Merge pull request github-linguist#162 from horchler/detect-matlab

More robust heuristics for .m files and 3 new Matlab tests. Finally.

lib/linguist/blob_helper.rb

@@ -451,24 +451,26 @@ def guess_h_language
     # Internal: Guess language of .m files.
     #
     # Objective-C heuristics:
-    # * Keywords
+    # * Keywords  ("#import", "#include", "#ifdef", #define, "@end") or "//" and opening "\*" comments
     #
     # Matlab heuristics:
-    # * Leading function keyword
+    # * Leading "function " of "classdef " keyword
     # * "%" comments
     #
+    # Note: All "#" keywords, e.g., "#import", are guaranteed to be Objective-C. Because the ampersand
+    # is used to created function handles and anonymous functions in Matlab, most "@" keywords are not
+    # safe heuristics. However, "end" is a reserved term in Matlab and can't be used to create a valid
+    # function handle. Because @end is required to close any @implementation, @property, @interface,
+    # @synthesize, etc. directive in Objective-C, only @end needs to be checked for.
+    #
     # Returns a Language.
     def guess_m_language
-      # Objective-C keywords
-      if lines.grep(/^#import|@(interface|implementation|property|synthesize|end)/).any?
+      # Objective-C keywords or comments
+      if lines.grep(/^#(import|include|ifdef|define)|@end/).any? || lines.grep(/^\s*\/\//).any? || lines.grep(/^\s*\/\*/).any?
         Language['Objective-C']
 
-      # File function
-      elsif lines.first.to_s =~ /^function /
-        Language['Matlab']
-
-      # Matlab comment
-      elsif lines.grep(/^%/).any?
+      # Matlab file function or class or comments
+      elsif lines.first.grep(/^\s*(function |classdef )/).any? || lines.grep(/^\s*%/).any?
         Language['Matlab']
 
       # Fallback to Objective-C, don't want any Matlab false positives

test/fixtures/matlab_class.m

@@ -0,0 +1,29 @@
+classdef matlab_class
+    properties
+        R;
+        G;
+        B;
+    end
+    methods
+        function obj = matlab_class(r,g,b)
+            obj.R = r;
+            obj.G = g;
+            obj.B = b;
+        end
+        function disp(obj)
+            disp(['Red: ' num2str(obj.R) ...
+                ', Green: ' num2str(obj.G) ...
+                ', Blue: ' num2str(obj.B)]);
+        end
+    end
+    enumeration
+        red     (1,0,0)
+        green   (0,1,0)
+        blue    (0,0,1)
+        cyan    (0,1,1)
+        magenta (1,0,1)
+        yellow  (1,1,0)
+        black   (0,0,0)
+        white   (1,1,1)
+    end
+end

test/fixtures/matlab_function2.m

@@ -0,0 +1,33 @@
+   function ret = matlab_function2(A,B)
+% Simple function that combines two values using function handles and displays
+% the return value
+
+% create function handles
+fun1=@interface;
+fun2=@implementation;
+fun3=@property;
+fun4=@synthesize;
+
+% use function handles
+ret = fun1(A)+fun2(A)+fun3(B)+fun4(B);
+
+% Display the return value
+disp('Return value in function');
+disp(ret);
+
+
+function A=interface(A)
+% simple sub-function with same name Objective-C @keyword
+A=2*A;
+
+function A=implementation(A)
+% simple sub-function with same name Objective-C @keyword
+A=A^2;
+
+function B=property(B)
+% simple sub-function with same name Objective-C @keyword
+B=2*B;
+
+function B=synthesize(B)
+% simple sub-function with same name Objective-C @keyword
+B=B^2;

test/fixtures/matlab_script2.m

@@ -0,0 +1,13 @@
+  % Matlab example script 2
+  % Comments precended with arbitrary whitespace (spaces or tabs)
+
+  %Call matlab_function function which resides in the same directory
+
+value1 = 5 % semicolon at end of line is not mandatory, only suppresses output to command line.
+value2 = 3
+
+  % Calculate sum of value1 and value2
+result = matlab_function(value1,value2);
+
+disp('called from script')
+disp(result);

test/test_blob.rb

@@ -311,6 +311,9 @@ def test_language
     assert_equal Language['Objective-C'], blob("hello.m").language
     assert_equal Language['Matlab'], blob("matlab_function.m").language
     assert_equal Language['Matlab'], blob("matlab_script.m").language
+    assert_equal Language['Matlab'], blob("matlab_function2.m").language
+    assert_equal Language['Matlab'], blob("matlab_script2.m").language
+    assert_equal Language['Matlab'], blob("matlab_class.m").language
 
     # .r disambiguation
     assert_equal Language['R'],           blob("hello-r.R").language