Context extraction is too slow when a file is big.

[sanastasiou]

Open a large file and try e.g. reference lookup on the 10th line.

The result should be instant.

Instead, it takes several seconds for the token to be highlighted.