[PATCH] Spelling suggestions for misspelled preprocessor directives

David Malcolm dmalcolm@redhat.com
Thu Aug 18 14:59:00 GMT 2016


This patch allows the preprocessor to offer suggestions for misspelled
directives, taking us from e.g.:

test.c:5:2: error: invalid preprocessing directive #endfi
 #endfi
  ^~~~~

to:

test.c:5:2: error: invalid preprocessing directive #endfi; did you mean #endif?
 #endfi
  ^~~~~
  endif

I can self-approve all of it apart from the changes to c-family.

Successfully bootstrapped&regrtested on x86_64-pc-linux-gnu.

OK for trunk?

gcc/c-family/ChangeLog:
	* c-common.c: Include "spellcheck.h".
	(cb_get_suggestion): New function.
	* c-common.h (cb_get_suggestion): New decl.
	* c-lex.c (init_c_lex): Initialize cb->get_suggestion to
	cb_get_suggestion.

gcc/testsuite/ChangeLog:
	* gcc.dg/cpp/misspelled-directive-1.c: New testcase.
	* gcc.dg/cpp/misspelled-directive-2.c: New testcase.

libcpp/ChangeLog:
	* directives.c (directive_names): New array.
	(_cpp_handle_directive): Offer spelling suggestions for misspelled
	directives.
	* errors.c (cpp_diagnostic_at_richloc): New function.
	(cpp_error_at_richloc): New function.
	* include/cpplib.h (struct cpp_callbacks): Add field
	"get_suggestion".
	(cpp_error_at_richloc): New decl.
---
 gcc/c-family/c-common.c                           | 17 ++++++++++
 gcc/c-family/c-common.h                           |  5 +++
 gcc/c-family/c-lex.c                              |  1 +
 gcc/testsuite/gcc.dg/cpp/misspelled-directive-1.c | 12 +++++++
 gcc/testsuite/gcc.dg/cpp/misspelled-directive-2.c | 21 ++++++++++++
 libcpp/directives.c                               | 41 +++++++++++++++++++++--
 libcpp/errors.c                                   | 36 ++++++++++++++++++++
 libcpp/include/cpplib.h                           |  8 +++++
 8 files changed, 139 insertions(+), 2 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/cpp/misspelled-directive-1.c
 create mode 100644 gcc/testsuite/gcc.dg/cpp/misspelled-directive-2.c

diff --git a/gcc/c-family/c-common.c b/gcc/c-family/c-common.c
index 569f000..a8ef595 100644
--- a/gcc/c-family/c-common.c
+++ b/gcc/c-family/c-common.c
@@ -46,6 +46,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "opts.h"
 #include "gimplify.h"
 #include "substring-locations.h"
+#include "spellcheck.h"
 
 cpp_reader *parse_in;		/* Declared in c-pragma.h.  */
 
@@ -12934,6 +12935,22 @@ cb_get_source_date_epoch (cpp_reader *pfile ATTRIBUTE_UNUSED)
   return (time_t) epoch;
 }
 
+/* Callback for libcpp for offering spelling suggestions for misspelled
+   directives.  GOAL is an unrecognized string; CANDIDATES is a
+   NULL-terminated array of candidate strings.  Return the closest
+   match to GOAL within CANDIDATES, or NULL if none are good
+   suggestions.  */
+
+const char *
+cb_get_suggestion (cpp_reader *, const char *goal,
+		   const char * const *candidates)
+{
+  best_match<const char *, const char *> bm (goal);
+  while (*candidates)
+    bm.consider (*(candidates++));
+  return bm.get_best_meaningful_candidate ();
+}
+
 /* Check and possibly warn if two declarations have contradictory
    attributes, such as always_inline vs. noinline.  */
 
diff --git a/gcc/c-family/c-common.h b/gcc/c-family/c-common.h
index 61f9ced..a3da631 100644
--- a/gcc/c-family/c-common.h
+++ b/gcc/c-family/c-common.h
@@ -1110,6 +1110,11 @@ extern time_t cb_get_source_date_epoch (cpp_reader *pfile);
    __TIME__ can store.  */
 #define MAX_SOURCE_DATE_EPOCH HOST_WIDE_INT_C (253402300799)
 
