www/compatibility.html - platform/external/clang - Gitiles

 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
           "http://www.w3.org/TR/html4/strict.dtd">
 <html>
 <head>
   <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1" />
   <title>Language Compatibility</title>
   <link type="text/css" rel="stylesheet" href="menu.css" />
   <link type="text/css" rel="stylesheet" href="content.css" />
   <style type="text/css">
 </style>
 </head>
 <body>

 <!--#include virtual="menu.html.incl"-->

 <div id="content">

 <!-- ======================================================================= -->
 <h1>Language Compatibility</h1>
 <!-- ======================================================================= -->

 <p>Clang strives to both conform to current language standards (C99,
   C++98) and also to implement many widely-used extensions available
   in other compilers, so that most correct code will "just work" when
   compiler with Clang. However, Clang is more strict than other
   popular compilers, and may reject incorrect code that other
   compilers allow. This page documents common compatibility and
   portability issues with Clang to help you understand and fix the
   problem in your code when Clang emits an error message.</p>

 <ul>
   <li><a href="#c">C compatibility</a>
     <ul>
       <li><a href="#inline">C99 inline functions</a></li>
       <li><a href="#lvalue-cast">Lvalue casts</a></li>
     </ul>
   </li>
   <li><a href="#objective-c">Objective-C compatibility</a>
     <ul>
       <li><a href="#super-cast">Cast of super</a></li>
       <li><a href="#sizeof-interface">Size of interfaces</a></li>
     </ul>
   </li>
   <li><a href="#c++">C++ compatibility</a>
     <ul>
       <li><a href="#vla">Variable-length arrays</a></li>
       <li><a href="#init_static_const">Initialization of non-integral static const data members within a class definition</a></li>
       <li><a href="#dep_lookup">Unqualified lookup in templates</a></li>
       <li><a href="#dep_lookup_bases">Unqualified lookup into dependent bases of class templates</a></li>
       <li><a href="#undep_incomplete">Incomplete types in templates</a></li>
       <li><a href="#bad_templates">Templates with no valid instantiations</a></li>
       <li><a href="#default_init_const">Default initialization of const
       variable of a class type requires user-defined default
       constructor</a></li>
     </ul>
   </li>
   <li><a href="#objective-c++">Objective-C++ compatibility</a>
     <ul>
       <li><a href="#implicit-downcasts">Implicit downcasts</a></li>
     </ul>
   </li>
 </ul>

 <!-- ======================================================================= -->
 <h2 id="c">C compatibility</h3>
 <!-- ======================================================================= -->

 <!-- ======================================================================= -->
 <h3 id="inline">C99 inline functions</h3>
 <!-- ======================================================================= -->
 <p>By default, Clang builds C code according to the C99 standard,
 which provides different inlining semantics than GCC's default
 behavior. For example, when compiling the following code with no optimization:</p>
 <pre>
 inline int add(int i, int j) { return i + j; }

 int main() {
   int i = add(4, 5);
   return i;
 }
 </pre>

 <p>In C99, this is an incomplete (incorrect) program because there is
 no external definition of the <code>add</code> function: the inline
 definition is only used for optimization, if the compiler decides to
 perform inlining. Therefore, we will get a (correct) link-time error
 with Clang, e.g.:</p>

 <pre>
 Undefined symbols:
   "_add", referenced from:
       _main in cc-y1jXIr.o
 </pre>

 <p>There are several ways to fix this problem:</p>

 <ul>
   <li>Change <code>add</code> to a <code>static inline</code>
   function. Static inline functions are always resolved within the
   translation unit, so you won't have to add an external, non-inline
   definition of the function elsewhere in your program.</li>

   <li>Provide an external (non-inline) definition of <code>add</code>
   somewhere in your program.</li>

   <li>Compile with the GNU89 dialect by adding
   <code>-std=gnu89</code> to the set of Clang options. This option is
   only recommended if the program source cannot be changed or if the
   program also relies on additional C89-specific behavior that cannot
   be changed.</li>
 </ul>

 <!-- ======================================================================= -->
 <h3 id="lvalue-cast">Lvalue casts</h3>
 <!-- ======================================================================= -->

