Extension of leaf nodes not working

[lausdahl]

The test case SimpleExtendTest does not generate the expected output:

base

exp {-> package='org.overture.ast.expressions'}
    =   {aa} [type]:stm.st1
    ;

extention

exp {-> package='org.overture.interpreter.ast.expressions'}
    =   {aa}  [ee]:stm.st2
    ;

We would expect the extended tree to include an extended AAaExp (extended) which extends AAaExp (base), having the following fields:

type
ee

and the analysis to include analysis for this new class as well where the depth first visitor takes inherited fields into account.

The node is skipped in the extended generation in ExtensionGenerator2.includeClassesFromExtension where willGenerateExtensionFor excluded it.

hack to include it:

private static void includeClassesFromExtension(Environment result,
            Environment ext, Environment base)
    {
        for (IClassDefinition cDef : ext.getClasses())
        {
            if (!willGenerateExtensionFor(cDef, base))
            {
                if (!isUtilityOrTemplateClass(cDef, ext))
                {
                    cDef.setIsExtTree(true);
                    result.getClasses().add(cDef);
                }
            }

            if(cDef.getName().toString().startsWith("AAaExp"))
            {
                cDef.setIsExtTree(true);
                result.getClasses().add(cDef);
            }
        }
    }

[ldcouto]

Looking at it, seems like both classes have the same tag "". Might that be it? Or maybe we need to consider the package in the calculation of whether to generate?

[lausdahl]

I think the tag is the name it has in the parse tree like exp. I mean the name before it gets an A, S or P in front but I need to look at it first.

Sent from my iPhone

On 02/03/2015, at 16.11, Luis Diogo Couto notifications@github.com wrote:

Looking at it, seems like both classes have the same tag "". Might that be it? Or maybe we need to consider the package in the calculation of whether to generate?

—
Reply to this email directly or view it on GitHub.

[ldcouto]

When this is done, we need to merge it in master. We branched off a few commits back when we started it. Oops...

[ldcouto]

Ok, so #11 was a duplicate of this. As I said there...

We need to write a few tests for it and make sure Symphony etc. still build

[ldcouto]

We have re-merged the branch into master.

But there's some sort of problem with the Symphony AST now.

It generates a "PExp" node for CML, which it should not.

Need to look more into this.