Adding the --include and --exclude options to lcov and geninfo

* The --include command-line option allows the user to specify a regular
  expression for the source files to be included. The command-line
  option can be repeated to specify multiple patterns. The coverage
  information is only included for the source files matching at least
  one of the patterns.

  The "lcov --capture --include" (or "geninfo --include") option is
  similar in functionality to the "lcov --extract" command-line
  option. But, by directly using applying the pattern while capturing
  coverage data one can often avoid having to run "lcov --extract" as a
  second pass.

* The --exclude command-line option allows the user to specify a regular
  expression for the source files to be excluded. The command-line
  option can be repeated to specify multiple patterns. The coverage
  information is excluded for source files matching at least one of the
  patterns.

  The "lcov --capture --exclude" (or "geninfo --exclude") option is
  similar in functionality to the "lcov --extract" command-line
  option. But, by directly using applying the pattern while capturing
  coverage data one can often avoid having to run "lcov --remove" as a
  second pass.

* On one of our code base at Autodesk, this speeds-up the generation of
  HTML code coverage reports by a factor of 3X.

Signed-off-by: Benoit Belley <benoit.belley@autodesk.com>

Author: adsk-belleyb

bin/geninfo

@@ -151,6 +151,7 @@ our $UNNAMED_BLOCK	= -1;
 
 # Prototypes
 sub print_usage(*);
+sub transform_pattern($);
 sub gen_info($);
 sub process_dafile($$);
 sub match_filename($@);
@@ -239,6 +240,9 @@ our $config;		# Configuration file contents
 our @ignore_errors;	# List of errors to ignore (parameter)
 our @ignore;		# List of errors to ignore (array)
 our $initial;
+our @include_patterns; # List of source file patterns to include
+our @exclude_patterns; # List of source file patterns to exclude
+our %excluded_files; # Files excluded due to include/exclude options
 our $no_recursion = 0;
 our $maxdepth;
 our $no_markers = 0;
@@ -382,11 +386,13 @@ if (!GetOptions("test-name|t=s" => \$test_name,
 		"gcov-tool=s" => \$gcov_tool,
 		"ignore-errors=s" => \@ignore_errors,
 		"initial|i" => \$initial,
+		"include=s" => \@include_patterns,
+		"exclude=s" => \@exclude_patterns,
 		"no-recursion" => \$no_recursion,
 		"no-markers" => \$no_markers,
 		"derive-func-data" => \$opt_derive_func_data,
 		"debug" => \$debug,
-		"external" => \$opt_external,
+		"external|e" => \$opt_external,
 		"no-external" => \$opt_no_external,
 		"compat=s" => \$opt_compat,
 		"config-file=s" => \$opt_config_file,
@@ -415,6 +421,16 @@ else
 		$opt_external = 0;
 		$opt_no_external = undef;
 	}
+
+	if(@include_patterns) {
+		# Need perlreg expressions instead of shell pattern
+		@include_patterns = map({ transform_pattern($_); } @include_patterns);
+	}
+
+	if(@exclude_patterns) {
+		# Need perlreg expressions instead of shell pattern
+		@exclude_patterns = map({ transform_pattern($_); } @exclude_patterns);
+	}
 }
 
 @data_directory = @ARGV;
@@ -622,12 +638,53 @@ sequentially.
       --config-file FILENAME        Specify configuration file location
       --rc SETTING=VALUE            Override configuration file setting
       --compat MODE=on|off|auto     Set compat MODE (libtool, hammer, split_crc)
+      --include PATTERN             Include files matching PATTERN
+      --exclude PATTERN             Exclude files matching PATTERN
 
