Groovy: Why do I need to double escape square brackets?

Question

Why doesn't this work?

"hello[world".split("\[")

and this does:

"hello[world".split("\\[")

Answer 1

This is actually two escapes in different contexts, due to the fact that the argument is a regular expression represented as a string.

The [ has to be escaped because otherwise it would have a special meaning in the regular expression. The escape for the regular expression would make it \[. But then the \ has to be escaped as it has special meaning in a string (for escaping and for representing characters by numeric value).

It can be much worse, as the \ character is used for escaping in both contexts. If you want to split by the \ character, you have to escape it (\\) for the regular expression usage, but then you have two \ characters, which both have to be escaped in string context. The usage in the original string you are splitting would also need the escape if you're writing it as a constant, so the analogous split would look like:

"hello\\world".split("\\\\")

Answer 2

Different versions of Groovy give different results.

For Groovy 1.1-BETA-1 and onwards,

assert "hello[world".split("\\[") == ["hello", "world"]

asserts OK, but Groovy 1.0 and before gives an assertion error.

assert "hello[world".split(/\[/) == ["hello", "world"]

asserts OK for all versions from 1.1-BETA-1 onwards, Groovy 1.0-jsr-01 to 1.0 gives an assertion error, while Groovy 1.0-beta-10 and before gives a lexical error.

"hello[world".split("\\[").each{println it}

for all versions of Groovy from 1.0-beta-5, and beta-3, prints

hello
world

but 1.0-beta-4 prints

[.]
[.]
hello
world