On Thu, 18 Oct 2001, Jens Grabowski wrote:
> Hi together,
>
> some experiments with TTCN-3 have shown that the naming
> rules for groups are not appropriate.
>
> Currently it is not allowed to specify something like this:
>
> group FirstTestcaseGroup {
> :
> group ValidBehviour {
> :
> }
> :
> }
>
> group SecondTestcaseGroup {
> :
> group ValidBehviour {
> :
> }
> :
> }
>
> Because of the identifier 'ValidBehaviour' is not unique.
After a first cursory reading of the TTCN-3 standard, I also
thought that the above was possible with the meaning that
the groupd 'ValidBehavior' is spread over more than one
other group (of course, the uniqueness of names in one
scope would make that impossible, although I don't see
the necessity for that restriction for group names).
This would make possible grouping mechanisms
beside those of scoping, ie. I could import only
FirstTestcaseGroup, or only SecondTestcaseGroup or
only ValidBehavior which would get me _all_ ValidBehavior
groups inside all other groups.
> We would like to change this, i.e., generalize the naming rules
> for groups. If necessary the dot notation should be used to
> select a specific group (e.g., for importing a group). For example,
> in the example above the groups named 'ValidBehaviour' would be
> uniquely identified by:
>
> FirstTestcaseGroup.ValidBehaviour
> and
> SecondTestcaseGroup.ValidBehaviour
For finer structuring, i.e. if I only want the ValidBehavior
of SecondTestcaseGroup, an additional '.' notation would be
nice, but I would not like to have to always use it for
nested groups.
Jacob Wieland, TU-Berlin