From 3b9750f0a4640fd54e7dd2714c997b62ef8df33c Mon Sep 17 00:00:00 2001
From: Dayna Blackwell <dayna@blackwell-systems.com>
Date: Mon, 27 Apr 2026 10:48:19 -0700
Subject: [PATCH] fix(text-splitters): remove incorrect C# and Elixir separator
 keywords (#37037)

## Summary

Removes two incorrect separators from `get_separators_for_language()` in
`RecursiveCharacterTextSplitter`:

- **C#**: `"\nimplements "` is a Java keyword. C# uses `:` for interface
implementation. This separator never matches valid C# source code.
- **Elixir**: `"\nwhile "` does not exist in Elixir. The language uses
recursion and `Enum.reduce_while/3` instead of while loops.

Both are dead separators that silently degrade chunking quality by
occupying positions in the separator priority list without contributing
useful split points.

## Tests

Added two targeted tests:
- `test_csharp_separators_no_java_keywords`: verifies `"\nimplements "`
is not in the C# separator list
- `test_elixir_separators_no_while`: verifies `"\nwhile "` is not in the
Elixir separator list

Existing `test_csharp_code_splitter` continues to pass (no change to
expected output since `implements` never matched valid C# code).

Full suite: 129 passed, 0 failed.

Fixes #37030
---
 .../langchain_text_splitters/character.py       |  2 --
 .../tests/unit_tests/test_text_splitters.py     | 17 +++++++++++++++++
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/libs/text-splitters/langchain_text_splitters/character.py b/libs/text-splitters/langchain_text_splitters/character.py
index 469afdb9132..dd2378a7876 100644
--- a/libs/text-splitters/langchain_text_splitters/character.py
+++ b/libs/text-splitters/langchain_text_splitters/character.py
@@ -440,7 +440,6 @@ class RecursiveCharacterTextSplitter(TextSplitter):
                 # Split along control flow statements
                 "\nif ",
                 "\nunless ",
-                "\nwhile ",
                 "\ncase ",
                 "\ncond ",
                 "\nwith ",
@@ -593,7 +592,6 @@ class RecursiveCharacterTextSplitter(TextSplitter):
             return [
                 "\ninterface ",
                 "\nenum ",
-                "\nimplements ",
                 "\ndelegate ",
                 "\nevent ",
                 # Split along class definitions
diff --git a/libs/text-splitters/tests/unit_tests/test_text_splitters.py b/libs/text-splitters/tests/unit_tests/test_text_splitters.py
index c7d7e2c480d..cae9f16d103 100644
--- a/libs/text-splitters/tests/unit_tests/test_text_splitters.py
+++ b/libs/text-splitters/tests/unit_tests/test_text_splitters.py
@@ -1011,6 +1011,23 @@ class Program
     ]
 
 
+def test_csharp_separators_no_java_keywords() -> None:
+    """C# separators should not contain Java-only keywords."""
+    splitter = RecursiveCharacterTextSplitter.from_language(
+        Language.CSHARP, chunk_size=CHUNK_SIZE, chunk_overlap=0
+    )
+    # "implements" is a Java keyword; C# uses ":" for interface implementation
+    assert "\nimplements " not in splitter._separators
+
+
+def test_elixir_separators_no_while() -> None:
+    """Elixir has no while loop; the separator should not be present."""
+    splitter = RecursiveCharacterTextSplitter.from_language(
+        Language.ELIXIR, chunk_size=CHUNK_SIZE, chunk_overlap=0
+    )
+    assert "\nwhile " not in splitter._separators
+
+
 def test_cpp_code_splitter() -> None:
     splitter = RecursiveCharacterTextSplitter.from_language(
         Language.CPP, chunk_size=CHUNK_SIZE, chunk_overlap=0