[rkward-tracker] [ rkward-Feature Requests-3499299 ] More powerful code completion

[SourceForge.net]

Feature Requests item #3499299, was opened at 2012-03-07 21:53
Message generated for change (Comment added) made by m-eik
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=459010&aid=3499299&group_id=50231

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Command Editor
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: François Maurice (cratyle)
Assigned to: Nobody/Anonymous (nobody)
Summary: More powerful code completion

Initial Comment:
Hi,

The auto-completion feature in RKWard is powerful but we need more. I'm not sure that the following criticism applies to all functions. For example, when I use the function matchit (), I have access to all sort of things and not only arguments for that function.  Also, I do not have access to the list of options for an argument. For example, there is an argument method= with the following options: "exact", "full", "genetic", "nearest", "optimal", "subclass". But I do not have access to those options (and only those) with Auto-completion. Sure, if I type "gen", the "genetic" options is in the list (wiht other irrelevant things). It would also be appreciated that the Auto-completion is a dynamic tool. When I chose an argument, or the option of an argument, arguments and options that are incompatible with those choices should not appear in subsequent choices. For example, when I choose the "subclass" option from the argument method =, it is not possible to specify the ratio = argument in the matchit() function. So Auto-completion should not display the argument ratio= among the possible choices of arguments.

Thanks to consider this request.


----------------------------------------------------------------------

>Comment By: m.eik (m-eik)
Date: 2012-03-08 10:40

Message:
hi,

thanks for your suggestions, most welcome :-)

i think the problem with completion like you'd like to have is that the
part that does the completion needs to know which options it can offer, and
how they interact. the matchit() function of the MatchIt package seems to
have method="nearest" set as its default value, so method="nearest" would
be about all that code completion could offer, because it's the only usable
information which can be extracted from the function's signature
automatically.

you, as a human being, can read and comprehend the manual to understand
which other options are valid, but unfortunately, RKWard isn't that
intelligent yet ;-)

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=459010&aid=3499299&group_id=50231