Re: Wisent grammar issues

Also, interestingly this grammar produces strange automaton:

zonk (30)
    on left: 1


state 0

    QVARID shift, and go to state 1

    zonk go to state 2



state 1

    zonk  ->  QVARID .   (rule 1)

    $default reduce using rule 1 (zonk)



state 2

    $EOI shift, and go to state 3



state 3

    $EOI shift, and go to state 4



state 4

    $default accept

Note that $EOI is mentioned twice in state 2 and state 3. Why is that?


2016-03-12 18:42 GMT+01:00 Gracjan Polak <[email protected]>:

> Hi,
>
> As advised in other email on this list I try to wrap my grammar in a
> couple of ERT tests. I'm stuck trying to get wisent grammar do anything for
> me in context of ERT tests.
>
> I have a lexer that works correctly and I have this grammar:
>
> %start zonk
> %token QVARID
>
> %%
>
> zonk : QVARID
>      ;
>
> And I have this ert-test declaration:
>
> (defmacro ws (start source expected)
>   `(with-temp-buffer
>     (haskell-mode)
>     (insert ,source)
>     (haskell-default-setup)
>     (semantic-mode)
>     (let ((wisent-parse-verbose-flag t)
>           (wisent-lex-istream
>            (funcall semantic-lex-analyzer (point-min) (point-max)))
>           (wisent-result (wisent-parse semantic--parse-table #'wisent-lex
> (lambda (msg) (message "%s" msg)) ',start)))
>       (message "%S" wisent-lex-istream)
>       (message "%S" wisent-result)
>       (should (equal ',expected wisent-result)))))
>
> (ert-deftest zzzz ()
>   (ws zonk "
> fun"
>       (zonk)))
>
> What I get is strange and I do not know how to interpret this:
>
> Running 1 tests (2016-03-12 18:38:25+0100)
> Syntax error, unexpected $EOI@13(""), expecting QVARID
> ((QVARID 2 . 5))
> nil
>
> Test zzzz condition:
>     (ert-test-failed
>      ((should
>        (equal '... wisent-result))
>       :form
>       (equal
>        (zonk)
>        nil)
>       :value nil :explanation
>       (different-types
>        (zonk)
>        nil)))
>    FAILED  1/1  zzzz
>
> I would expect:
> 1. That is parsers 'fun' a QVARID and zonk rule fires properly.
> 2. That it does not try to parse further.
> 3. That it does not report $EOI issues.
> 4. That it returns a result of any kind, not nil.
>
> Is there anything obvious I'm doing wrong?
>
> --
> Gracjan
>
>

------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785111&iu=/4140

_______________________________________________
cedet-semantic mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cedet-semantic