Improve the c-format guessing heuristic.

21 files changed, 99 insertions, 21 deletions

diff --git a/gettext-tools/src/ChangeLog b/gettext-tools/src/ChangeLog
index 56de0e8..588f186 100644
--- a/gettext-tools/src/ChangeLog
+++ b/gettext-tools/src/ChangeLog
@@ -1,5 +1,37 @@
 2006-02-12  Bruno Haible  <bruno@clisp.org>
 
+	Reduce the number of false positives produced by the c-format guessing
+	heuristics.
+	* format.h (struct formatstring_parser): Add 'is_unlikely_intentional'
+	field.
+	* format-awk.c (formatstring_awk): Update.
+	* format-c.c (struct spec): Add 'unlikely_unintentional' field.
+	(format_parse): Set unlikely_unintentional to true when encountering
+	a "%...%" directive other than "%%".
+	(format_is_unlikely_intentional): New function.
+	(formatstring_c, formatstring_objc): Use it.
+	* format-csharp.c (formatstring_csharp): Update.
+	* format-elisp.c (formatstring_elisp): Update.
+	* format-gcc-internal.c (formatstring_gcc_internal): Update.
+	* format-java.c (formatstring_java): Update.
+	* format-librep.c (formatstring_librep): Update.
+	* format-lisp.c (formatstring_lisp): Update.
+	* format-pascal.c (formatstring_pascal): Update.
+	* format-perl.c (formatstring_perl): Update.
+	* format-perl-brace.c (formatstring_perl_brace): Update.
+	* format-php.c (format_php): Update.
+	* format-python.c (formatstring_python): Update.
+	* format-qt.c (formatstring_qt): Update.
+	* format-scheme.c (formatstring_scheme): Update
+	* format-sh.c (formatstring_sh): Update.
+	* format-tcl.c (formatstring_tcl): Update.
+	* format-ycp.c (formatstring_ycp, formatstring_smalltalk): Update.
+	* xgettext.c (remember_a_message, remember_a_message_plural): Don't
+	mark the message as c-format if the validity as c-format looks
+	unintentional.
+
+2006-02-12  Bruno Haible  <bruno@clisp.org>
+
 	* msgattrib.c (main): Update year in --version output.
 	* msgcat.c (main): Likewise.
 	* msgcmp.c (main): Likewise.
diff --git a/gettext-tools/src/format-awk.c b/gettext-tools/src/format-awk.c
index b290f40..d540da6 100644
--- a/gettext-tools/src/format-awk.c
+++ b/gettext-tools/src/format-awk.c
@@ -1,5 +1,5 @@
 /* awk format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2002.
 
    This program is free software; you can redistribute it and/or modify
@@ -537,6 +537,7 @@ struct formatstring_parser formatstring_awk =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-c.c b/gettext-tools/src/format-c.c
index be12d24..e31bebf 100644
--- a/gettext-tools/src/format-c.c
+++ b/gettext-tools/src/format-c.c
@@ -1,5 +1,5 @@
 /* C format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -152,6 +152,7 @@ struct spec
   unsigned int unnumbered_arg_count;
   unsigned int allocated;
   struct unnumbered_arg *unnumbered;
+  bool unlikely_intentional;
   unsigned int sysdep_directives_count;
   const char **sysdep_directives;
 };
@@ -190,6 +191,7 @@ format_parse (const char *format, bool translated, bool objc_extensions,
   spec.allocated = 0;
   numbered = NULL;
   spec.unnumbered = NULL;
+  spec.unlikely_intentional = false;
   spec.sysdep_directives_count = 0;
   spec.sysdep_directives = NULL;
 
@@ -608,6 +610,13 @@ format_parse (const char *format, bool translated, bool objc_extensions,
 	    switch (*format)
 	      {
 	      case '%':
+		/* Programmers writing _("%2%") most often will not want to
+		   use this string as a c-format string, but rather as a
+		   literal or as a different kind of format string.  */
+		if (format[-1] != '%')
+		  spec.unlikely_intentional = true;
+		type = FAT_NONE;
+		break;
 	      case 'm': /* glibc extension */
 		type = FAT_NONE;
 		break;
@@ -820,6 +829,14 @@ format_free (void *descr)
   free (spec);
 }
 
+static bool
+format_is_unlikely_intentional (void *descr)
+{
+  struct spec *spec = (struct spec *) descr;
+
+  return spec->unlikely_intentional;
+}
+
 static int
 format_get_number_of_directives (void *descr)
 {
@@ -867,6 +884,7 @@ struct formatstring_parser formatstring_c =
   format_c_parse,
   format_free,
   format_get_number_of_directives,
+  format_is_unlikely_intentional,
   format_check
 };
 
@@ -876,6 +894,7 @@ struct formatstring_parser formatstring_objc =
   format_objc_parse,
   format_free,
   format_get_number_of_directives,
+  format_is_unlikely_intentional,
   format_check
 };
 
diff --git a/gettext-tools/src/format-csharp.c b/gettext-tools/src/format-csharp.c
index 8e71fb1..966ede2 100644
--- a/gettext-tools/src/format-csharp.c
+++ b/gettext-tools/src/format-csharp.c
@@ -1,5 +1,5 @@
 /* C# format strings.
-   Copyright (C) 2003-2004 Free Software Foundation, Inc.
+   Copyright (C) 2003-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2003.
 
