Using a capturing group of the match pattern of a regsub in the substitution itself?-CodePudding

In the code below, $html is a string of HTML. The code captures the full match and the capturing group for each match in a list. Then, if the list is not empty, it iterates through the list to replace the span tags with em tags along with the original text that was between them.

For example if the HTML is: This is span 1 and this is span 2. then $a would be a list of length 4: {span 1} {span 1} {span 2} {span 2}.

The sample code generates: This is span 1 and this is span 2. as expected; but it seems that this must be an inefficient way to do this and, somehow, the capturing group should be usable directly within the regsub expression.

Is this true and how is it done?

Something like:

set html [regsub "(. ?)" $html ".../em>"]

where the ... is something that points to the captured group.

Thank you.

set a [regexp -all -inline -- {<span class='add'>(. ?)</span>} $html]
if { [llength $a] > 0 } {
  foreach {x y} $a {
    set html [regsub "<span class='add'>${y}</span>" $html "<em>${y}</em>"]
  }
}

CodePudding user response：

You can use a backreference in the replacement:

regsub -all {<span class='add'>(. ?)</span>} $html {<em>\1</em>}