[Tcl-bugs] [ tcl-Bugs-1022151 ] 'array unset' inconsistency ?

[SourceForge.net <no...@so...>]

Bugs item #1022151, was opened at 2004-09-04 00:20
Message generated for change (Comment added) made by dkf
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110894&aid=1022151&group_id=10894

Category: 07. Variables
Group: current: 8.4.7
Status: Open
Resolution: None
Priority: 5
Submitted By: Andreas Kupries (andreas_kupries)
Assigned to: miguel sofer (msofer)
Summary: 'array unset' inconsistency ?

Initial Comment:
IMHO the behaviour of the command
    'array unset ?pattern?'
is inconsistent, and the problem is exacerbated by an
ambiguous description in its manpage.

Take an array variable, add some elements, then execute

(1)     unset a
(2)     array unset a
(3)	array unset a *

[This has to be done independently, not one after the
other]

Tracing and checking the results of the commands shows
that the scenarios (1) and (2) are equivalent. Whereas
the regular pattern for this type of command is to make
(2) and (3) equivalent.

However when reading the documentation for 'array unset'
it seems to imply the (2)/(3) equivalence:

        Unsets all of the elements in the array that match
        pattern	(using the matching rules of string match).
        [...] If pattern is omitted and arrayName is an
array
        variable, then the command unsets the entire array.

Because of the first sentence it can be assumed while
reading the last sentence that 'unsetting the entire array'
means the removal of all elements in the array, and not
the removal of the variable.

Only then reading them disconnected it becomes clear
that the last sentence means that the variable is unset,
not its elements.

Either documentation or behaviour should be fixed.
Preferably both.


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

>Comment By: Donal K. Fellows (dkf)
Date: 2004-09-06 12:03

Message:
Logged In: YES 
user_id=79902

IMO the behaviour is inconsistent with the way the rest of
the core operates (in terms of filtering) and so is *wrong*.
 I think we should fix the code and make sure that the
documentation says what it should as well.

Note that this may be a POTENTIAL INCOMPATABILITY, but it is
at least following the Principle of Least Surprise...

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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110894&aid=1022151&group_id=10894