From 65d6bba9254d514a16e0bf89d76cbd9b90740e5c Mon Sep 17 00:00:00 2001
From: Daniel Parks <dp+git@oxidized.org>
Date: Mon, 18 Nov 2019 01:26:20 -0800
Subject: [PATCH] api: fix for splitn("a", 2) returning extra ""

Corrects `/-/.splitn("a", 2)` to return `["a"]` instead of `["a", ""]`.
(`/-/` is shorthand for `Regex::new("-").unwrap()`.)

Fixes #521, Closes #606, Closes #628
---
 CHANGELOG.md      |  4 +++-
 src/re_bytes.rs   | 15 +++++++++++----
 src/re_unicode.rs | 15 +++++++++++----
 tests/api.rs      |  4 +---
 4 files changed, 26 insertions(+), 12 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index db36d21328..2606af8bd5 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -9,10 +9,12 @@ New features:
 
 Bug fixes:
 
+* [BUG #521](https://github.com/rust-lang/regex/issues/521):
+  Corrects `/-/.splitn("a", 2)` to return `["a"]` instead of `["a", ""]`.
 * [BUG #594](https://github.com/rust-lang/regex/pull/594):
   Improve error reporting when writing `\p\`.
 * [BUG #627](https://github.com/rust-lang/regex/issues/627):
-  Corrects `re.split("a-")` to return `["a", ""]` instead of `["a"]`.
+  Corrects `/-/.split("a-")` to return `["a", ""]` instead of `["a"]`.
 * [BUG #633](https://github.com/rust-lang/regex/pull/633):
   Squash deprecation warnings for the `std::error::Error::description` method.
 
diff --git a/src/re_bytes.rs b/src/re_bytes.rs
index dff70c0ba9..69f0b335de 100644
--- a/src/re_bytes.rs
+++ b/src/re_bytes.rs
@@ -774,12 +774,19 @@ impl<'r, 't> Iterator for SplitN<'r, 't> {
         if self.n == 0 {
             return None;
         }
+
         self.n -= 1;
-        if self.n == 0 {
-            let text = self.splits.finder.0.text();
-            Some(&text[self.splits.last..])
+        if self.n > 0 {
+            return self.splits.next();
+        }
+
+        let text = self.splits.finder.0.text();
+        if self.splits.last > text.len() {
+            // We've already returned all substrings.
+            None
         } else {
-            self.splits.next()
+            // self.n == 0, so future calls will return None immediately
+            Some(&text[self.splits.last..])
         }
     }
 }
diff --git a/src/re_unicode.rs b/src/re_unicode.rs
index 7498757830..b746599088 100644
--- a/src/re_unicode.rs
+++ b/src/re_unicode.rs
@@ -814,12 +814,19 @@ impl<'r, 't> Iterator for SplitN<'r, 't> {
         if self.n == 0 {
             return None;
         }
+
         self.n -= 1;
-        if self.n == 0 {
-            let text = self.splits.finder.0.text();
-            Some(&text[self.splits.last..])
+        if self.n > 0 {
+            return self.splits.next();
+        }
+
+        let text = self.splits.finder.0.text();
+        if self.splits.last > text.len() {
+            // We've already returned all substrings.
+            None
         } else {
-            self.splits.next()
+            // self.n == 0, so future calls will return None immediately
+            Some(&text[self.splits.last..])
         }
     }
 }
diff --git a/tests/api.rs b/tests/api.rs
index e16f7264bc..0d4962cc9f 100644
--- a/tests/api.rs
+++ b/tests/api.rs
@@ -213,9 +213,7 @@ split!(split_trailing_blank, r"-", r"a-", &[t!("a"), t!("")]);
 split!(split_trailing_blanks, r"-", r"a--", &[t!("a"), t!(""), t!("")]);
 split!(split_empty, r"-", r"", &[t!("")]);
 
-// See: https://github.com/rust-lang/regex/issues/521
-// splitn!(splitn_below_limit, r"-", r"a", 2, &[t!("a")]);
-
+splitn!(splitn_below_limit, r"-", r"a", 2, &[t!("a")]);
 splitn!(splitn_at_limit, r"-", r"a-b", 2, &[t!("a"), t!("b")]);
 splitn!(splitn_above_limit, r"-", r"a-b-c", 2, &[t!("a"), t!("b-c")]);
 splitn!(splitn_zero_limit, r"-", r"a-b", 0, empty_vec!());