3 c8661ffa 2005-11-29 devnull <!-- This HTML file has been created by texi2html 1.52
4 c8661ffa 2005-11-29 devnull from gxxint.texi on 27 August 1999 -->
6 c8661ffa 2005-11-29 devnull <TITLE>G++ internals - Mangling</TITLE>
9 c8661ffa 2005-11-29 devnull Go to the <A HREF="gxxint_1.html">first</A>, <A HREF="gxxint_14.html">previous</A>, <A HREF="gxxint_16.html">next</A>, <A HREF="gxxint_16.html">last</A> section, <A HREF="gxxint_toc.html">table of contents</A>.
10 c8661ffa 2005-11-29 devnull <P><HR><P>
13 c8661ffa 2005-11-29 devnull <H2><A NAME="SEC20" HREF="gxxint_toc.html#TOC20">Function name mangling for C++ and Java</A></H2>
16 c8661ffa 2005-11-29 devnull Both C++ and Jave provide overloaded function and methods,
17 c8661ffa 2005-11-29 devnull which are methods with the same types but different parameter lists.
18 c8661ffa 2005-11-29 devnull Selecting the correct version is done at compile time.
19 c8661ffa 2005-11-29 devnull Though the overloaded functions have the same name in the source code,
20 c8661ffa 2005-11-29 devnull they need to be translated into different assembler-level names,
21 c8661ffa 2005-11-29 devnull since typical assemblers and linkers cannot handle overloading.
22 c8661ffa 2005-11-29 devnull This process of encoding the parameter types with the method name
23 c8661ffa 2005-11-29 devnull into a unique name is called <EM>name mangling</EM>. The inverse
24 c8661ffa 2005-11-29 devnull process is called <EM>demangling</EM>.
28 c8661ffa 2005-11-29 devnull It is convenient that C++ and Java use compatible mangling schemes,
29 c8661ffa 2005-11-29 devnull since the makes life easier for tools such as gdb, and it eases
30 c8661ffa 2005-11-29 devnull integration between C++ and Java.
34 c8661ffa 2005-11-29 devnull Note there is also a standard "Jave Native Interface" (JNI) which
35 c8661ffa 2005-11-29 devnull implements a different calling convention, and uses a different
36 c8661ffa 2005-11-29 devnull mangling scheme. The JNI is a rather abstract ABI so Java can call methods
37 c8661ffa 2005-11-29 devnull written in C or C++;
38 c8661ffa 2005-11-29 devnull we are concerned here about a lower-level interface primarily
39 c8661ffa 2005-11-29 devnull intended for methods written in Java, but that can also be used for C++
40 c8661ffa 2005-11-29 devnull (and less easily C).
45 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC21" HREF="gxxint_toc.html#TOC21">Method name mangling</A></H3>
48 c8661ffa 2005-11-29 devnull C++ mangles a method by emitting the function name, followed by <CODE>__</CODE>,
49 c8661ffa 2005-11-29 devnull followed by encodings of any method qualifiers (such as <CODE>const</CODE>),
50 c8661ffa 2005-11-29 devnull followed by the mangling of the method's class,
51 c8661ffa 2005-11-29 devnull followed by the mangling of the parameters, in order.
55 c8661ffa 2005-11-29 devnull For example <CODE>Foo::bar(int, long) const</CODE> is mangled
56 c8661ffa 2005-11-29 devnull as <SAMP>`bar__C3Fooil'</SAMP>.
60 c8661ffa 2005-11-29 devnull For a constructor, the method name is left out.
61 c8661ffa 2005-11-29 devnull That is <CODE>Foo::Foo(int, long) const</CODE> is mangled
62 c8661ffa 2005-11-29 devnull as <SAMP>`__C3Fooil'</SAMP>.
66 c8661ffa 2005-11-29 devnull GNU Java does the same.
71 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC22" HREF="gxxint_toc.html#TOC22">Primitive types</A></H3>
74 c8661ffa 2005-11-29 devnull The C++ types <CODE>int</CODE>, <CODE>long</CODE>, <CODE>short</CODE>, <CODE>char</CODE>,
75 c8661ffa 2005-11-29 devnull and <CODE>long long</CODE> are mangled as <SAMP>`i'</SAMP>, <SAMP>`l'</SAMP>,
76 c8661ffa 2005-11-29 devnull <SAMP>`s'</SAMP>, <SAMP>`c'</SAMP>, and <SAMP>`x'</SAMP>, respectively.
77 c8661ffa 2005-11-29 devnull The corresponding unsigned types have <SAMP>`U'</SAMP> prefixed
78 c8661ffa 2005-11-29 devnull to the mangling. The type <CODE>signed char</CODE> is mangled <SAMP>`Sc'</SAMP>.
