Pattern Matcher finds match but cannot get text due to \r - works fine in regex testers

Question

For the given text obtained from a text area field:

Apex script unhandled exception by user/organization: xxx/xxx

Visualforce Page: /apex/pagename

caused by: System.NullPointerException: Attempt to de-reference a null >object

trace line 1

trace line 2

I want to match and capture

Visualforce Page: /apex/pagename

Note: This line can be anything but will always be a single line of text with followed by one or more \n then the line starting with caused by

I have tried the following regex at regex101 and it shows the FULL MATCH being "Visualforce Page: /apex/pagename "

In salesforce I am using the following pattern

(?im).*(?=\\n*caused by.*)

Which in regex101 grabs the first non empty line prior to the "caused by"

Now when I run the following code I get a true value debugged

System.debug(
     Pattern.compile('(?m).*(?=\\n*caused by.*)')
        .matcher(t).find()
);

But I cannot seem to get the actual text of the match.....

I am using the following code and get the following debug

System.debug(t);

Pattern p = Pattern.compile('(?im).*(?=\\n*caused by.*)');

Matcher m = p.matcher(t);

while(m.find()){
    System.debug(m.groupCount());
    System.debug(m.start());
    System.debug(m.end());
    System.debug(m.group(0));
}

Note: I also tried the following pattern with same results (?im)(.*)(?=\\n*caused by)

Update -- Still not working but

I added a check for charAt to see what group 1 was returning when I used the pattern: (?im)(.*)(?=\\n*caused by.*)

doing a m.group(1).length gets 0

so I changed the pattern to (?im).*[^\\n](?=\\n*caused by.*)

doing a m.group(0).length gets 0 for the first find and 1 for the second find. The second find is char(13)

System.debug(m.groupCount());
String x = m.group(0);
System.debug(x);
System.debug(x.length());
System.debug(x.charAt(0));

trying to do the pattern (?im).*[^\\r\\n](?=\\n*caused by.*)results in no matches...

When copied and pasted into regex101 it works fine but seems apex is having issues. I cannot seem to get anything but the CR or LF back and when I try to exclude both I get no matches

BUT If I do this before the matching process it works fine:

t = t.replaceAll('\\r','');

I am not the best at regex so I cannot seem to get it to exclude the \r as when I include [^\\r\\n] when not replacing the \r I get no matches

I am not the best at regex and pattern and matcher so any help on getting the matched text would be great

Answer 1

Ok, so you've got carriage returns in the mix here too.

Newlines and carriage returns aren't matched by the wildcard . by default. You can add the 's' flag to your regex options to include newline characters in ., but that'd make it more difficult to isolate your substring match to just the second line.

Instead, you can use a capturing group inside of your lookahead to check for 0 or more \r or \n

String t = 'Apex script unhandled exception by user/organization: xxx/xxx\r\n\r\nVisualforce Page: /apex/pagename\r\n\r\ncaused by: System.NullPointerException: Attempt to de-reference a null >object\r\n\r\ntrace line 1\r\n\r\ntrace line 2';

System.debug(t);

Pattern p = Pattern.compile('(?im).*(?=(\\r|\\n)*caused by.*)');

Matcher m = p.matcher(t);

while(m.find()){
    System.debug(m.start());
    System.debug(m.end());
    System.debug(m.group());
}