Done:
https://pharo.fogbugz.com/f/cases/17179/Semantic-analysis-doesn-t-compute-RBArgumentNode-instances

On Thu, Dec 3, 2015 at 2:30 PM, Marcus Denker <marcus.denker@inria.fr> wrote:
yes !
> On 03 Dec 2015, at 14:16, Martin Dias <tinchodias@gmail.com> wrote:
>
> Hi,
>
> In Pharo 4, the RBExplicitVariableParser computed the argument variables as RBArgumentNodes:
>
> (Integer >> #bitAnd:) parseTree arguments
> "--> an OrderedCollection(RBArgumentNode(n))"
>
> In Pharo 50191, such parser was removed (case 16026) because "semantic analysis" does the same job. However, it does not exactly the same job: argument variables are now parsed as generic variable nodes:
>
> (Integer >> #bitAnd:) parseTree doSemanticAnalysis arguments
>�� "an OrderedCollection(RBVariableNode(n))"
>
> Anyway, it works well with other variables, e.g. temporaries:
>
> (Integer >> #bitAnd:) parseTree doSemanticAnalysis temporaries
>�� "an OrderedCollection(RBTemporaryNode(norm))"
>
> Should I create a bug entry?
>
> Cheers,
> Martin