Command for inspecting the representation of a [Tcl_Obj] for debugging purposes. This command is not officially supported and the format of its output is deliberately chosen to be suitable for human use and not mechanical use. ''Do not base the behavior of any command on the results of this one; it does not conform to Tcl's value semantics!''

    :   '''[tcl::unsupported]::representation''' ''value''

----
**Examples**
[DKF]: Here are some examples of output. Note that the refcounts can be sometimes confusing; be aware that references to literals can be retained in the [history].

======
% tcl::unsupported::representation 1
value is a int with a refcount of 8, object pointer at 0x32eb0, internal representation 0x1:0x0, string representation "1".
% tcl::unsupported::representation 123
value is a pure string with a refcount of 3, object pointer at 0x34230, string representation "123".
% tcl::unsupported::representation abc
value is a pure string with a refcount of 3, object pointer at 0x32f40, string representation "abc".
% set s abcdef
abcdef
% string length $s
6
% tcl::unsupported::representation $s
value is a string with a refcount of 4, object pointer at 0x34338, internal representation 0x93c50:0x0, string representation "abcdef".
% tcl::unsupported::representation [list 1 2 3]
value is a list with a refcount of 1, object pointer at 0x34500, internal representation 0x8a050:0x0, no string representation.
% tcl::unsupported::representation [dict create 1 2 3 4]
value is a dict with a refcount of 1, object pointer at 0x97428, internal representation 0x92210:0x33960, no string representation.
======

[AMG]: I'm curious about why "1" has a large refcount.  Does Tcl maintain a cache of common Tcl_Objs? 

'''[DGP]''' Tcl keeps tables of literal values so they may be shared.

[AMG]: Also, what about this?

======
% # Test with "1"
% set data 1
% tcl::unsupported::representation $data
value is a int with a refcount of 20, object pointer at 00E138D0, internal representation 00000001:00000000, string representation "1".
% llength $data
% tcl::unsupported::representation $data
value is a int with a refcount of 19, object pointer at 00E138D0, internal representation 00000001:00000000, string representation "1".
% # Test with "12345" rigged to have an "int" intrep
% set data [expr {12345}]
% tcl::unsupported::representation $data
value is a int with a refcount of 3, object pointer at 00DB61C8, internal representation 00003039:00000000, string representation "12345".
% llength $data
% tcl::unsupported::representation $data
value is a list with a refcount of 2, object pointer at 00DB61C8, internal representation 00DD8988:00000000, string representation "12345".
% info patchlevel
8.6b1.2
======

[[[llength]]] causes 12345 to shimmer from int to list, but it does ''not'' cause 1 to shimmer from int to list.  Why?

'''[DGP]''' That is probably an effect of shared literals, but honestly I cannot reproduce it.

[APN] Could someone explain the ref counts from the following sequence ? Why does the refcount for literal "a" go up on the second assignment to x? And after unsetting of x, should the refcount not be back to the original 3 ?

======
(tcl) 49 % ::tcl::unsupported::representation a
value is a pure string with a refcount of 3, object pointer at 00F2A030, string representation "a".
(tcl) 49 % set x a
a
(tcl) 50 % ::tcl::unsupported::representation a
value is a pure string with a refcount of 6, object pointer at 00F29A00, string representation "a".
(tcl) 51 % set x a
a
(tcl) 52 % ::tcl::unsupported::representation a
value is a pure string with a refcount of 8, object pointer at 00F29A00, string representation "a".
(tcl) 53 % unset x
(tcl) 54 % ::tcl::unsupported::representation a
value is a pure string with a refcount of 7, object pointer at 00F29A00, string representation "a".
======

<<categories>> Command | Internals