From b8867a0239b1930a16f9ef3f7f3e864b01416dff Mon Sep 17 00:00:00 2001
From: "Nicolas B. Pierron" <nicolas.b.pierron@gmail.com>
Date: Tue, 15 Aug 2017 18:08:41 +0000
Subject: Add builtins.string function.

The function 'builtins.split' takes a POSIX extended regular expression
and an arbitrary string. It returns a list of non-matching substring
interleaved by lists of matched groups of the regular expression.

```nix
with builtins;
assert split "(a)b" "abc"      == [ "" [ "a" ] "c" ];
assert split "([ac])" "abc"    == [ "" [ "a" ] "b" [ "c" ] "" ];
assert split "(a)|(c)" "abc"   == [ "" [ "a" null ] "b" [ null "c" ] "" ];
assert split "([[:upper:]]+)" "  FOO   "
                               == [ "  " [ "FOO" ] "   " ];
```
---
 tests/lang/eval-okay-regex-split.nix | 48 ++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)
 create mode 100644 tests/lang/eval-okay-regex-split.nix

(limited to 'tests/lang/eval-okay-regex-split.nix')

diff --git a/tests/lang/eval-okay-regex-split.nix b/tests/lang/eval-okay-regex-split.nix
new file mode 100644
index 000000000000..0073e057787d
--- /dev/null
+++ b/tests/lang/eval-okay-regex-split.nix
@@ -0,0 +1,48 @@
+with builtins;
+
+# Non capturing regex returns empty lists
+assert  split "foobar" "foobar"  == ["" [] ""];
+assert  split "fo*" "f"          == ["" [] ""];
+assert  split "fo+" "f"          == ["f"];
+assert  split "fo*" "fo"         == ["" [] ""];
+assert  split "fo*" "foo"        == ["" [] ""];
+assert  split "fo+" "foo"        == ["" [] ""];
+assert  split "fo{1,2}" "foo"    == ["" [] ""];
+assert  split "fo{1,2}" "fooo"   == ["" [] "o"];
+assert  split "fo*" "foobar"     == ["" [] "bar"];
+
+# Capturing regex returns a list of sub-matches
+assert  split "(fo*)" "f"        == ["" ["f"] ""];
+assert  split "(fo+)" "f"        == ["f"];
+assert  split "(fo*)" "fo"       == ["" ["fo"] ""];
+assert  split "(f)(o*)" "f"      == ["" ["f" ""] ""];
+assert  split "(f)(o*)" "foo"    == ["" ["f" "oo"] ""];
+assert  split "(fo+)" "foo"      == ["" ["foo"] ""];
+assert  split "(fo{1,2})" "foo"  == ["" ["foo"] ""];
+assert  split "(fo{1,2})" "fooo" == ["" ["foo"] "o"];
+assert  split "(fo*)" "foobar"   == ["" ["foo"] "bar"];
+
+# Matches are greedy.
+assert  split "(o+)" "oooofoooo" == ["" ["oooo"] "f" ["oooo"] ""];
+
+# Matches multiple times.
+assert  split "(b)" "foobarbaz"  == ["foo" ["b"] "ar" ["b"] "az"];
+
+# Split large strings containing newlines. null are inserted when a
+# pattern within the current did not match anything.
+assert  split "[[:space:]]+|([',.!?])" ''
+  Nix Rocks!
+  That's why I use it.
+''  == [
+  "Nix" [ null ] "Rocks" ["!"] "" [ null ]
+  "That" ["'"] "s" [ null ] "why" [ null ] "I" [ null ] "use" [ null ] "it" ["."] "" [ null ]
+  ""
+];
+
+# Documentation examples
+assert  split  "(a)b" "abc"      == [ "" [ "a" ] "c" ];
+assert  split  "([ac])" "abc"    == [ "" [ "a" ] "b" [ "c" ] "" ];
+assert  split  "(a)|(c)" "abc"   == [ "" [ "a" null ] "b" [ null "c" ] "" ];
+assert  split  "([[:upper:]]+)" "  FOO   " == [ "  " [ "FOO" ] "   " ];
+
+true
-- 
cgit 1.4.1