regexp — {b?} abcd match

[Bob Schmid]

the above command comes back with one hit but no value…..? why

non-greedy search…shouldn’t it find the “b” ?

tcl>regexp — {b?} abcd match

1

tcl>echo $match

tcl>

tcl>regexp — {a?} abcd match

1

tcl>echo $match

a

tcl>

Come on Charlie….lay it on me….Im almost afraid to ask!

Topic

[Max Drown (Infor)]

What are you trying to do exactly?

The return value of “1” means it did match on your expression.

I’m actually not sure what regexp stores in the match variable. It wasn’t clear, at least to my eyes, in the documentation … “match will be set to the range of string that matched all of exp.“

If you want to capture the value, you’ll need to use parens and variables like this: regexp — {(b)} abdc match var1; puts $var1.

If you want to use the expression in a decision, do it like this: if {[regexp {b} abdc]} {…}.

-- Max Drown (Infor)

[Tom Rioux]

Not sure what you are trying to do but one way (provided the pattern you are looking for isnt at the beginning or the end of a word),

hcitcl>regexp — {bY} abcd match

1

hcitcl>echo $match

b

What ARE you trying to do?

Tom

[Bob Schmid]

we are going thru some regexp examples….understanding greedy vs non-greedy…and was wondering if in fact the above got a hit…..what did it get a hit on ?

[Max Drown (Infor)]

Took me years of experience to get a good grasp on greediness.

-- Max Drown (Infor)

[Max Drown (Infor)]

Robert Schmid wrote:

if b? is greedy (match as much as possible) why did it not find the b in “abcd”..

It *DID* find it. That’s why it returned a “1”.

-- Max Drown (Infor)

[Max Drown (Infor)]

Oh. I think I see your problem. The pattern “b?” means 0 or 1 b’s. You have to use a modifier like . * + for greediness modifications.

Ex. b*?, b+?

Code:

regexp — {b} abcd match
puts $match
>b
regexp — {a} abcd match
puts $match
>a

-- Max Drown (Infor)

[David Barr]

Max Drown wrote:

It *DID* find it. That’s why it returned a “1”.

Maybe it didn’t find the “b”.  This expression returns a one as well:

Code:

regexp — {b?} acd match

That’s because the question mark means to match zero or one occurrences of the previous character.  The pattern {b?} will always match any string.  

Robert, I think that the reason you’re seeing this behavior is due to confusion about what greedy means.  I think that it means that if it finds a match, it will try to use as many characters around that match as possible.  However, if there are multiple matches, a greedy operator won’t match the longest one, it will just match the first one.  In your case, every single character in the string matches your pattern, so it tries to match the first character, a.  The greedy nature of the expression causes it to evaluate if it can match one occurance of “b” against the “a” instead of zero, but that fails.

You’ll see slightly different results with this test:

Code:

regexp — {b?} bcde match

In this case, a “b” will be stored in match.

[Charlie Bursell]

Don’t you just hate it when it does exactly what you told it to do?  🙄

[Author]

Replies