 <p>Old versions of GCC permit casting the left-hand side of an assignment to a
 different type. Clang produces an error on similar code, e.g.,</p>

 <pre>
 lvalue.c:2:3: error: assignment to cast is illegal, lvalue casts are not
       supported
   (int*)addr = val;
   ^~~~~~~~~~ ~
 </pre>

 <p>To fix this problem, move the cast to the right-hand side. In this
 example, one could use:</p>

 <pre>
   addr = (float *)val;
 </pre>

 <!-- ======================================================================= -->
 <h2 id="objective-c">Objective-C compatibility</h3>
 <!-- ======================================================================= -->

 <!-- ======================================================================= -->
 <h3 id="super-cast">Cast of super</h3>
 <!-- ======================================================================= -->

 <p>GCC treats the <code>super</code> identifier as an expression that
 can, among other things, be cast to a different type. Clang treats
 <code>super</code> as a context-sensitive keyword, and will reject a
 type-cast of <code>super</code>:</p>

 <pre>
 super.m:11:12: error: cannot cast 'super' (it isn't an expression)
   [(Super*)super add:4];
    ~~~~~~~~^
 </pre>

 <p>To fix this problem, remove the type cast, e.g.</p>
 <pre>
   [super add:4];
 </pre>

 <!-- ======================================================================= -->
 <h3 id="sizeof-interface">Size of interfaces</h3>
 <!-- ======================================================================= -->

 <p>When using the "non-fragile" Objective-C ABI in use, the size of an
 Objective-C class may change over time as instance variables are added
 (or removed). For this reason, Clang rejects the application of the
 <code>sizeof</code> operator to an Objective-C class when using this
 ABI:</p>

 <pre>
 sizeof.m:4:14: error: invalid application of 'sizeof' to interface 'NSArray' in
       non-fragile ABI
   int size = sizeof(NSArray);
              ^     ~~~~~~~~~
 </pre>

 <p>Code that relies on the size of an Objective-C class is likely to
 be broken anyway, since that size is not actually constant. To address
 this problem, use the Objective-C runtime API function
 <code>class_getInstanceSize()</code>:</p>

 <pre>
   class_getInstanceSize([NSArray class])
 </pre>

 <!-- ======================================================================= -->
 <h2 id="c++">C++ compatibility</h3>
 <!-- ======================================================================= -->

 <!-- ======================================================================= -->
 <h3 id="vla">Variable-length arrays</h3>
 <!-- ======================================================================= -->

 <p>GCC and C99 allow an array's size to be determined at run
 time. This extension is not permitted in standard C++. However, Clang
 supports such variable length arrays in very limited circumstances for
 compatibility with GNU C and C99 programs:</p>

 <ul>
   <li>The element type of a variable length array must be a POD
   ("plain old data") type, which means that it cannot have any
   user-declared constructors or destructors, base classes, or any
   members if non-POD type. All C types are POD types.</li>

   <li>Variable length arrays cannot be used as the type of a non-type
 template parameter.</li> </ul>

 <p>If your code uses variable length arrays in a manner that Clang doesn't support, there are several ways to fix your code:

 <ol>
 <li>replace the variable length array with a fixed-size array if you can
     determine a
     reasonable upper bound at compile time; sometimes this is as
     simple as changing <tt>int size = ...;</tt> to <tt>const int size
     = ...;</tt> (if the definition of <tt>size</tt> is a compile-time
     integral constant);</li>
 <li>use an <tt>std::string</tt> instead of a <tt>char []</tt>;</li>
 <li>use <tt>std::vector</tt> or some other suitable container type;
     or</li>
 <li>allocate the array on the heap instead using <tt>new Type[]</tt> -
     just remember to <tt>delete[]</tt> it.</li>
 </ol>

 <!-- ======================================================================= -->
 <h3 id="init_static_const">Initialization of non-integral static const data members within a class definition</h3>
 <!-- ======================================================================= -->

 The following code is ill-formed in C++'03:

 <pre>
 class SomeClass {
  public:
   static const double SomeConstant = 0.5;
 };

 const double SomeClass::SomeConstant;
 </pre>

 Clang errors with something similar to:

 <pre>
 .../your_file.h:42:42: error: 'SomeConstant' can only be initialized if it is a static const integral data member
   static const double SomeConstant = 0.5;
                       ^              ~~~
 </pre>

 Only <i>integral</i> constant expressions are allowed as initializers
 within the class definition. See C++'03 [class.static.data] p4 for the
 details of this restriction.  The fix here is straightforward: move
 the initializer to the definition of the static data member, which
 must exist outside of the class definition:

 <pre>
 class SomeClass {
  public:
   static const double SomeConstant;
 };

 const double SomeClass::SomeConstant<b> = 0.5</b>;
 </pre>

 Note that the forthcoming C++0x standard will allow this.

