Re: preprocessor/10613: Please make -Wtrigraphs the default

[Neil Booth <neil@daikokuya.co.uk>]

The following reply was made to PR preprocessor/10613; it has been noted by GNATS.

From: Neil Booth <neil@daikokuya.co.uk>
To: Hallvard B Furuseth <h.b.furuseth@usit.uio.no>
Cc: gcc-gnats@gcc.gnu.org, gcc-patches@gcc.gnu.org,
	Zack Weinberg <zack@codesourcery.com>
Subject: Re: preprocessor/10613: Please make -Wtrigraphs the default
Date: Sun, 4 May 2003 21:00:17 +0100

 This is a patch for preprocessor/10613 and preprocessor/10614.  We warn
 about trigraphs, unless an explicit -trigraphs (or -std implying
 -trigraphs) is given.  -Wall still implies -Wtrigraphs.
 
 Further, we want about trigraphs that imply an escaped newline sequence
 in a comment, but no other trigraphs in a comment are warned about.
 This is a good enough approximation to the "warn about a trigraph iff
 it changes the meaning of the program" rule.
 
 Tests included.  Tested on NetBSD-x86.
 
 Neil.
 
 	* cppinit.c (cpp_create_reader, post_options): Warn about
 	trigraphs unless explicity set or -trigraphs.
 	* cpplex.c (warn_in_comment): New.
 	(_cpp_process_line_notes): Better handling of -Wtrigraphs.
 	(_cpp_skip_block_comment): Add call to _cpp_process_line_notes.
 	* doc/cppopts.texi, doc/cpp.texi: Update.
 testsuite:
 	* gcc.dg/cpp/Wtrigraphs.c: Update.
 	* gcc.dg/cpp/Wtrigraphs-2.c: New tests.
 
 Index: cppinit.c
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/cppinit.c,v
 retrieving revision 1.276
 diff -u -p -r1.276 cppinit.c
 --- cppinit.c	19 Apr 2003 00:22:48 -0000	1.276
 +++ cppinit.c	4 May 2003 19:56:49 -0000
 @@ -149,6 +149,7 @@ cpp_create_reader (lang, table)
    CPP_OPTION (pfile, show_column) = 1;
    CPP_OPTION (pfile, tabstop) = 8;
    CPP_OPTION (pfile, operator_names) = 1;
 +  CPP_OPTION (pfile, warn_trigraphs) = 2;
    CPP_OPTION (pfile, warn_endif_labels) = 1;
    CPP_OPTION (pfile, warn_deprecated) = 1;
    CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
 @@ -553,6 +554,9 @@ post_options (pfile)
        pfile->state.prevent_expansion = 1;
        CPP_OPTION (pfile, traditional) = 0;
      }
 +
 +  if (CPP_OPTION (pfile, warn_trigraphs) == 2)
 +    CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
  
    if (CPP_OPTION (pfile, traditional))
      {
 Index: cpplex.c
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/cpplex.c,v
 retrieving revision 1.233
 diff -u -p -r1.233 cpplex.c
 --- cpplex.c	26 Apr 2003 21:03:51 -0000	1.233
 +++ cpplex.c	4 May 2003 19:56:49 -0000
 @@ -63,6 +63,7 @@ static void save_comment PARAMS ((cpp_re
  				  cppchar_t));
  static void create_literal PARAMS ((cpp_reader *, cpp_token *, const uchar *,
  				    unsigned int, enum cpp_ttype));
 +static bool warn_in_comment PARAMS ((cpp_reader *, _cpp_line_note *));
  static int name_p PARAMS ((cpp_reader *, const cpp_string *));
  static cppchar_t maybe_read_ucn PARAMS ((cpp_reader *, const uchar **));
  static tokenrun *next_tokenrun PARAMS ((tokenrun *));
 @@ -180,6 +181,36 @@ _cpp_clean_line (pfile)
    buffer->next_line = s + 1;
  }
  
