Add bounds checks to extractRangeText and split into helper functions

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2 files changed, 126 insertions, 14 deletions

diff --git a/internal/lsp/extract_range_text_test.go b/internal/lsp/extract_range_text_test.go
new file mode 100644
index 0000000..1fe29ae
--- /dev/null
+++ b/internal/lsp/extract_range_text_test.go
@@ -0,0 +1,83 @@
+package lsp
+
+import "testing"
+
+func TestExtractRangeText_NilDocument(t *testing.T) {
+	got := extractRangeText(nil, Range{})
+	if got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}
+
+func TestExtractRangeText_EmptyLines(t *testing.T) {
+	d := &document{lines: []string{}}
+	got := extractRangeText(d, Range{Start: Position{0, 0}, End: Position{0, 5}})
+	if got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}
+
+func TestExtractRangeText_NegativeStartLine(t *testing.T) {
+	d := &document{lines: []string{"hello"}}
+	got := extractRangeText(d, Range{Start: Position{-1, 0}, End: Position{0, 3}})
+	if got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}
+
+func TestExtractRangeText_NegativeEndLine(t *testing.T) {
+	d := &document{lines: []string{"hello"}}
+	got := extractRangeText(d, Range{Start: Position{0, 0}, End: Position{-1, 3}})
+	if got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}
+
+func TestExtractRangeText_StartLineBeyondEnd(t *testing.T) {
+	d := &document{lines: []string{"hello", "world"}}
+	got := extractRangeText(d, Range{Start: Position{5, 0}, End: Position{6, 0}})
+	if got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}
+
+func TestExtractRangeText_EndLineBeyondEnd(t *testing.T) {
+	// End line is clamped to the last valid line.
+	d := &document{lines: []string{"hello", "world"}}
+	got := extractRangeText(d, Range{Start: Position{0, 0}, End: Position{10, 5}})
+	if got != "hello\nworld" {
+		t.Fatalf("expected clamped result, got %q", got)
+	}
+}
+
+func TestExtractRangeText_StartAfterEnd(t *testing.T) {
+	d := &document{lines: []string{"aaa", "bbb"}}
+	got := extractRangeText(d, Range{Start: Position{1, 0}, End: Position{0, 2}})
+	if got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}
+
+func TestExtractRangeText_SingleLine(t *testing.T) {
+	d := &document{lines: []string{"hello world"}}
+	got := extractRangeText(d, Range{Start: Position{0, 6}, End: Position{0, 11}})
+	if got != "world" {
+		t.Fatalf("expected %q, got %q", "world", got)
+	}
+}
+
+func TestExtractRangeText_SingleLineNegativeChar(t *testing.T) {
+	d := &document{lines: []string{"hello"}}
+	got := extractRangeText(d, Range{Start: Position{0, -3}, End: Position{0, 3}})
+	if got != "hel" {
+		t.Fatalf("expected %q, got %q", "hel", got)
+	}
+}
+
+func TestExtractRangeText_MultiLine(t *testing.T) {
+	d := &document{lines: []string{"aaa", "bbb", "ccc"}}
+	got := extractRangeText(d, Range{Start: Position{0, 1}, End: Position{2, 2}})
+	if got != "aa\nbbb\ncc" {
+		t.Fatalf("expected %q, got %q", "aa\nbbb\ncc", got)
+	}
+}
diff --git a/internal/lsp/handlers_utils.go b/internal/lsp/handlers_utils.go
index 4a5bf4a..620b3a9 100644
--- a/internal/lsp/handlers_utils.go
+++ b/internal/lsp/handlers_utils.go
@@ -550,23 +550,52 @@ func labelForCompletion(cleaned, filter string) string {
 }
 
 // extractRangeText returns the exact text within the given document range.
+// It performs bounds checks on line indices and character offsets, returning
+// an empty string when the range is invalid (e.g. negative lines, out-of-bounds
+// lines, or an empty document).
 func extractRangeText(d *document, r Range) string {
+	if d == nil || len(d.lines) == 0 {
+		return ""
+	}
+	// Clamp line indices to valid bounds.
+	if r.Start.Line < 0 || r.End.Line < 0 || r.Start.Line >= len(d.lines) {
+		return ""
+	}
+	if r.End.Line >= len(d.lines) {
+		r.End.Line = len(d.lines) - 1
+		r.End.Character = len(d.lines[r.End.Line])
+	}
+	if r.Start.Line > r.End.Line {
+		return ""
+	}
+
 	if r.Start.Line == r.End.Line {
-		line := d.lines[r.Start.Line]
-		if r.Start.Character < 0 {
-			r.Start.Character = 0
-		}
-		if r.End.Character > len(line) {
-			r.End.Character = len(line)
-		}
-		if r.Start.Character > r.End.Character {
-			return ""
-		}
-		return line[r.Start.Character:r.End.Character]
+		return extractSingleLineRange(d.lines[r.Start.Line], r)
+	}
+	return extractMultiLineRange(d.lines, r)
+}
+
+// extractSingleLineRange handles the case where start and end are on the same line.
+// Character offsets are clamped to the line length.
+func extractSingleLineRange(line string, r Range) string {
+	if r.Start.Character < 0 {
+		r.Start.Character = 0
 	}
+	if r.End.Character > len(line) {
+		r.End.Character = len(line)
+	}
+	if r.Start.Character > r.End.Character {
+		return ""
+	}
+	return line[r.Start.Character:r.End.Character]
+}
+
+// extractMultiLineRange handles ranges spanning multiple lines, clamping
+// character offsets on the first and last lines.
+func extractMultiLineRange(lines []string, r Range) string {
 	var b strings.Builder
 	// first line
-	first := d.lines[r.Start.Line]
+	first := lines[r.Start.Line]
 	if r.Start.Character < 0 {
 		r.Start.Character = 0
 	}
@@ -577,13 +606,13 @@ func extractRangeText(d *document, r Range) string {
 	b.WriteString("\n")
 	// middle lines
 	for i := r.Start.Line + 1; i < r.End.Line; i++ {
-		b.WriteString(d.lines[i])
+		b.WriteString(lines[i])
 		if i+1 <= r.End.Line {
 			b.WriteString("\n")
 		}
 	}
 	// last line
-	last := d.lines[r.End.Line]
+	last := lines[r.End.Line]
 	if r.End.Character < 0 {
 		r.End.Character = 0
 	}