GoogleCodeExporter
commented
9 years ago This is a good idea. The main question is do we want this to be included in
2.8.2 or not. That release is already long overdue so we don't want to wait too
much for new features. We won't be able to get the release out this week,
though, so there is some time still.
Some comments about the current code and this feature in general.
1) Tests and docs needed.
2) It would probably be a good idea to wait until we get cleanup related to
issue 1500 ready before the final implementation. There are going to be changes
to same code that may make merging complicated and at least mess the history.
3) With Jython you should always import Java classes/interfaces like `from
foo.bar import SomeStuff` instead of just using `import foo.bar`. The former is
the style we use in Robot code in general and, more importantly, the latter
doesn't always work when the code is packaged to JAR. I don't remember the
details about this latter issue but better safe than sorry.
4) Static Java keywords cannot support named argument syntax even though they
could support kwargs. Thus your change not to create ArgumentResolver with
`resolve_named=False` is not a good idea at least alone. With that change you
probably could call Java keywords like `arg1=foo` or `arg2=bar`
(JavaArgumentParser uses generic names `arg1`, `arg2`, etc.) but that would
likely fail later. I'm not entirely sure what would be the best way to avoid
this. Perhaps we should discuss that in IRC.
5) If kwargs can be supported by having Map as the last argument, it would be
logical if varargs would work with List. If I remember correctly, currently
only arrays are supported. Jython should automatically coerce both to array and
List, so implementing this should only require recognizing List as varargs like
arrays. This should probably be taken care by a separate issue.
3)Original comment by pekka.klarck on 20 Nov 2013 at 8:59
- Changed state: Pending
Original issue reported on code.google.com by
user.stefanzon 20 Nov 2013 at 3:34