[Alpha-Bugzilla Bug 1454] New - RFE: -indices flag to [listpick]

[Alpha-Bugzilla <alp...@ic...>]

[This e-mail has been automatically generated.]

http://www.maths.mq.edu.au/~steffen/Alpha/bugzilla/show_bug.cgi?id=1454

           Summary: RFE: -indices flag to [listpick]
           Product: Alpha
           Version: 8.0b15
          Platform: Macintosh PowerPC
        OS/Version: All
            Status: NEW
 Status Whiteboard:
          Keywords:
        Resolution:
          Severity: enhancement
          Priority: P2
         Component: Core
        AssignedTo: alp...@li...
        ReportedBy: ko...@ma...

It would be nice if [listpick] had an option -indices which would make the
proc return the index of the chosen list item instead of the item itself.
This is needed in many places where the list presented to the user is some
sort of primitive version of the list that really matters internally.

For example, when electric-completing a file name, the relevant list
internally is of course the list of full path names, but for the user it
would be much more convenient to see only the tail of each item.  In fact,
if you are more than around 32 chars down in a file hierarchy and you try
to complete a file name, the amount of text shown in the listpick doesn't
even show the tail because truncation at listpick-width takes place...

Other examples: when completing a bibitem, what matters internally is
the citekey, but what you would like to present in the listpick is the
list of titles (and possibly authors).  

In general, whenever you are requested to select a record in a database,
internally you want to refer to the record by its uid, but at the
user interface level you will rather present a list of some other key
from the database, since the uid might be pretty useless for the user to
choose from.

Sure, it is not extremely difficult to perform an additional [lsearch] for
the return value of [listpick] and get the index in this way, but it is
sort of backwards and awkward.

Without knowing what I am talking about I guess that it would be easy to 
implement in the core: probably the core proc knows the index before it 
knows the value at that index...?

--
This e-mail has been sent to you by the Alpha-Bugzilla bugsystem
(http://www.maths.mq.edu.au/~steffen/Alpha/bugzilla/)
I am an automated system, please direct your replies or
complaints to my maintainer <st...@ma...>