CODEWITHSUNDEEP
javaregex
OnTheFly
OnTheFly

Reputation: 2101

Regex - get code between two comments

How can I get

Text t;
Graph g;

from

// Co
Text t;
Graph g;
// Co

?

// Co\r?\n(.*)\r?\n// Co

works when there is one line, but does not work with multiple lines.

Upvotes: 2

Views: 200

Answers (1)

ruakh
ruakh

Reputation: 183602

By default, . matches any character except newlines (\n), but you can use the Pattern.DOTALL flag to make it match newlines as well. So, for example, instead of this:

final Pattern pat = Pattern.compile("// Co\r?\n(.*)\r?\n// Co");

you would use this:

final Pattern pat = Pattern.compile("// Co\r?\n(.*)\r?\n// Co", Pattern.DOTALL);

Equivalently, you can set the s flag inside the regex itself, using either of these styles:

final Pattern pat = Pattern.compile("// Co\r?\n((?s).*)\r?\n// Co");
final Pattern pat = Pattern.compile("// Co\r?\n(?s:(.*))\r?\n// Co");

which lets you control exactly which part of the regex accepts a newline for .. (But I'd stick with the DOTALL approach unless you have a regex where . sometimes means the one thing, sometimes the other.)

Edited to add: It seems very likely — but you'll have to determine this for yourself, based on your use-case — that you need to use non-greedy quantification, .*?, rather than greedy quantification, .*.

Upvotes: 5

Related Questions