 For more information see: $lcov_url
 END_OF_USAGE
 	;
 }
 
+
+#
+# transform_pattern(pattern)
+#
+# Transform shell wildcard expression to equivalent Perl regular expression.
+# Return transformed pattern.
+#
+
+sub transform_pattern($)
+{
+	my $pattern = $_[0];
+
+	# Escape special chars
+
+	$pattern =~ s/\\/\\\\/g;
+	$pattern =~ s/\//\\\//g;
+	$pattern =~ s/\^/\\\^/g;
+	$pattern =~ s/\$/\\\$/g;
+	$pattern =~ s/\(/\\\(/g;
+	$pattern =~ s/\)/\\\)/g;
+	$pattern =~ s/\[/\\\[/g;
+	$pattern =~ s/\]/\\\]/g;
+	$pattern =~ s/\{/\\\{/g;
+	$pattern =~ s/\}/\\\}/g;
+	$pattern =~ s/\./\\\./g;
+	$pattern =~ s/\,/\\\,/g;
+	$pattern =~ s/\|/\\\|/g;
+	$pattern =~ s/\+/\\\+/g;
+	$pattern =~ s/\!/\\\!/g;
+
+	# Transform ? => (.) and * => (.*)
+
+	$pattern =~ s/\*/\(\.\*\)/g;
+	$pattern =~ s/\?/\(\.\)/g;
+
+	return $pattern;
+}
+
+
 #
 # get_common_prefix(min_dir, filenames)
 #
@@ -743,6 +800,12 @@ sub gen_info($)
 			process_dafile($file, $prefix);
 		}
 	}
+
+	# Report whether files were excluded.
+	if (%excluded_files) {
+		info("Excluded data for %d files due to include/exclude options\n",
+			 scalar keys %excluded_files);
+	}
 }
 
 
@@ -1176,6 +1239,40 @@ sub process_dafile($$)
 						\@matches, \@gcov_content);
 		}
 
+		if (@include_patterns)
+		{
+			my $keep = 0;
+
+			foreach my $pattern (@include_patterns)
+			{
+				$keep ||= ($source_filename =~ (/^$pattern$/));
+			}
+
+			if (!$keep)
+			{
+				$excluded_files{$source_filename} = ();
+				unlink($gcov_file);
+				next;
+			}
+		}
+
+		if (@exclude_patterns)
+		{
+			my $exclude = 0;
+
+			foreach my $pattern (@exclude_patterns)
+			{
+				$exclude ||= ($source_filename =~ (/^$pattern$/));
+			}
+
+			if ($exclude)
+			{
+				$excluded_files{$source_filename} = ();
+				unlink($gcov_file);
+				next;
+			}
+		}
+
 		# Skip external files if requested
 		if (!$opt_external) {
 			if (is_external($source_filename)) {

bin/lcov

@@ -166,6 +166,8 @@ our $no_compat_libtool;	# If set, indicates that libtool mode is to be disabled
 our $gcov_tool;
 our @opt_ignore_errors;
 our $initial;
+our @include_patterns; # List of source file patterns to include
+our @exclude_patterns; # List of source file patterns to exclude
 our $no_recursion = 0;
 our $to_package;
 our $from_package;
@@ -281,6 +283,8 @@ if (!GetOptions("directory|d|di=s" => \@directory,
 		"gcov-tool=s" => \$gcov_tool,
 		"ignore-errors=s" => \@opt_ignore_errors,
 		"initial|i" => \$initial,
+		"include=s" => \@include_patterns,
+		"exclude=s" => \@exclude_patterns,
 		"no-recursion" => \$no_recursion,
 		"to-package=s" => \$to_package,
 		"from-package=s" => \$from_package,
@@ -538,6 +542,8 @@ Options:
       --config-file FILENAME      Specify configuration file location
       --rc SETTING=VALUE          Override configuration file setting
       --compat MODE=on|off|auto   Set compat MODE (libtool, hammer, split_crc)
+      --include PATTERN           Include files matching PATTERN
+      --exclude PATTERN           Exclude files matching PATTERN
 
