I want to extract sequences of words according to how many characters in each word.

So I want to extract for instance a sequence based on the number of characters (here defined as letters of the alphabet - not punctuation, numbers, white space).

For instance: I want sequences of 2, 4 and 3 character words - in that order only (but it could be any numbers of characters in any order I choose).

Say my text is: "xxxx yy zzzzz xxxx qqq"

I should extract the sequence: "yy xxxx qqq"

and keep on doing it. So from "xxxx yy zzzzz xxxx qqq xxxx yy zzzzz xxxx qqq"

I should extract

"yy xxxx qqq yy xxxx qqq"

 
my $string = "xxxx yy zzzzz xxxx qqq";  

my @array = ( $string =~  /(\b..?\b) (\b....?\b)  (\b...?\b)/sg );

print @array;


# produces nothing. 

# I have also tried rewriting it without success:  it may 
# produce results, but not the right ones! (not the exact 
# sequence)

# also if the string were longer it should produce
# the sequence repeated: 
# "xxxx yy zzzzz xxxx qqq xxxx yy zzzzz xxxx qqq" 
# should produce "yy xxxx qqq  yy xxxx qqq" etc etc 
# until we run out of text.
[download]

I have also tried running adaptions of remiah's code, but without success: http://www.perlmonks.org/?node_id=996670. The problem/task differs and I cannot adapt the code to it. Inability! nicemank thanks in advance!