 <!-- ======================================================================= -->
 <h3 id="dep_lookup">Unqualified lookup in templates</h3>
 <!-- ======================================================================= -->

 <p>Some versions of GCC accept the following invalid code:

 <pre>
 template &lt;typename T&gt; T Squared(T x) {
   return Multiply(x, x);
 }

 int Multiply(int x, int y) {
   return x * y;
 }

 int main() {
   Squared(5);
 }
 </pre>

 <p>Clang complains:

 <pre>  <b>my_file.cpp:2:10: <span class="error">error:</span> use of undeclared identifier 'Multiply'</b>
     return Multiply(x, x);
   <span class="caret">         ^</span>

   <b>my_file.cpp:10:3: <span class="note">note:</span> in instantiation of function template specialization 'Squared&lt;int&gt;' requested here</b>
     Squared(5);
   <span class="caret">  ^</span>
 </pre>

 <p>The C++ standard says that unqualified names like <q>Multiply</q>
 are looked up in two ways.

 <p>First, the compiler does <i>unqualified lookup</i> in the scope
 where the name was written.  For a template, this means the lookup is
 done at the point where the template is defined, not where it's
 instantiated.  Since <tt>Multiply</tt> hasn't been declared yet at
 this point, unqualified lookup won't find it.

 <p>Second, if the name is called like a function, then the compiler
 also does <i>argument-dependent lookup</i> (ADL).  (Sometimes
 unqualified lookup can suppress ADL; see [basic.lookup.argdep]p3 for
 more information.)  In ADL, the compiler looks at the types of all the
 arguments to the call.  When it finds a class type, it looks up the
 name in that class's namespace; the result is all the declarations it
 finds in those namespaces, plus the declarations from unqualified
 lookup.  However, the compiler doesn't do ADL until it knows all the
 argument types.

 <p>In our example, <tt>Multiply</tt> is called with dependent
 arguments, so ADL isn't done until the template is instantiated.  At
 that point, the arguments both have type <tt>int</tt>, which doesn't
 contain any class types, and so ADL doesn't look in any namespaces.
 Since neither form of lookup found the declaration
 of <tt>Multiply</tt>, the code doesn't compile.

 <p>Here's another example, this time using overloaded operators,
 which obey very similar rules.

 <pre>#include &lt;iostream&gt;

 template&lt;typename T&gt;
 void Dump(const T&amp; value) {
   std::cout &lt;&lt; value &lt;&lt; "\n";
 }

 namespace ns {
   struct Data {};
 }

 std::ostream&amp; operator&lt;&lt;(std::ostream&amp; out, ns::Data data) {
   return out &lt;&lt; "Some data";
 }

 void Use() {
   Dump(ns::Data());
 }</pre>

 <p>Again, Clang complains about not finding a matching function:</p>

 <pre>
 <b>my_file.cpp:5:13: <span class="error">error:</span> invalid operands to binary expression ('ostream' (aka 'basic_ostream&lt;char&gt;') and 'ns::Data const')</b>
   std::cout &lt;&lt; value &lt;&lt; "\n";
   <span class="caret">~~~~~~~~~ ^  ~~~~~</span>
 <b>my_file.cpp:17:3: <span class="note">note:</span> in instantiation of function template specialization 'Dump&lt;ns::Data&gt;' requested here</b>
   Dump(ns::Data());
   <span class="caret">^</span>
 </pre>

 <p>Just like before, unqualified lookup didn't find any declarations
 with the name <tt>operator&lt;&lt;</tt>.  Unlike before, the argument
 types both contain class types: one of them is an instance of the
 class template type <tt>std::basic_ostream</tt>, and the other is the
 type <tt>ns::Data</tt> that we declared above.  Therefore, ADL will
 look in the namespaces <tt>std</tt> and <tt>ns</tt> for
 an <tt>operator&lt;&lt;</tt>.  Since one of the argument types was
 still dependent during the template definition, ADL isn't done until
 the template is instantiated during <tt>Use</tt>, which means that
 the <tt>operator&lt;&lt;</tt> we want it to find has already been
 declared.  Unfortunately, it was declared in the global namespace, not
 in either of the namespaces that ADL will look in!

