Fix locations within raw strings

Whilst investigating PR preprocessor/78324 I noticed that the
substring location code currently doesn't handle raw strings
correctly, by not skipping the 'R', opening quote, delimiter
and opening parenthesis.

For example, an attempt to underline chars 4-7 with caret at 6 of
this raw string yields this erroneous output:
   __emit_string_literal_range (R"foo(0123456789)foo",
                                    ~~^~

With the patch, the correct range/caret is printed:

   __emit_string_literal_range (R"foo(0123456789)foo",
                                          ~~^~

gcc/ChangeLog:
	* input.c (selftest::test_lexer_string_locations_long_line): New
	function.
	(selftest::test_lexer_string_locations_raw_string_multiline): New
	function.
	(selftest::input_c_tests): Call the new functions, via
	for_each_line_table_case.

gcc/testsuite/ChangeLog:
	* gcc.dg/plugin/diagnostic-test-string-literals-1.c
	(test_raw_string_one_liner): New function.
	(test_raw_string_multiline): New function.

libcpp/ChangeLog:
	* charset.c (cpp_interpret_string_1): Skip locations from
	loc_reader when advancing 'p' when handling raw strings.

From-SVN: r242552

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index f69f911e4811327014eb9f92be4b180229a759d9..1bb29c0ab03a535d396c1bbd6e47655c031ad676 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,12 @@
+2016-11-17  David Malcolm  <dmalcolm@redhat.com>
+
+       * input.c (selftest::test_lexer_string_locations_long_line): New
+       function.
+       (selftest::test_lexer_string_locations_raw_string_multiline): New
+       function.
+       (selftest::input_c_tests): Call the new functions, via
+       for_each_line_table_case.
+
 2016-11-17  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 
        * config/aarch64/aarch64.md (mov<mode>): Call

diff --git a/gcc/input.c b/gcc/input.c
index c2042e8e211b1f2f591663233b6c20d03a3d4238..728f4dd65d7716642d8f474140e49abb6e1fa848 100644 (file)
--- a/gcc/input.c
+++ b/gcc/input.c
@@ -3156,6 +3156,78 @@ test_lexer_string_locations_long_line (const line_table_case &case_)
                          i, 2, 7 + i, 7 + i);
 }
 
+/* Test of locations within a raw string that doesn't contain a newline.  */
+
+static void
+test_lexer_string_locations_raw_string_one_line (const line_table_case &case_)
+{
+  /* .....................00.0000000111111111122.
+     .....................12.3456789012345678901.  */
+  const char *content = ("R\"foo(0123456789)foo\"\n");
+  lexer_test test (case_, content, NULL);
+
+  /* Verify that we get the expected token back.  */
+  const cpp_token *tok = test.get_token ();
+  ASSERT_EQ (tok->type, CPP_STRING);
+
+  /* Verify that cpp_interpret_string works.  */
+  cpp_string dst_string;
+  const enum cpp_ttype type = CPP_STRING;
+  bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1,
+                                     &dst_string, type);
+  ASSERT_TRUE (result);
+  ASSERT_STREQ ("0123456789", (const char *)dst_string.text);
+  free (const_cast <unsigned char *> (dst_string.text));
+
+  if (!should_have_column_data_p (line_table->highest_location))
+    return;
+
+  /* 0-9, plus the nil terminator.  */
+  ASSERT_NUM_SUBSTRING_RANGES (test, tok->src_loc, CPP_STRING, 11);
+  for (int i = 0; i < 11; i++)
+    ASSERT_CHAR_AT_RANGE (test, tok->src_loc, CPP_STRING,
+                         i, 1, 7 + i, 7 + i);
+}
+
+/* Test of locations within a raw string that contains a newline.  */
+
+static void
+test_lexer_string_locations_raw_string_multiline (const line_table_case &case_)
+{
+  /* .....................00.0000.
+     .....................12.3456.  */
+  const char *content = ("R\"foo(\n"
+  /* .....................00000.
+     .....................12345.  */
+                        "hello\n"
+                        "world\n"
+  /* .....................00000.
+     .....................12345.  */
+                        ")foo\"\n");
+  lexer_test test (case_, content, NULL);
+
+  /* Verify that we get the expected token back.  */
+  const cpp_token *tok = test.get_token ();
+  ASSERT_EQ (tok->type, CPP_STRING);
+
+  /* Verify that cpp_interpret_string works.  */
+  cpp_string dst_string;
+  const enum cpp_ttype type = CPP_STRING;
+  bool result = cpp_interpret_string (test.m_parser, &tok->val.str, 1,
+                                     &dst_string, type);
+  ASSERT_TRUE (result);
+  ASSERT_STREQ ("\nhello\nworld\n", (const char *)dst_string.text);
+  free (const_cast <unsigned char *> (dst_string.text));
+
+  if (!should_have_column_data_p (line_table->highest_location))
+    return;
+
+  /* Currently we don't support locations within raw strings that
+     contain newlines.  */
+  ASSERT_HAS_NO_SUBSTRING_RANGES (test, tok->src_loc, tok->type,
+                                 "range endpoints are on different lines");
+}
+
 /* Test of lexing char constants.  */
 
