Re: [gobo-eiffel-develop] Precondition violated in XPath test

[Colin A. <col...@go...>]

Then the correct thing to do is to remove the call (after the
precursor call) to initialize_special_properties.

What I don't understand is why I didn't get this error when running
the tests with ge. I am afraid as to what the state of my local
repository is compared with yours. Can this wait until tomorrow?

On 01/11/2007, Eric Bezault <er...@go...> wrote:
> Colin Adams wrote:
> > It's up to you.
> > 6155 has fewer bugs.
>
> Hmm, this is not making things easy.
>
> Here is what I found:
>
> In XM_XPATH_COMPUTED_EXPRESSION:
>
> ~~~~~~~~~~~~~~
>         compute_special_properties is
>                         -- Compute special properties.
>                 require
>                         not_yet_computed: not are_special_properties_computed
>                         all_sub_expressions_computed: sub_expressions_have_special_properties
>                 do
>                         initialize_special_properties
>                 ensure
>                         computed: are_special_properties_computed and then special_properties
> /= Void
>                 end
>
> ~~~~~~~~~~~~~~
>
> And in XM_XPATH_MAPPED_PATH_EXPRESSION (6155):
>
> ~~~~~~~~~~~~~~
>         compute_special_properties is
>                         -- Compute special properties.
>                 local
>                         l_expression: XM_XPATH_COMPUTED_EXPRESSION
>                 do
>                         Precursor
>                         initialize_special_properties
>
>                         if not are_cardinalities_computed then compute_cardinality end
>                         if not start.are_special_properties_computed then
>                                 l_expression := start.as_computed_expression
>                                 l_expression.compute_special_properties
>                         end
>                         if not step.are_special_properties_computed then
>                                 l_expression := step.as_computed_expression
>                                 l_expression.compute_special_properties
>                         end
>
>                         if start.non_creating and then step.non_creating then
>                                 set_non_creating
>                         end
>                 end
> ~~~~~~~~~~~~~~
>
> So `initialize_special_properties' is called twice (in the Precursor
> and on the next line), which explains the precondition violation:
> -------------------------------------------------------------------------------
> Class / Object      Routine                Nature of exception
>    Effect
> -------------------------------------------------------------------------------
> XM_XPATH_MAPPED_PATH_EXPRESSION
>                       initialize_special_properties @5
>
> special_properties_not_computed:
> <00000000046E47B8>  (From XM_XPATH_STATIC_PROPERTY)
>                                              Precondition violated.
>     Fail
> -------------------------------------------------------------------------------
> XM_XPATH_MAPPED_PATH_EXPRESSION
>                       compute_special_properties @4
> <00000000046E47B8>                         Routine failure.
>    Fail
> -------------------------------------------------------------------------------
> XM_XPATH_MAPPED_PATH_EXPRESSION
>                       compute_static_properties @29
> <00000000046E47B8>  (From XM_XPATH_COMPUTED_EXPRESSION)
>                                              Routine failure.
>     Fail
> -------------------------------------------------------------------------------
> XM_XPATH_MAPPED_PATH_EXPRESSION
>                       make @7
> <00000000046E47B8>                         Routine failure.
>    Fail
> -------------------------------------------------------------------------------
>
> Therefore, I suggest removing the call to
> `initialize_special_properties' after the call
> to Precursor.
>
> --
> Eric Bezault
> mailto:er...@go...
> http://www.gobosoft.com
>