Regex pattern matching cursor-position

Question

Considering the following regex pattern:

/\s*<<.\{-}>>/

I want to match those patterns only when the cursor is somewhere on this pattern.

So, if have the line (the character enclosed in [] is the cusor position):

This is <<on[e]>> and this is <<another one>>.

I want to only match <<one>>, not <<another one>>.

I know that \%# specifies the cursor position, but I don't know how to apply it to state that the cursor may be anywhere on that pattern.

Is this possible somehow?

Answer 1

One solution is to use a lookahead before the regex and a lookbehind after the regex. That ensures the cursor is after the start of the match, and before the end of the match.

One issue with the regex you're using is .\{-} which is a non-greedy operator. If you add anchors for the cursor position, this will end up matching from the beginning of <<one to the end of another one>>. To solve that, I'm replacing it with \([^>]\|>[^>]\)*, which will match any character other than >, or a single > followed by a different character. (Perhaps a negative match would have worked here instead.)

Putting it all together:

/\(.*\%#\)\@=\s*<<\([^>]\|>[^>]\)*>>\(\%#.*\)\@<=

UPDATE: Using \@> for the middle bit is able to preserve the -\{-}, resulting in:

 /\(.*\%#\)\@=\(\s*<<.\{-}>>\)\@>\(\%#.*\)\@<=

Answer 2

You can use the atom /\&. This is like forcing two different regular atoms to match at the same position.

Now it becomes a bit complicated, since the cursor position can be anywhere inside the <<..>> and then you also want to allow whitespace inside those <<..>>. This makes it a bit complex to match the correct item if there are several matches in a line.

So instead of using .\{-\} to allow an arbitrarily long number of matches (even other <<), you need to restrict your match to only match until the >>.

To make the pattern simpler, let's for now assume all characters are allowed except for >. That would make the pattern like this:

/<[^>]\+>>\&[^>]*\%#[^>]*>>

which means < followed by any character except > ([^>]) and at the same time (\&) allow any character except >, followed by the cursor position (\%#), followed by any character except > followed by >>.

Now you can make this more complicated by enforcing the >> delimiter. Possible search patterns are \(>>\)\@! (not followed by >>) or the already mentioned \([^>]\|>[^>]\). However this will make your search pattern a lot more complex, so this is left as an exercise to the reader ;)

You might also want to have a look at the distributed plugin LogiPat, which allows to construct patterns using logical operators (and, or, not, etc).