 static void
@@ -3297,6 +3369,8 @@ input_c_tests ()
   for_each_line_table_case (test_lexer_string_locations_stringified_macro_argument);
   for_each_line_table_case (test_lexer_string_locations_non_string);
   for_each_line_table_case (test_lexer_string_locations_long_line);
+  for_each_line_table_case (test_lexer_string_locations_raw_string_one_line);
+  for_each_line_table_case (test_lexer_string_locations_raw_string_multiline);
   for_each_line_table_case (test_lexer_char_constants);
 
   test_reading_source_line ();

diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 77fa771ab179a5183a6fb49dd2e20acbddde2a3c..fb0bd45f431184bc7e8156bde4386d51b54c1185 100644 (file)
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,9 @@
+2016-11-17  David Malcolm  <dmalcolm@redhat.com>
+
+       * gcc.dg/plugin/diagnostic-test-string-literals-1.c
+       (test_raw_string_one_liner): New function.
+       (test_raw_string_multiline): New function.
+
 2016-11-17  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>
 
        * gcc.target/aarch64/store_repeating_constant_1.c: New test.

diff --git a/gcc/testsuite/gcc.dg/plugin/diagnostic-test-string-literals-1.c b/gcc/testsuite/gcc.dg/plugin/diagnostic-test-string-literals-1.c
index 3d5ff6d9414fdfb2a8dc138016cad0b13abce000..3e449360f756ed4f7c2031564231146fb7360cff 100644 (file)
--- a/gcc/testsuite/gcc.dg/plugin/diagnostic-test-string-literals-1.c
+++ b/gcc/testsuite/gcc.dg/plugin/diagnostic-test-string-literals-1.c
@@ -193,6 +193,39 @@ test_L (void)
    { dg-end-multiline-output "" } */
 }
 
+void
+test_raw_string_one_liner (void)
+{
+  /* Digits 0-9.  */
+  __emit_string_literal_range (R"foo(0123456789)foo", /* { dg-warning "range" } */
+                              6, 4, 7);
+/* { dg-begin-multiline-output "" }
+   __emit_string_literal_range (R"foo(0123456789)foo",
+                                          ~~^~
+   { dg-end-multiline-output "" } */
+}
+
+void
+test_raw_string_multiline (void)
+{
+  __emit_string_literal_range (R"foo(
+hello
+world
+)foo",
+                              6, 4, 7);
+  /* { dg-error "unable to read substring location: range endpoints are on different lines" "" { target *-*-* } .-5 } */
+  /* { dg-begin-multiline-output "" }
+   __emit_string_literal_range (R"foo(
+                                ^~~~~~
+ hello
+ ~~~~~                           
+ world
+ ~~~~~                           
+ )foo",
+ ~~~~~                           
+   { dg-end-multiline-output "" } */
+}
+
 void
 test_macro (void)
 {

diff --git a/libcpp/ChangeLog b/libcpp/ChangeLog
index 7690b5deba9a548d5af35490608b5892c3f56de7..6b928078636bc24cc978816467720e91ad47263d 100644 (file)
--- a/libcpp/ChangeLog
+++ b/libcpp/ChangeLog
@@ -1,3 +1,8 @@
+2016-11-17  David Malcolm  <dmalcolm@redhat.com>
+
+       * charset.c (cpp_interpret_string_1): Skip locations from
+       loc_reader when advancing 'p' when handling raw strings.
+
 2016-11-16  Jakub Jelinek  <jakub@redhat.com>
 
        PR bootstrap/72823

diff --git a/libcpp/charset.c b/libcpp/charset.c
index e77270a18eb957e29f4a5e391a4969f1a2fb0ce0..9cd1e10ae89ac7407d9c5d2c7aefdc61b4e8d997 100644 (file)
--- a/libcpp/charset.c
+++ b/libcpp/charset.c
@@ -1564,10 +1564,21 @@ cpp_interpret_string_1 (cpp_reader *pfile, const cpp_string *from, size_t count,
 
          /* Skip over 'R"'.  */
          p += 2;
+         if (loc_reader)
+           {
+             loc_reader->get_next ();
+             loc_reader->get_next ();
+           }
          prefix = p;
          while (*p != '(')
-           p++;
+           {
+             p++;
+             if (loc_reader)
+               loc_reader->get_next ();
+           }
          p++;
+         if (loc_reader)
+           loc_reader->get_next ();
          limit = from[i].text + from[i].len;
          if (limit >= p + (p - prefix) + 1)
            limit -= (p - prefix) + 1;