 +/* Return true if the trigraph indicated by NOTE should be warned
 +   about in a comment.  */
 +static bool
 +warn_in_comment (pfile, note)
 +     cpp_reader *pfile;
 +     _cpp_line_note *note;
 +{
 +  const uchar *p;
 +
 +  /* Within comments we don't warn about trigraphs, unless the
 +     trigraph forms an escaped newline, as that may change
 +     behaviour.  */
 +  if (note->type != '/')
 +    return false;
 +
 +  /* If -trigraphs, then this was an escaped newline iff the next note
 +     is coincident.  */
 +  if (CPP_OPTION (pfile, trigraphs))
 +    return note[1].pos == note->pos;
 +
 +  /* Otherwise, see if this forms an escaped newline.  */
 +  p = note->pos + 3;
 +  while (is_nvspace (*p))
 +    p++;
 +
 +  /* There might have been escaped newlines between the trigraph and the
 +     newline we found.  Hence the position test.  */
 +  return (*p == '\n' && p < note[1].pos);
 +}
 +
  /* Process the notes created by add_line_note as far as the current
     location.  */
  void
 @@ -219,7 +250,8 @@ _cpp_process_line_notes (pfile, in_comme
  	}
        else if (_cpp_trigraph_map[note->type])
  	{
 -	  if (!in_comment && CPP_OPTION (pfile, warn_trigraphs))
 +	  if (CPP_OPTION (pfile, warn_trigraphs)
 +	      && (!in_comment || warn_in_comment (pfile, note)))
  	    {
  	      if (CPP_OPTION (pfile, trigraphs))
  		cpp_error_with_line (pfile, DL_WARNING, pfile->line, col,
 @@ -284,6 +316,7 @@ _cpp_skip_block_comment (pfile)
  	}
      }
  
 +  _cpp_process_line_notes (pfile, true);
    return false;
  }
  
 Index: testsuite/gcc.dg/cpp/Wtrigraphs-2.c
 ===================================================================
 RCS file: testsuite/gcc.dg/cpp/Wtrigraphs-2.c
 diff -N testsuite/gcc.dg/cpp/Wtrigraphs-2.c
 --- /dev/null	1 Jan 1970 00:00:00 -0000
 +++ testsuite/gcc.dg/cpp/Wtrigraphs-2.c	4 May 2003 19:56:49 -0000
 @@ -0,0 +1,20 @@
 +/* { dg-do preprocess } */
 +/* { dg-options "-std=c99 -Wtrigraphs -fno-show-column" } */
 +
 +/* Test we don't double warn for trigraphs immediately after preceding
 +   text.  Source Neil Booth.  4 May 2003.  */
 +
 +/* { dg-bogus "ignored" } Test ??< ??= a few ??/ random things in
 +   { dg-warning "converted" } some ??/
 +   { dg-bogus "ignored" } ??< comments.  */
 +
 +// { dg-bogus "ignored" } More ??/ comment ??> tests.
 +
 +// { dg-warning "converted" } Another ??/
 +   Test
 +
 +// { dg-warning "converted" } And another with space after ??/  
 +   the escape
 +
 +// { dg-bogus "ignored" } A tricky one ??/\
 +
 Index: testsuite/gcc.dg/cpp/Wtrigraphs.c
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/testsuite/gcc.dg/cpp/Wtrigraphs.c,v
 retrieving revision 1.1
 diff -u -p -r1.1 Wtrigraphs.c
 --- testsuite/gcc.dg/cpp/Wtrigraphs.c	22 Nov 2000 20:37:44 -0000	1.1
 +++ testsuite/gcc.dg/cpp/Wtrigraphs.c	4 May 2003 19:56:49 -0000
 @@ -1,5 +1,5 @@
  /* { dg-do preprocess } */
 -/* { dg-options "-Wtrigraphs -fno-show-column" } */
 +/* { dg-options "-std=gnu99 -Wtrigraphs -fno-show-column" } */
  
