regextclword-boundaries

regexp with word boundaries not working in TCL

I have following code:

set variable1 "5"
set variable2 "6-7"
if {[regexp {\m\d+\M} $variable1 matchall]} {
  puts "only a number has been provided and that is $matchall"
}
if {[regexp {\m\d+\M} $variable2 matchall]} {
  puts "only a number has been provided and that is $matchall"
}

Basically, i want to only match variable1 which is just a number. variable2 does have a hyphen inside which should not match because i am using the \m and \M which are basically doing the same as \b in normal regex which indicates word boundaries. So when i run the code i get:

only a number has been provided and that is 5
only a number has been provided and that is 6

can anyone give me tip in the right direction? I am stuck with TCL 8.3 and cannot upgrade. But regex should basically work the same way as with never versions.

Solution

  • In addition to Barmar’s answer, it is worth noting that you do not need a regular expression here. Just test with the string command:

    string is integer -strict $s

    This permits leading and trailing whitespace, though. If that is significant, you can trim it away with

    string trim $s

    before using it anywhere. To wit:

    if {[string is integer -strict $variable1]} {
    puts "only a number has been provided and that is [string trim $variable1]"
}