    This program is free software; you can redistribute it and/or modify
@@ -206,6 +206,7 @@ struct formatstring_parser formatstring_csharp =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-elisp.c b/gettext-tools/src/format-elisp.c
index 4f5005c..17c2e62 100644
--- a/gettext-tools/src/format-elisp.c
+++ b/gettext-tools/src/format-elisp.c
@@ -1,5 +1,5 @@
 /* Emacs Lisp format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2002.
 
    This program is free software; you can redistribute it and/or modify
@@ -385,6 +385,7 @@ struct formatstring_parser formatstring_elisp =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-gcc-internal.c b/gettext-tools/src/format-gcc-internal.c
index 7d3be08..43713ae 100644
--- a/gettext-tools/src/format-gcc-internal.c
+++ b/gettext-tools/src/format-gcc-internal.c
@@ -1,5 +1,5 @@
 /* GCC internal format strings.
-   Copyright (C) 2003-2005 Free Software Foundation, Inc.
+   Copyright (C) 2003-2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2003.
 
    This program is free software; you can redistribute it and/or modify
@@ -357,6 +357,7 @@ struct formatstring_parser formatstring_gcc_internal =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-java.c b/gettext-tools/src/format-java.c
index 4574574..b7f6de1 100644
--- a/gettext-tools/src/format-java.c
+++ b/gettext-tools/src/format-java.c
@@ -1,5 +1,5 @@
 /* Java format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -777,6 +777,7 @@ struct formatstring_parser formatstring_java =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-librep.c b/gettext-tools/src/format-librep.c
index bc51379..e64b92c 100644
--- a/gettext-tools/src/format-librep.c
+++ b/gettext-tools/src/format-librep.c
@@ -1,5 +1,5 @@
 /* librep format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -349,6 +349,7 @@ struct formatstring_parser formatstring_librep =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-lisp.c b/gettext-tools/src/format-lisp.c
index 25c39c8..4b746b2 100644
--- a/gettext-tools/src/format-lisp.c
+++ b/gettext-tools/src/format-lisp.c
@@ -1,5 +1,5 @@
 /* Lisp format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -3370,6 +3370,7 @@ struct formatstring_parser formatstring_lisp =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-pascal.c b/gettext-tools/src/format-pascal.c
index cb4b88f..c863e30 100644
--- a/gettext-tools/src/format-pascal.c
+++ b/gettext-tools/src/format-pascal.c
@@ -1,5 +1,5 @@
 /* Object Pascal format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -438,6 +438,7 @@ struct formatstring_parser formatstring_pascal =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-perl-brace.c b/gettext-tools/src/format-perl-brace.c
index 9743754..1d3877b 100644
--- a/gettext-tools/src/format-perl-brace.c
+++ b/gettext-tools/src/format-perl-brace.c
@@ -1,5 +1,5 @@
 /* Perl brace format strings.
-   Copyright (C) 2004 Free Software Foundation, Inc.
+   Copyright (C) 2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2003.
 
    This program is free software; you can redistribute it and/or modify
@@ -213,6 +213,7 @@ struct formatstring_parser formatstring_perl_brace =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-perl.c b/gettext-tools/src/format-perl.c
index b1d15d3..20d6632 100644
--- a/gettext-tools/src/format-perl.c
+++ b/gettext-tools/src/format-perl.c
@@ -1,5 +1,5 @@
 /* Perl format strings.
-   Copyright (C) 2004 Free Software Foundation, Inc.
+   Copyright (C) 2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2003.
 
    This program is free software; you can redistribute it and/or modify
@@ -615,6 +615,7 @@ struct formatstring_parser formatstring_perl =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-php.c b/gettext-tools/src/format-php.c
index c32f3f5..434808e 100644
--- a/gettext-tools/src/format-php.c
+++ b/gettext-tools/src/format-php.c
@@ -1,5 +1,5 @@
 /* PHP format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2002.
 
    This program is free software; you can redistribute it and/or modify
@@ -385,6 +385,7 @@ struct formatstring_parser formatstring_php =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-python.c b/gettext-tools/src/format-python.c
index 610cbf0..b86de59 100644
--- a/gettext-tools/src/format-python.c
+++ b/gettext-tools/src/format-python.c
@@ -1,5 +1,5 @@
 /* Python format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -507,6 +507,7 @@ struct formatstring_parser formatstring_python =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-qt.c b/gettext-tools/src/format-qt.c
index 7f395be..f4da230 100644
--- a/gettext-tools/src/format-qt.c
+++ b/gettext-tools/src/format-qt.c
@@ -1,5 +1,5 @@
 /* Qt format strings.
-   Copyright (C) 2003-2004 Free Software Foundation, Inc.
+   Copyright (C) 2003-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2003.
 