+/* Callback for libcpp for offering spelling suggestions for misspelled
+   directives.  */
+extern const char *cb_get_suggestion (cpp_reader *, const char *,
+				      const char * const *);
+
 extern GTY(()) string_concat_db *g_string_concat_db;
 
 /* libcpp can calculate location information about a range of characters
diff --git a/gcc/c-family/c-lex.c b/gcc/c-family/c-lex.c
index 4c7e385..c904ee6 100644
--- a/gcc/c-family/c-lex.c
+++ b/gcc/c-family/c-lex.c
@@ -81,6 +81,7 @@ init_c_lex (void)
   cb->read_pch = c_common_read_pch;
   cb->has_attribute = c_common_has_attribute;
   cb->get_source_date_epoch = cb_get_source_date_epoch;
+  cb->get_suggestion = cb_get_suggestion;
 
   /* Set the debug callbacks if we can use them.  */
   if ((debug_info_level == DINFO_LEVEL_VERBOSE
diff --git a/gcc/testsuite/gcc.dg/cpp/misspelled-directive-1.c b/gcc/testsuite/gcc.dg/cpp/misspelled-directive-1.c
new file mode 100644
index 0000000..f79670a
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/cpp/misspelled-directive-1.c
@@ -0,0 +1,12 @@
+#ifndef SOME_GUARD /* { dg-error "unterminated" } */
+
+#if 1
+/* Typo here: "endfi" should have been "endif".  */
+#endfi /* { dg-error "invalid preprocessing directive #endfi; did you mean #endif?" } */
+
+int make_non_empty;
+
+/* Another transposition typo:  */
+#deifne FOO /* { dg-error "invalid preprocessing directive #deifne; did you mean #define?" } */ 
+
+#endif /* #ifndef SOME_GUARD */
diff --git a/gcc/testsuite/gcc.dg/cpp/misspelled-directive-2.c b/gcc/testsuite/gcc.dg/cpp/misspelled-directive-2.c
new file mode 100644
index 0000000..7ec5dee
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/cpp/misspelled-directive-2.c
@@ -0,0 +1,21 @@
+/* { dg-options "-fdiagnostics-show-caret" } */
+
+#endfi /* { dg-error "invalid preprocessing directive #endfi; did you mean #endif?" } */
+
+/* Verify that we offer fix-it hints.  */
+/* { dg-begin-multiline-output "" }
+ #endfi
+  ^~~~~
+  endif
+  { dg-end-multiline-output "" } */
+
+/* Test coverage for the case of an unrecognized directive where no suggestion
+   is offered.  */
+
+#this_does_not_match_anything /* { dg-error "invalid preprocessing directive #this_does_not_match_anything" } */
+/* { dg-begin-multiline-output "" }
+ #this_does_not_match_anything
+  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
+  { dg-end-multiline-output "" } */
+
+int make_non_empty;
diff --git a/libcpp/directives.c b/libcpp/directives.c
index 772b835..c0006a4 100644
--- a/libcpp/directives.c
+++ b/libcpp/directives.c
@@ -188,6 +188,16 @@ static const directive dtable[] =
 DIRECTIVE_TABLE
 };
 #undef D
+
+/* A NULL-terminated array of directive names for use
+   when suggesting corrections for misspelled directives.  */
+#define D(name, t, origin, flags) #name,
+static const char * const directive_names[] = {
+DIRECTIVE_TABLE
+  NULL
+};
+#undef D
+
 #undef DIRECTIVE_TABLE
 
