Do no longer escape backslashes when printing strings or bytes.

Escaping backslashes in `print` output seems both unnecessary and confusing. Closes #1895.
zeek · Oct 23, 2024 · edf96cb · edf96cb
1 parent b439d1d
commit edf96cb
Show file tree

Hide file tree

Showing 9 changed files with 55 additions and 15 deletions.
diff --git a/hilti/runtime/include/types/bytes.h b/hilti/runtime/include/types/bytes.h
@@ -572,7 +572,7 @@ class Bytes : protected std::string {
 };
 
 inline std::ostream& operator<<(std::ostream& out, const Bytes& x) {
-    out << escapeBytes(x.str());
+    out << escapeBytes(x.str(), render_style::Bytes::NoEscapeBackslash);
     return out;
 }
 
@@ -584,7 +584,7 @@ inline Bytes operator"" _b(const char* str, size_t size) { return Bytes(Bytes::B
 
 template<>
 inline std::string detail::to_string_for_print<Bytes>(const Bytes& x) {
-    return escapeBytes(x.str());
+    return escapeBytes(x.str(), render_style::Bytes::NoEscapeBackslash);
 }
 
 namespace detail::adl {

diff --git a/hilti/runtime/include/types/string.h b/hilti/runtime/include/types/string.h
@@ -108,12 +108,14 @@ inline std::string to_string(const CharT (&x)[N], adl::tag /*unused*/) {
 
 template<>
 inline std::string detail::to_string_for_print<std::string>(const std::string& x) {
-    return escapeUTF8(x, render_style::UTF8::NoEscapeHex | render_style::UTF8::NoEscapeControl);
+    return escapeUTF8(x, render_style::UTF8::NoEscapeHex | render_style::UTF8::NoEscapeControl |
+                             render_style::UTF8::NoEscapeBackslash);
 }
 
 template<>
 inline std::string detail::to_string_for_print<std::string_view>(const std::string_view& x) {
-    return escapeUTF8(x, render_style::UTF8::NoEscapeHex | render_style::UTF8::NoEscapeControl);
+    return escapeUTF8(x, render_style::UTF8::NoEscapeHex | render_style::UTF8::NoEscapeControl |
+                             render_style::UTF8::NoEscapeBackslash);
 }
 
 // Specialization for string literals. Since `to_string_for_print` is not

diff --git a/hilti/runtime/include/util.h b/hilti/runtime/include/util.h
@@ -297,9 +297,10 @@ namespace render_style {
  * the default style accordingly.
  */
 enum class Bytes {
-    Default = 0,               /**< name for unmodified default style */
-    EscapeQuotes = (1U << 1U), /**< escape double quotes with backslashes */
-    UseOctal = (1U << 2U),     /**< escape non-printables with `\NNN` instead of `\xNN` */
+    Default = 0,                    /**< name for unmodified default style */
+    EscapeQuotes = (1U << 1U),      /**< escape double quotes with backslashes */
+    UseOctal = (1U << 2U),          /**< escape non-printables with `\NNN` instead of `\xNN` */
+    NoEscapeBackslash = (1U << 3U), /**< do not escape backslashes */
 };
 
 /**
@@ -311,11 +312,12 @@ enum class Bytes {
  * through `expandUTF8Escapes()`.
  */
 enum class UTF8 {
-    Default = 0,                  /**< name for unmodified default style */
-    EscapeQuotes = (1U << 1U),    /**< escape double quotes with backslashes */
-    NoEscapeControl = (1U << 2U), /**< do not escape control characters and null bytes */
+    Default = 0,                    /**< name for unmodified default style */
+    EscapeQuotes = (1U << 1U),      /**< escape double quotes with backslashes */
+    NoEscapeBackslash = (1U << 2U), /**< do not escape backslashes; this may leave the result non-reversible */
+    NoEscapeControl = (1U << 3U),   /**< do not escape control characters and null bytes */
     NoEscapeHex =
-        (1U << 3U), /**< do not escape already existing `\xNN` escape codes; this may leave the result non-reversible */
+        (1U << 4U), /**< do not escape already existing `\xNN` escape codes; this may leave the result non-reversible */
 };
 
 } // namespace render_style

diff --git a/hilti/runtime/src/tests/bytes.cc b/hilti/runtime/src/tests/bytes.cc
@@ -653,4 +653,13 @@ TEST_CASE("issue 599") {
     CHECK_EQ(a->toInt(), 31);
 }
 
