A function to substitute parts of a string matching a regular expression with the result of a script called using the match as argument. Code and examples will be much more clear than my english:

(new version more comfortable and faster)

    # regmap re string var script
    # substitute all the occurrences of 're' in 'string' with
    # the result of the evaluation of 'script'. Script is called
    # for every match with 'var' variable set to the matching string.
    #
    # Example: regmap {[0-9]+} "1 2 hello 3 4 world" x {expr {$x+1}}
    # Result:  "2 3 hello 4 5 world"
    proc regmap {re string var script} {
        set submatches [lindex [regexp -about $re] 0]
        lappend varlist idx
        while {[incr submatches -1] >= 0} {
            lappend varlist _
        }
        set res $string
        set indices [regexp -all -inline -indices $re $string]
        set delta 0
        foreach $varlist $indices {
            foreach {start end} $idx break
            set substr [string range $string $start $end]
            uplevel [list set $var $substr]
            set subresult [uplevel $script]
            incr start $delta
            incr end $delta
            set res [string replace $res $start $end $subresult]
            incr delta [expr {[string length $subresult]-[string length $substr]}]
        }
        return $res
    }

Note that you can use regsub and subst combined to get the same effect, but
often it is less secure (against nontrusted inputs) and usually more tricky.

Comments to [SS]