 /* Wrapper struct directive for linemarkers.
@@ -498,8 +508,35 @@ _cpp_handle_directive (cpp_reader *pfile, int indented)
       if (CPP_OPTION (pfile, lang) == CLK_ASM)
 	skip = 0;
       else if (!pfile->state.skipping)
-	cpp_error (pfile, CPP_DL_ERROR, "invalid preprocessing directive #%s",
-		   cpp_token_as_text (pfile, dname));
+	{
+	  const char *unrecognized
+	    = (const char *)cpp_token_as_text (pfile, dname);
+	  const char *hint = NULL;
+
+	  /* Call back into gcc to get a spelling suggestion.  Ideally
+	     we'd just use best_match from gcc/spellcheck.h (and filter
+	     out the uncommon directives), but that requires moving it
+	     to a support library.  */
+	  if (pfile->cb.get_suggestion)
+	    hint = pfile->cb.get_suggestion (pfile, unrecognized,
+					     directive_names);
+
+	  if (hint)
+	    {
+	      rich_location richloc (pfile->line_table, dname->src_loc);
+	      source_range misspelled_token_range
+		= get_range_from_loc (pfile->line_table, dname->src_loc);
+	      richloc.add_fixit_replace (misspelled_token_range, hint);
+	      cpp_error_at_richloc (pfile, CPP_DL_ERROR, &richloc,
+				    "invalid preprocessing directive #%s;"
+				    " did you mean #%s?",
+				    unrecognized, hint);
+	    }
+	  else
+	    cpp_error (pfile, CPP_DL_ERROR,
+		       "invalid preprocessing directive #%s",
+		       unrecognized);
+	}
     }
 
   pfile->directive = dir;
diff --git a/libcpp/errors.c b/libcpp/errors.c
index f7d4112..3b0a0b4 100644
--- a/libcpp/errors.c
+++ b/libcpp/errors.c
@@ -31,6 +31,23 @@ along with this program; see the file COPYING3.  If not see
 
 ATTRIBUTE_FPTR_PRINTF(5,0)
 static bool
+cpp_diagnostic_at_richloc (cpp_reader * pfile, int level, int reason,
+			   rich_location *richloc,
+			   const char *msgid, va_list *ap)
+{
+  bool ret;
+
+  if (!pfile->cb.error)
+    abort ();
+  ret = pfile->cb.error (pfile, level, reason, richloc, _(msgid), ap);
+
+  return ret;
+}
+
+/* Print a diagnostic at the given location.  */
+
+ATTRIBUTE_FPTR_PRINTF(5,0)
+static bool
 cpp_diagnostic_at (cpp_reader * pfile, int level, int reason,
 		   source_location src_loc,
 		   const char *msgid, va_list *ap)
@@ -255,6 +272,25 @@ cpp_error_at (cpp_reader * pfile, int level, source_location src_loc,
   return ret;
 }
 
+/* As cpp_error, but use RICHLOC as the location of the error, without
+   a column override.  */
+
+bool
+cpp_error_at_richloc (cpp_reader * pfile, int level, rich_location *richloc,
+		      const char *msgid, ...)
+{
+  va_list ap;
+  bool ret;
+
+  va_start (ap, msgid);
+
+  ret = cpp_diagnostic_at_richloc (pfile, level, CPP_W_NONE, richloc,
+				   msgid, &ap);
+
+  va_end (ap);
+  return ret;
+}
+
 /* Print a warning or error, depending on the value of LEVEL.  Include
    information from errno.  */
 
diff --git a/libcpp/include/cpplib.h b/libcpp/include/cpplib.h
index 659686b..fae19d2 100644
--- a/libcpp/include/cpplib.h
+++ b/libcpp/include/cpplib.h
@@ -597,6 +597,9 @@ struct cpp_callbacks
 
   /* Callback to parse SOURCE_DATE_EPOCH from environment.  */
   time_t (*get_source_date_epoch) (cpp_reader *);
+
+  /* Callback for providing suggestions for misspelled directives.  */
+  const char *(*get_suggestion) (cpp_reader *, const char *, const char * const *);
 };
 
 #ifdef VMS
@@ -1066,6 +1069,11 @@ extern bool cpp_error_at (cpp_reader * pfile, int level,
 			  source_location src_loc, const char *msgid, ...)
   ATTRIBUTE_PRINTF_4;
 
+extern bool cpp_error_at_richloc (cpp_reader * pfile, int level,
+				  rich_location *richloc, const char *msgid,
+				  ...)
+  ATTRIBUTE_PRINTF_4;
+
 /* In lex.c */
 extern int cpp_ideq (const cpp_token *, const char *);
 extern void cpp_output_line (cpp_reader *, FILE *);
-- 
1.8.5.3



More information about the Gcc-patches mailing list