 <p>There are two ways to fix this problem:</p>
 <ol><li>Make sure the function you want to call is declared before the
 template that might call it.  This is the only option if none of its
 argument types contain classes.  You can do this either by moving the
 template definition, or by moving the function definition, or by
 adding a forward declaration of the function before the template.</li>
 <li>Move the function into the same namespace as one of its arguments
 so that ADL applies.</li></ol>

 <p>For more information about argument-dependent lookup, see
 [basic.lookup.argdep].  For more information about the ordering of
 lookup in templates, see [temp.dep.candidate].

 <!-- ======================================================================= -->
 <h3 id="dep_lookup_bases">Unqualified lookup into dependent bases of class templates</h3>
 <!-- ======================================================================= -->

 Some versions of GCC accept the following invalid code:

 <pre>
 template &lt;typename T&gt; struct Base {
   void DoThis(T x) {}
   static void DoThat(T x) {}
 };

 template &lt;typename T&gt; struct Derived : public Base&lt;T&gt; {
   void Work(T x) {
     DoThis(x);  // Invalid!
     DoThat(x);  // Invalid!
   }
 };
 </pre>

 Clang correctly rejects it with the following errors
 (when <tt>Derived</tt> is eventually instantiated):

 <pre>
 my_file.cpp:8:5: error: use of undeclared identifier 'DoThis'
     DoThis(x);
     ^
     this-&gt;
 my_file.cpp:2:8: note: must qualify identifier to find this declaration in dependent base class
   void DoThis(T x) {}
        ^
 my_file.cpp:9:5: error: use of undeclared identifier 'DoThat'
     DoThat(x);
     ^
     this-&gt;
 my_file.cpp:3:15: note: must qualify identifier to find this declaration in dependent base class
   static void DoThat(T x) {}
 </pre>

 Like we said <a href="#dep_lookup">above</a>, unqualified names like
 <tt>DoThis</tt> and <tt>DoThat</tt> are looked up when the template
 <tt>Derived</tt> is defined, not when it's instantiated.  When we look
 up a name used in a class, we usually look into the base classes.
 However, we can't look into the base class <tt>Base&lt;T&gt;</tt>
 because its type depends on the template argument <tt>T</tt>, so the
 standard says we should just ignore it.  See [temp.dep]p3 for details.

 <p>The fix, as Clang tells you, is to tell the compiler that we want a
 class member by prefixing the calls with <tt>this-&gt;</tt>:

 <pre>
   void Work(T x) {
     <b>this-&gt;</b>DoThis(x);
     <b>this-&gt;</b>DoThat(x);
   }
 </pre>

 Alternatively, you can tell the compiler exactly where to look:

 <pre>
   void Work(T x) {
     <b>Base&lt;T&gt;</b>::DoThis(x);
     <b>Base&lt;T&gt;</b>::DoThat(x);
   }
 </pre>

 This works whether the methods are static or not, but be careful:
 if <tt>DoThis</tt> is virtual, calling it this way will bypass virtual
 dispatch!

 <!-- ======================================================================= -->
 <h3 id="undep_incomplete">Incomplete types in templates</h3>
 <!-- ======================================================================= -->

 The following code is invalid, but compilers are allowed to accept it:

 <pre>
   class IOOptions;
   template &lt;class T&gt; bool read(T &amp;value) {
     IOOptions opts;
     return read(opts, value);
   }

   class IOOptions { bool ForceReads; };
   bool read(const IOOptions &amp;opts, int &amp;x);
   template bool read&lt;&gt;(int &amp;);
 </pre>

 The standard says that types which don't depend on template parameters
 must be complete when a template is defined if they affect the
 program's behavior.  However, the standard also says that compilers
 are free to not enforce this rule.  Most compilers enforce it to some
 extent; for example, it would be an error in GCC to
 write <tt>opts.ForceReads</tt> in the code above.  In Clang, we feel
 that enforcing the rule consistently lets us provide a better
 experience, but unfortunately it also means we reject some code that
 other compilers accept.

