEPTH START-DEPTH @ < IF
         DEPTH START-DEPTH @ SWAP DO 0 LOOP          DEPTH START-DEPTH @ SWAP DO 0 LOOP
     THEN      THEN
Line 243  HAS-FLOATING-STACK [IF] Line 252  HAS-FLOATING-STACK [IF]
     THEN      THEN
     EMPTY-FSTACK ;      EMPTY-FSTACK ;
   
 : ERROR1        \ ( C-ADDR U -- ) DISPLAY AN ERROR MESSAGE FOLLOWED BY  : ERROR1        \ ( C-ADDR U -- ) display an error message 
                 \ THE LINE THAT HAD THE ERROR.                  \ followed by the line that had the error.
    TYPE SOURCE TYPE CR                  \ DISPLAY LINE CORRESPONDING TO ERROR     TYPE SOURCE TYPE CR                  \ display line corresponding to error
    EMPTY-STACK                          \ THROW AWAY EVERY THING ELSE     EMPTY-STACK                          \ throw away everything else
 ;  ;
   
 ' ERROR1 ERROR-XT !  ' ERROR1 ERROR-XT !
   
 : T{            \ ( -- ) SYNTACTIC SUGAR.  : T{            \ ( -- ) syntactic sugar.
    DEPTH START-DEPTH ! 0 XCURSOR ! F{ ;     DEPTH START-DEPTH ! 0 XCURSOR ! F{ ;
   
 : ->            \ ( ... -- ) RECORD DEPTH AND CONTENT OF STACK.  : ->            \ ( ... -- ) record depth and contents of stack.
    DEPTH DUP ACTUAL-DEPTH !             \ RECORD DEPTH     DEPTH DUP ACTUAL-DEPTH !             \ record depth
    START-DEPTH @ > IF           \ IF THERE IS SOMETHING ON STACK     START-DEPTH @ > IF           \ if there is something on the stack
        DEPTH START-DEPTH @ - 0 DO ACTUAL-RESULTS I CELLS + ! LOOP \ SAVE THEM         DEPTH START-DEPTH @ - 0 DO ACTUAL-RESULTS I CELLS + ! LOOP \ save them
    THEN     THEN
    F-> ;     F-> ;
   
 : }T            \ ( ... -- ) COMPARE STACK (EXPECTED) CONTENTS WITH SAVED  : }T            \ ( ... -- ) COMPARE STACK (EXPECTED) CONTENTS WITH SAVED
                 \ (ACTUAL) CONTENTS.                  \ (ACTUAL) CONTENTS.
    DEPTH ACTUAL-DEPTH @ = IF            \ IF DEPTHS MATCH     DEPTH ACTUAL-DEPTH @ = IF            \ if depths match
       DEPTH START-DEPTH @ > IF          \ IF THERE IS SOMETHING ON THE STACK        DEPTH START-DEPTH @ > IF          \ if there is something on the stack
          DEPTH START-DEPTH @ - 0 DO     \ FOR EACH STACK ITEM           DEPTH START-DEPTH @ - 0 DO     \ for each stack item
             ACTUAL-RESULTS I CELLS + @  \ COMPARE ACTUAL WITH EXPECTED              ACTUAL-RESULTS I CELLS + @  \ compare actual with expected
             <> IF S" INCORRECT RESULT: " ERROR LEAVE THEN              <> IF S" INCORRECT RESULT: " ERROR LEAVE THEN
          LOOP           LOOP
       THEN        THEN
    ELSE                                 \ DEPTH MISMATCH     ELSE                                 \ depth mismatch
       S" WRONG NUMBER OF RESULTS: " ERROR        S" WRONG NUMBER OF RESULTS: " ERROR
    THEN     THEN
    F} ;     F} ;
Line 325  HAS-FLOATING [IF] Line 334  HAS-FLOATING [IF]
 : RRRR}T FTESTER FTESTER FTESTER FTESTER ...}T ;  : RRRR}T FTESTER FTESTER FTESTER FTESTER ...}T ;
 [THEN]  [THEN]
   
   \ Set the following flag to TRUE for more verbose output; this may
   \ allow you to tell which test caused your system to hang.
   VARIABLE VERBOSE
      FALSE VERBOSE !
   
 : TESTING       \ ( -- ) TALKING COMMENT.  : TESTING       \ ( -- ) TALKING COMMENT.
    SOURCE VERBOSE @     SOURCE VERBOSE @
    IF DUP >R TYPE CR R> >IN !     IF DUP >R TYPE CR R> >IN !
Line 332  HAS-FLOATING [IF] Line 346  HAS-FLOATING [IF]
    THEN ;     THEN ;
   
 BASE !  BASE !
   \ end of ttester.fs

Removed from v.1.13  
changed lines
  Added in v.1.14


FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>