namespace foo {
code puts "Hello World!"
}
produces the scoped value:
@scope ::foo {puts {Hello World!}}
Note that the code command captures the current namespace
context. If the -namespace flag is specified, then the
current context is ignored, and the name string is used
as the namespace context.
Extensions like Tk execute ordinary code fragments in the global namespace. A scoped value captures a code fragment together with its namespace context in a way that allows it to be executed properly later. It is needed, for example, to wrap up code fragments when a Tk widget is used within a namespace:
namespace foo {
private proc report {mesg} {
puts "click: $mesg"
}
button .b1 -text "Push Me"
-command [code report "Hello World!"]
pack .b1
}
The code fragment associated with button .b1 only makes
sense in the context of namespace "foo". Furthermore, the
"report" procedure is private, and can only be accessed within
that namespace. The code command wraps up the code
fragment in a way that allows it to be executed properly
when the button is pressed.
Also, note that the code command preserves the integrity of arguments on the command line. This makes it a natural replacement for the list command, which is often used to format Tcl code fragments. In other words, instead of using the list command like this:
after 1000 [list puts "Hello $name!"]use the code command like this:
after 1000 [code puts "Hello $name!"]This not only formats the command correctly, but also captures its namespace context.
Scoped commands can be invoked like ordinary code fragments, with or without the eval command. For example, the following statements work properly:
set cmd {@scope ::foo .b1}
$cmd configure -background red
set opts {-bg blue -fg white}
eval $cmd configure $opts
Note that scoped commands by-pass the usual protection mechanisms;
the command:
@scope ::foo {report {Hello World!}}
can be used to access the "foo::report" proc from any namespace
context, even though it is private.