+TEST_CASE("to_string") {
+    CHECK_EQ(to_string("abc"_b), "b\"abc\"");
+    CHECK_EQ(to_string("\"\\"_b), "b\"\\\"\\\\\"");
+}
+
+TEST_CASE("to_string_for_print") {
+    CHECK_EQ(to_string_for_print("abc"_b), "abc");
+    CHECK_EQ(to_string_for_print("\\\""_b), "\\\"");
+}
 TEST_SUITE_END();
diff --git a/hilti/runtime/src/tests/deferred-expression.cc b/hilti/runtime/src/tests/deferred-expression.cc
@@ -89,8 +89,8 @@ TEST_CASE("to_string_for_print") {
     auto expr = make_deferred<Bytes>([&i]() { return Bytes(fmt("\\x0%d", ++i)); });
 
     // Stringification evaluates the expression.
-    CHECK_EQ(to_string_for_print(expr), R"#(\\x01)#");
-    CHECK_EQ(to_string_for_print(expr), R"#(\\x02)#");
+    CHECK_EQ(to_string_for_print(expr), R"#(\x01)#");
+    CHECK_EQ(to_string_for_print(expr), R"#(\x02)#");
 }
 
 TEST_SUITE_END();
diff --git a/hilti/runtime/src/tests/string.cc b/hilti/runtime/src/tests/string.cc
@@ -66,12 +66,16 @@ TEST_CASE("to_string") {
     CHECK_EQ(to_string(std::string("abc")), "\"abc\"");
     CHECK_EQ(to_string(std::string_view("abc")), "\"abc\"");
     CHECK_EQ(to_string("abc"), "\"abc\"");
+    CHECK_EQ(to_string("\"\\"), "\"\\\"\\\\\"");
 }
 
 TEST_CASE("to_string_for_print") {
     CHECK_EQ(to_string_for_print(std::string("abc")), "abc");
     CHECK_EQ(to_string_for_print(std::string_view("abc")), "abc");
     CHECK_EQ(to_string_for_print("abc"), "abc");
+    CHECK_EQ(to_string_for_print(std::string("\\\"")), "\\\"");
+    CHECK_EQ(to_string_for_print(std::string_view("\\\"")), "\\\"");
+    CHECK_EQ(to_string_for_print("\\\""), "\\\"");
 }
 
 TEST_CASE("split") {

diff --git a/hilti/runtime/src/util.cc b/hilti/runtime/src/util.cc
@@ -294,8 +294,10 @@ std::string hilti::rt::escapeUTF8(std::string_view s, bitmask<render_style::UTF8
         if ( cp == '\\' ) {
             if ( (style & render_style::UTF8::NoEscapeHex) && (p + n) < e && *(p + n) == 'x' )
                 esc += "\\";
-            else
+            else if ( ! (style & render_style::UTF8::NoEscapeBackslash) )
                 esc += "\\\\";
+            else
+                esc += "\\";
         }
 
         else if ( cp == '"' && (style & render_style::UTF8::EscapeQuotes) )
@@ -346,7 +348,7 @@ std::string hilti::rt::escapeBytes(std::string_view s, bitmask<render_style::Byt
     std::string esc;
 
     while ( p < e ) {
-        if ( *p == '\\' )
+        if ( *p == '\\' && ! (style & render_style::Bytes::NoEscapeBackslash) )
             esc += "\\\\";
 
         else if ( *p == '"' && (style & render_style::Bytes::EscapeQuotes) )

diff --git a/tests/Baseline/spicy.rt.print-escaping/output b/tests/Baseline/spicy.rt.print-escaping/output
@@ -0,0 +1,4 @@
+### BTest baseline data generated by btest-diff. Do not edit. Use "btest -U/-u" to update. Requires BTest >= 0.63.
+\", \"
+\", \"
+5C22, 5C22
diff --git a/tests/spicy/rt/print-escaping.spicy b/tests/spicy/rt/print-escaping.spicy
@@ -0,0 +1,17 @@
+# @TEST-EXEC: spicyc -j -d %INPUT >output
+# @TEST-EXEC: btest-diff output
+#
+# @TEST-DOC: Check that we escape certain special characters in strings correctly when printing them.
+
+module test;
+
+import spicy;
+
+global x = b"\\\"";
+global y = "\\\"";
+
+print x, y;
+print "%s, %s" % (x, y);
+
+# for comparison
+print spicy::bytes_to_hexstring(x), spicy::bytes_to_hexstring(y.encode());