 <p>We've explained the rule here in very imprecise terms; see
 [temp.res]p8 for details.

 <!-- ======================================================================= -->
 <h3 id="bad_templates">Templates with no valid instantiations</h3>
 <!-- ======================================================================= -->

 The following code contains a typo: the programmer
 meant <tt>init()</tt> but wrote <tt>innit()</tt> instead.

 <pre>
   template &lt;class T&gt; class Processor {
     ...
     void init();
     ...
   };
   ...
   template &lt;class T&gt; void process() {
     Processor&lt;T&gt; processor;
     processor.innit();       // <-- should be 'init()'
     ...
   }
 </pre>

 Unfortunately, we can't flag this mistake as soon as we see it: inside
 a template, we're not allowed to make assumptions about "dependent
 types" like <tt>Processor&lt;T&gt;</tt>.  Suppose that later on in
 this file the programmer adds an explicit specialization
 of <tt>Processor</tt>, like so:

 <pre>
   template &lt;&gt; class Processor&lt;char*&gt; {
     void innit();
   };
 </pre>

 Now the program will work &mdash; as long as the programmer only ever
 instantiates <tt>process()</tt> with <tt>T = char*</tt>!  This is why
 it's hard, and sometimes impossible, to diagnose mistakes in a
 template definition before it's instantiated.

 <p>The standard says that a template with no valid instantiations is
 ill-formed.  Clang tries to do as much checking as possible at
 definition-time instead of instantiation-time: not only does this
 produce clearer diagnostics, but it also substantially improves
 compile times when using pre-compiled headers.  The downside to this
 philosophy is that Clang sometimes fails to process files because they
 contain broken templates that are no longer used.  The solution is
 simple: since the code is unused, just remove it.

 <!-- ======================================================================= -->
 <h3 id="default_init_const">Default initialization of const variable of a class type requires user-defined default constructor</h3>
 <!-- ======================================================================= -->

 If a <tt>class</tt> or <tt>struct</tt> has no user-defined default
 constructor, C++ doesn't allow you to default construct a <tt>const</tt>
 instance of it like this ([dcl.init], p9):

 <pre>
 class Foo {
  public:
   // The compiler-supplied default constructor works fine, so we
   // don't bother with defining one.
   ...
 };

 void Bar() {
   const Foo foo;  // Error!
   ...
 }
 </pre>

 To fix this, you can define a default constructor for the class:

 <pre>
 class Foo {
  public:
   Foo() {}
   ...
 };

 void Bar() {
   const Foo foo;  // Now the compiler is happy.
   ...
 }
 </pre>

 <!-- ======================================================================= -->
 <h2 id="objective-c++">Objective-C++ compatibility</h3>
 <!-- ======================================================================= -->

 <!-- ======================================================================= -->
 <h3 id="implicit-downcasts">Implicit downcasts</h3>
 <!-- ======================================================================= -->

 <p>Due to a bug in its implementation, GCC allows implicit downcasts
 (from base class to a derived class) when calling functions. Such code is
 inherently unsafe, since the object might not actually be an instance
 of the derived class, and is rejected by Clang. For example, given
 this code:</p>

 <pre>
 @interface Base @end
 @interface Derived : Base @end

 void f(Derived *);
 void g(Base *base) {
   f(base);
 }
 </pre>

 <p>Clang produces the following error:</p>

 <pre>
 downcast.mm:6:3: error: no matching function for call to 'f'
   f(base);
   ^
 downcast.mm:4:6: note: candidate function not viable: cannot convert from base
       class pointer 'Base *' to derived class pointer 'Derived *' for 1st
       argument
 void f(Derived *);
      ^
 </pre>

 <p>If the downcast is actually correct (e.g., because the code has
 already checked that the object has the appropriate type), add an
 explicit cast:</p>

 <pre>
   f((Derived *)base);
 </pre>

 </div>
 </body>
 </html>