CODEWITHSUNDEEP
regexruby-2.0
Bertaud
Bertaud

Reputation: 2918

regex: stop matching when end of line

I want to scan a text composed of lines. When I recognize a keyword I am using a regex for this type of line.

My text has this form:

text<eol>
...
Function1 parameter1 value1, parameter2 value2, .... parameterN value N<eol>
...
text<eol>

In this case, when I recognize "Function1", I use this regex:

(?:(\w+)\s|\G(?<!^))[ ]*(\S+[ ]+\S+)\s*,?

$1 = the keyword Function1
$3 = parameterx valuex

I would like to stop the matching when the end of line is found.

while mytext =~ /.../ do
 case $3
  when 'parameter1 value1'
   ...
 end #case
 $'
end #while

My code is correct excepted when the matching is close to the end of line:my regex captures a part a the next line.

The question is: what can I add in my regex to stop when end of line ? I suppose I must add $ somewhere ?

Note: Sorry I copied the wrong regex ([] gave me an error with Ruby 2.0). Note the space before the first star !

(?:(\w+)(\s+)|\G(?<!^)) *(\S+\s+\S+)\s*,?

Upvotes: 1

Views: 7955

Answers (3)

aliteralmind
aliteralmind

Reputation: 20163

To stop matching at the end of lines, you should first make sure that the dotall flag is not active.

If the regular expression 

(?:(\w+)\s|\G(?<!^))[ ]*(\S+[ ]+\S+)\s*,?

is working for you, then simply add start (^) and/or end-of-line ($) boundaries to it, such as:

^(?:(\w+)\s|\G(?<!^))[ ]*(\S+[ ]+\S+)\s*,?$

But it seems that your goal is to first find a line with a function-signature and capture the function name. So first you'll need to match the entire line with

^(\w+)\s+(?:(?:\w+\s+\w+)(?:, *)?)+$

Regular expression visualization

Debuggex Demo

(The function name is in capture group one.)

Then, starting at the first non-whitespace character following the function name, iterate through each parameter with

\b(\w+)\s+(\w+)\b

Regular expression visualization

Debuggex Demo

Upvotes: 1

rohit01
rohit01

Reputation: 227

Not sure about regex specific to ruby. But I think, adding [^<eol>] should help exclude new lines. Try this simple change:

(?:(\w+)(\s+)|\G(?<!^)) *(\S+\s+\S+)\s*[^<eol>],?

Upvotes: 0

Jerry
Jerry

Reputation: 71538

Okay, since you asked for it, here it is:

(?:(\w+) +|(?!^)\G) *(\S+ +[^\n, ]+) *,?

I changed all the \s into whitespaces and the last \S to [^\n, ] so that the comma after the value is not consumed during the matching.

I made some more minor changes regarding the capture groups and the way you were negating the ^ for the \G anchor.

It's working on PCRE flavoured regex like on this demo site but like I said, I'm not entirely sure \G works on Ruby. Would be glad if someone could confirm this.

Upvotes: 1

Related Questions