 For more information see: $lcov_url
 END_OF_USAGE
@@ -885,6 +891,12 @@ sub lcov_geninfo(@)
 	if (defined($opt_config_file)) {
 		@param = (@param, "--config-file", $opt_config_file);
 	}
+	foreach (@include_patterns) {
+		@param = (@param, "--include", $_);
+	}
+	foreach (@exclude_patterns) {
+		@param = (@param, "--exclude", $_);
+	}
 
 	system(@param) and exit($? >> 8);
 }

man/geninfo.1

@@ -45,6 +45,11 @@ geninfo \- Generate tracefiles from .da files
 .br
 .RB [ \-\-rc
 .IR keyword = value ]
+.br
+.RB [ \-\-include
+.IR pattern ]
+.RB [ \-\-exclude
+.IR pattern ]
 .RE
 .SH DESCRIPTION
 .B geninfo 
@@ -286,6 +291,28 @@ instead derive function coverage data from line coverage data and
 information about which lines belong to a function.
 .RE
 
+.B \-\-exclude
+.I pattern
+.br
+.RS
+Exclude source files matching
+.IR pattern .
+
+Use this switch if you want to exclude coverage data for a particular set
+of source files matching any of the given patterns. Multiple patterns can be
+specified by using multiple
+.B --exclude
+command line switches. The
+.I patterns
+will be interpreted as shell wildcard patterns (note that they may need to be
+escaped accordingly to prevent the shell from expanding them first).
+
+Can be combined with the
+.B --include
+command line switch. If a given file matches both the include pattern and the
+exclude pattern, the exclude pattern will take precedence.
+.RE
+
 .B \-\-external
 .br
 .B \-\-no\-external
@@ -324,6 +351,23 @@ Specify the location of the gcov tool.
 Print a short help text, then exit.
 .RE
 
+.B \-\-include
+.I pattern
+.br
+.RS
+Include source files matching
+.IR pattern .
+
+Use this switch if you want to include coverage data for only a particular set
+of source files matching any of the given patterns. Multiple patterns can be
+specified by using multiple
+.B --include
+command line switches. The
+.I patterns
+will be interpreted as shell wildcard patterns (note that they may need to be
+escaped accordingly to prevent the shell from expanding them first).
+.RE
+
 .B \-\-ignore\-errors
 .I errors
 .br

man/lcov.1

@@ -50,6 +50,11 @@ lcov \- a graphical GCOV front\-end
 .RB [ \-\-compat
 .IR  mode =on|off|auto]
 .br
+.RB [ \-\-include
+.IR pattern ]
+.RB [ \-\-exclude
+.IR pattern ]
+.br
 .RE
 
 .B lcov
@@ -476,6 +481,28 @@ where the counter files ending with .da will be stored).
 Note that you may specify this option more than once.
 .RE
 
+.B \-\-exclude
+.I pattern
+.br
+.RS
+Exclude source files matching
+.IR pattern .
+
+Use this switch if you want to exclude coverage data for a particular set
+of source files matching any of the given patterns. Multiple patterns can be
+specified by using multiple
+.B --exclude
+command line switches. The
+.I patterns
+will be interpreted as shell wildcard patterns (note that they may need to be
+escaped accordingly to prevent the shell from expanding them first).
+
+Can be combined with the
+.B --include
+command line switch. If a given file matches both the include pattern and the
+exclude pattern, the exclude pattern will take precedence.
+.RE
+
 .B \-\-external
 .br
 .B \-\-no\-external
@@ -556,6 +583,23 @@ Specify the location of the gcov tool.
 Print a short help text, then exit.
 .RE
 
+.B \-\-include
+.I pattern
+.br
+.RS
+Include source files matching
+.IR pattern .
+
+Use this switch if you want to include coverage data for only a particular set
+of source files matching any of the given patterns. Multiple patterns can be
+specified by using multiple
+.B --include
+command line switches. The
+.I patterns
+will be interpreted as shell wildcard patterns (note that they may need to be
+escaped accordingly to prevent the shell from expanding them first).
+.RE
+
 .B \-\-ignore\-errors
 .I errors
 .br