82 c8661ffa 2005-11-29 devnull The C++ and Java floating-point types <CODE>float</CODE> and <CODE>double</CODE>
83 c8661ffa 2005-11-29 devnull are mangled as <SAMP>`f'</SAMP> and <SAMP>`d'</SAMP> respectively.
87 c8661ffa 2005-11-29 devnull The C++ <CODE>bool</CODE> type and the Java <CODE>boolean</CODE> type are
88 c8661ffa 2005-11-29 devnull mangled as <SAMP>`b'</SAMP>.
92 c8661ffa 2005-11-29 devnull The C++ <CODE>wchar_t</CODE> and the Java <CODE>char</CODE> types are
93 c8661ffa 2005-11-29 devnull mangled as <SAMP>`w'</SAMP>.
97 c8661ffa 2005-11-29 devnull The Java integral types <CODE>byte</CODE>, <CODE>short</CODE>, <CODE>int</CODE>
98 c8661ffa 2005-11-29 devnull and <CODE>long</CODE> are mangled as <SAMP>`c'</SAMP>, <SAMP>`s'</SAMP>, <SAMP>`i'</SAMP>,
99 c8661ffa 2005-11-29 devnull and <SAMP>`x'</SAMP>, respectively.
103 c8661ffa 2005-11-29 devnull C++ code that has included <CODE>javatypes.h</CODE> will mangle
104 c8661ffa 2005-11-29 devnull the typedefs <CODE>jbyte</CODE>, <CODE>jshort</CODE>, <CODE>jint</CODE>
105 c8661ffa 2005-11-29 devnull and <CODE>jlong</CODE> as respectively <SAMP>`c'</SAMP>, <SAMP>`s'</SAMP>, <SAMP>`i'</SAMP>,
106 c8661ffa 2005-11-29 devnull and <SAMP>`x'</SAMP>. (This has not been implemented yet.)
111 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC23" HREF="gxxint_toc.html#TOC23">Mangling of simple names</A></H3>
114 c8661ffa 2005-11-29 devnull A simple class, package, template, or namespace name is
115 c8661ffa 2005-11-29 devnull encoded as the number of characters in the name, followed by
116 c8661ffa 2005-11-29 devnull the actual characters. Thus the class <CODE>Foo</CODE>
117 c8661ffa 2005-11-29 devnull is encoded as <SAMP>`3Foo'</SAMP>.
121 c8661ffa 2005-11-29 devnull If any of the characters in the name are not alphanumeric
122 c8661ffa 2005-11-29 devnull (i.e not one of the standard ASCII letters, digits, or '_'),
123 c8661ffa 2005-11-29 devnull or the initial character is a digit, then the name is
124 c8661ffa 2005-11-29 devnull mangled as a sequence of encoded Unicode letters.
125 c8661ffa 2005-11-29 devnull A Unicode encoding starts with a <SAMP>`U'</SAMP> to indicate
126 c8661ffa 2005-11-29 devnull that Unicode escapes are used, followed by the number of
127 c8661ffa 2005-11-29 devnull bytes used by the Unicode encoding, followed by the bytes
128 c8661ffa 2005-11-29 devnull representing the encoding. ASSCI letters and
129 c8661ffa 2005-11-29 devnull non-initial digits are encoded without change. However, all
130 c8661ffa 2005-11-29 devnull other characters (including underscore and initial digits) are
131 c8661ffa 2005-11-29 devnull translated into a sequence starting with an underscore,
132 c8661ffa 2005-11-29 devnull followed by the big-endian 4-hex-digit lower-case encoding of the character.
136 c8661ffa 2005-11-29 devnull If a method name contains Unicode-escaped characters, the
137 c8661ffa 2005-11-29 devnull entire mangled method name is followed by a <SAMP>`U'</SAMP>.
141 c8661ffa 2005-11-29 devnull For example, the method <CODE>X\u0319::M\u002B(int)</CODE> is encoded as
142 c8661ffa 2005-11-29 devnull <SAMP>`M_002b__U6X_0319iU'</SAMP>.
147 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC24" HREF="gxxint_toc.html#TOC24">Pointer and reference types</A></H3>
150 c8661ffa 2005-11-29 devnull A C++ pointer type is mangled as <SAMP>`P'</SAMP> followed by the
151 c8661ffa 2005-11-29 devnull mangling of the type pointed to.
155 c8661ffa 2005-11-29 devnull A C++ reference type as mangled as <SAMP>`R'</SAMP> followed by the
156 c8661ffa 2005-11-29 devnull mangling of the type referenced.
160 c8661ffa 2005-11-29 devnull A Java object reference type is equivalent
161 c8661ffa 2005-11-29 devnull to a C++ pointer parameter, so we mangle such an parameter type
162 c8661ffa 2005-11-29 devnull as <SAMP>`P'</SAMP> followed by the mangling of the class name.
