Regular expression doesn't match string when using MAWK

user_337149 0
Viewed 80

I have defined a regular expression which matches a list of words separated by one or more spaces where one of the words is followed by an asterisk. The strange thing is that for a list with only one word the expression doesn't match when I use mawk but it matches when I use gawk and nawk:

$ echo 'a*' | mawk '/([a-z]+ *)*[a-z]+ *[*]( *[a-z]+)*/'
$ echo 'a*' | gawk '/([a-z]+ *)*[a-z]+ *[*]( *[a-z]+)*/'
a*
$ echo 'a*' | nawk '/([a-z]+ *)*[a-z]+ *[*]( *[a-z]+)*/'
a*

If the word with the asterisk is followed by one or more words then the regular expression matches also when using mawk:

$ echo 'a* b' | mawk '/([a-z]+ *)*[a-z]+ *[*]( *[a-z]+)*/'
a* b

Any clues?

In Debian 11, mawk is the default implementation of AWK.

$ mawk -W version
mawk 1.3.4 20200120
Copyright 2008-2019,2020, Thomas E. Dickey
Copyright 1991-1996,2014, Michael D. Brennan

random-funcs:       srandom/random
regex-funcs:        internal
compiled limits:
sprintf buffer      8192
maximum-integer     2147483647
awk regex
1 Answers

no idea how to fix mawk-1 ittself, but if you want regex to circumvent its shortcomings :

echo 'a*' | mawk '/([a-z]+ *)*[a-z]+ *[*](( *[a-z]+)*)?/'

    a*

but since all the stuff before or prior are merely "zero or more", and you need the whole line anyway, then why not just

echo 'a*' | mawk '/[a-z]+([ ]+)?[*]/' 

a*

or even more minimalistically,

echo 'a*' | mawk '/[a-z] +?\*/'  

a*

if you want a tight criteria around it, then maybe

echo 'a*' | mawk '/([a-z]+ +?)+[*]( +?[a-z]+)+?/'

a*

using [...]+? instead of [...]* sometimes may be more friendly with regex engines.

most modern regex engines shouldn't have any issues with [...]* , but this would be one of those scenarios where the slightly less intuitive appearing syntax provides meaningful assistance to the engine

user_146721140
Related