  /* Test we don't double warn for trigraphs immediately after preceding
     text.  Source Neil Booth.  22 Nov 2000.  */
 @@ -7,3 +7,21 @@
  abcdef??<			/* { dg-warning "ignored" } */
  123456??>			/* { dg-warning "ignored" } */
  +??=				/* { dg-warning "ignored" } */
 +
 +/* Test we warn of escaped newlines only in comments.  Source Neil
 +   Booth.  4 May 2003.  */
 +
 +/* { dg-bogus "ignored" } Test ??< ??= a few ??/ random things in
 +   { dg-warning "ignored" } some ??/
 +   { dg-bogus "ignored" } ??< comments.  */
 +
 +// { dg-bogus "ignored" } More ??/ comment ??> tests.
 +
 +// { dg-warning "ignored" } Another ??/
 +   Test
 +
 +// { dg-warning "ignored" } And another with space after ??/  
 +   the escape
 +
 +// { dg-bogus "ignored" } A tricky one ??/\
 +
 Index: doc/cpp.texi
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/doc/cpp.texi,v
 retrieving revision 1.54
 diff -u -p -r1.54 cpp.texi
 --- doc/cpp.texi	12 Mar 2003 20:30:05 -0000	1.54
 +++ doc/cpp.texi	4 May 2003 19:56:51 -0000
 @@ -293,16 +293,18 @@ obsolete systems that lack some of C's p
  example, @samp{??/} stands for @samp{\}, so @t{'??/n'} is a character
  constant for a newline.
  
 -Trigraphs are not popular and many compilers implement them incorrectly.
 -Portable code should not rely on trigraphs being either converted or
 -ignored.  If you use the @option{-Wall} or @option{-Wtrigraphs} options,
 -GCC will warn you when a trigraph would change the meaning of your
 -program if it were converted.
 +Trigraphs are not popular and many compilers implement them
 +incorrectly.  Portable code should not rely on trigraphs being either
 +converted or ignored.  With @option{-Wtrigraphs} GCC will warn you
 +when a trigraph may change the meaning of your program if it were
 +converted.  @xref{Wtrigraphs}.
  
 -In a string constant, you can prevent a sequence of question marks from
 -being confused with a trigraph by inserting a backslash between the
 -question marks.  @t{"(??\?)"} is the string @samp{(???)}, not
 -@samp{(?]}.  Traditional C compilers do not recognize this idiom.
 +In a string constant, you can prevent a sequence of question marks
 +from being confused with a trigraph by inserting a backslash between
 +the question marks, or by separating the string literal at the
 +trigraph and making use of string literal concatenation.  @t{"(??\?)"}
 +is the string @samp{(???)}, not @samp{(?]}.  Traditional C compilers
 +do not recognize these idioms.
  
  The nine trigraphs and their replacements are
  
 Index: doc/cppopts.texi
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/doc/cppopts.texi,v
 retrieving revision 1.18
 diff -u -p -r1.18 cppopts.texi
 --- doc/cppopts.texi	16 Mar 2003 10:37:06 -0000	1.18
 +++ doc/cppopts.texi	4 May 2003 19:56:51 -0000
 @@ -69,10 +69,12 @@ use @option{-o} to specify the output fi
  
  @item -Wall
  @opindex Wall
 -Turns on all optional warnings which are desirable for normal code.  At
 -present this is @option{-Wcomment} and @option{-Wtrigraphs}.  Note that
 -many of the preprocessor's warnings are on by default and have no
 -options to control them.
 +Turns on all optional warnings which are desirable for normal code.
 +At present this is @option{-Wcomment}, @option{-Wtrigraphs},
 +@option{-Wmultichar} and a warning about integer promotion causing a
 +change of sign in @code{#if} expressions.  Note that many of the
 +preprocessor's warnings are on by default and have no options to
 +control them.
  
  @item -Wcomment
  @itemx -Wcomments
 @@ -84,10 +86,13 @@ comment, or whenever a backslash-newline
  
  @item -Wtrigraphs
  @opindex Wtrigraphs
 -Warn if any trigraphs are encountered.  This option used to take effect
 -only if @option{-trigraphs} was also specified, but now works
 -independently.  Warnings are not given for trigraphs within comments, as
 -they do not affect the meaning of the program.
 +@anchor{Wtrigraphs}
 +Warn if any trigraphs that may change the meaning of a program are
 +encountered.  This option is in effect unless trigraphs are turned on,
 +and is implied by @option{-Wall}.  With the exception of a trigraph
 +that would form an escaped newline, warnings are not given for
 +trigraphs within comments as they do not affect the meaning of the
 +program.
  
  @item -Wtraditional
  @opindex Wtraditional