    This program is free software; you can redistribute it and/or modify
@@ -145,6 +145,7 @@ struct formatstring_parser formatstring_qt =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-scheme.c b/gettext-tools/src/format-scheme.c
index b71d65b..7497a92 100644
--- a/gettext-tools/src/format-scheme.c
+++ b/gettext-tools/src/format-scheme.c
@@ -1,5 +1,5 @@
 /* Scheme format strings.
-   Copyright (C) 2001-2005 Free Software Foundation, Inc.
+   Copyright (C) 2001-2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -3308,6 +3308,7 @@ struct formatstring_parser formatstring_scheme =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-sh.c b/gettext-tools/src/format-sh.c
index d09f0be..176177f 100644
--- a/gettext-tools/src/format-sh.c
+++ b/gettext-tools/src/format-sh.c
@@ -1,5 +1,5 @@
 /* Shell format strings.
-   Copyright (C) 2003-2004 Free Software Foundation, Inc.
+   Copyright (C) 2003-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <bruno@clisp.org>, 2003.
 
    This program is free software; you can redistribute it and/or modify
@@ -314,6 +314,7 @@ struct formatstring_parser formatstring_sh =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-tcl.c b/gettext-tools/src/format-tcl.c
index 14e234c..8002832 100644
--- a/gettext-tools/src/format-tcl.c
+++ b/gettext-tools/src/format-tcl.c
@@ -1,5 +1,5 @@
 /* Tcl format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2002.
 
    This program is free software; you can redistribute it and/or modify
@@ -424,6 +424,7 @@ struct formatstring_parser formatstring_tcl =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format-ycp.c b/gettext-tools/src/format-ycp.c
index 62f4fab..cf9797d 100644
--- a/gettext-tools/src/format-ycp.c
+++ b/gettext-tools/src/format-ycp.c
@@ -1,5 +1,5 @@
 /* YCP and Smalltalk format strings.
-   Copyright (C) 2001-2004 Free Software Foundation, Inc.
+   Copyright (C) 2001-2004, 2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -146,6 +146,7 @@ struct formatstring_parser formatstring_ycp =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
@@ -155,6 +156,7 @@ struct formatstring_parser formatstring_smalltalk =
   format_parse,
   format_free,
   format_get_number_of_directives,
+  NULL,
   format_check
 };
 
diff --git a/gettext-tools/src/format.h b/gettext-tools/src/format.h
index a62a5e0..27bb4af 100644
--- a/gettext-tools/src/format.h
+++ b/gettext-tools/src/format.h
@@ -1,5 +1,5 @@
 /* Format strings.
-   Copyright (C) 2001-2005 Free Software Foundation, Inc.
+   Copyright (C) 2001-2006 Free Software Foundation, Inc.
    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
 
    This program is free software; you can redistribute it and/or modify
@@ -58,6 +58,11 @@ struct formatstring_parser
      A string that can be output literally has 0 format directives.  */
   int (*get_number_of_directives) (void *descr);
 
+  /* Return true if the format string, although valid, contains directives that
+     make it appear unlikely that the string was meant as a format string.
+     A NULL function is equivalent to a function that always returns false.  */
+  bool (*is_unlikely_intentional) (void *descr);
+
   /* Verify that the argument types/names in msgid_descr and those in
      msgstr_descr are the same (if equality=true), or (if equality=false)
      that those of msgid_descr extend those of msgstr_descr (i.e.
diff --git a/gettext-tools/src/xgettext.c b/gettext-tools/src/xgettext.c
index c0fa97c..9da8476 100644
--- a/gettext-tools/src/xgettext.c
+++ b/gettext-tools/src/xgettext.c
@@ -2079,7 +2079,9 @@ meta information, not the empty string.\n")));
 		 "xgettext: no-c-format" anywhere where a translator wishes
 		 to use a percent sign.  So, the msgfmt checking will not be
 		 perfect.  Oh well.  */
-	      if (parser->get_number_of_directives (descr) > 0)
+	      if (parser->get_number_of_directives (descr) > 0
+		  && !(parser->is_unlikely_intentional != NULL
+		       && parser->is_unlikely_intentional (descr)))
 		is_format[i] = possible;
 
 	      parser->free (descr);
@@ -2181,7 +2183,9 @@ remember_a_message_plural (message_ty *mp, char *string,
 	    if (descr != NULL)
 	      {
 		/* Same heuristic as in remember_a_message.  */
-		if (parser->get_number_of_directives (descr) > 0)
+		if (parser->get_number_of_directives (descr) > 0
+		    && !(parser->is_unlikely_intentional != NULL
+			 && parser->is_unlikely_intentional (descr)))
 		  mp->is_format[i] = possible;
 
 		parser->free (descr);