167 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC25" HREF="gxxint_toc.html#TOC25">Qualified names</A></H3>
170 c8661ffa 2005-11-29 devnull Both C++ and Java allow a class to be lexically nested inside another
171 c8661ffa 2005-11-29 devnull class. C++ also supports namespaces (not yet implemented by G++).
172 c8661ffa 2005-11-29 devnull Java also supports packages.
176 c8661ffa 2005-11-29 devnull These are all mangled the same way: First the letter <SAMP>`Q'</SAMP>
177 c8661ffa 2005-11-29 devnull indicates that we are emitting a qualified name.
178 c8661ffa 2005-11-29 devnull That is followed by the number of parts in the qualified name.
179 c8661ffa 2005-11-29 devnull If that number is 9 or less, it is emitted with no delimiters.
180 c8661ffa 2005-11-29 devnull Otherwise, an underscore is written before and after the count.
181 c8661ffa 2005-11-29 devnull Then follows each part of the qualified name, as described above.
185 c8661ffa 2005-11-29 devnull For example <CODE>Foo::\u0319::Bar</CODE> is encoded as
186 c8661ffa 2005-11-29 devnull <SAMP>`Q33FooU5_03193Bar'</SAMP>.
191 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC26" HREF="gxxint_toc.html#TOC26">Templates</A></H3>
194 c8661ffa 2005-11-29 devnull A class template instantiation is encoded as the letter <SAMP>`t'</SAMP>,
195 c8661ffa 2005-11-29 devnull followed by the encoding of the template name, followed
196 c8661ffa 2005-11-29 devnull the number of template parameters, followed by encoding of the template
197 c8661ffa 2005-11-29 devnull parameters. If a template parameter is a type, it is written
198 c8661ffa 2005-11-29 devnull as a <SAMP>`Z'</SAMP> followed by the encoding of the type.
202 c8661ffa 2005-11-29 devnull A function template specialization (either an instantiation or an
203 c8661ffa 2005-11-29 devnull explicit specialization) is encoded by an <SAMP>`H'</SAMP> followed by the
204 c8661ffa 2005-11-29 devnull encoding of the template parameters, as described above, followed by
205 c8661ffa 2005-11-29 devnull an <SAMP>`_'</SAMP>, the encoding of the argument types template function (not the
206 c8661ffa 2005-11-29 devnull specialization), another <SAMP>`_'</SAMP>, and the return type. (Like the
207 c8661ffa 2005-11-29 devnull argument types, the return type is the return type of the function
208 c8661ffa 2005-11-29 devnull template, not the specialization.) Template parameters in the argument
209 c8661ffa 2005-11-29 devnull and return types are encoded by an <SAMP>`X'</SAMP> for type parameters, or a
210 c8661ffa 2005-11-29 devnull <SAMP>`Y'</SAMP> for constant parameters, and an index indicating their position
211 c8661ffa 2005-11-29 devnull in the template parameter list declaration.
216 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC27" HREF="gxxint_toc.html#TOC27">Arrays</A></H3>
219 c8661ffa 2005-11-29 devnull C++ array types are mangled by emitting <SAMP>`A'</SAMP>, followed by
220 c8661ffa 2005-11-29 devnull the length of the array, followed by an <SAMP>`_'</SAMP>, followed by
221 c8661ffa 2005-11-29 devnull the mangling of the element type. Of course, normally
222 c8661ffa 2005-11-29 devnull array parameter types decay into a pointer types, so you
223 c8661ffa 2005-11-29 devnull don't see this.
227 c8661ffa 2005-11-29 devnull Java arrays are objects. A Java type <CODE>T[]</CODE> is mangled
228 c8661ffa 2005-11-29 devnull as if it were the C++ type <CODE>JArray<T></CODE>.
229 c8661ffa 2005-11-29 devnull For example <CODE>java.lang.String[]</CODE> is encoded as
230 c8661ffa 2005-11-29 devnull <SAMP>`Pt6JArray1ZPQ34java4lang6String'</SAMP>.
235 c8661ffa 2005-11-29 devnull <H3><A NAME="SEC28" HREF="gxxint_toc.html#TOC28">Table of demangling code characters</A></H3>
238 c8661ffa 2005-11-29 devnull The following special characters are used in mangling:
241 c8661ffa 2005-11-29 devnull <DL COMPACT>
243 c8661ffa 2005-11-29 devnull <DT><SAMP>`A'</SAMP>
245 c8661ffa 2005-11-29 devnull Indicates a C++ array type.
247 c8661ffa 2005-11-29 devnull <DT><SAMP>`b'</SAMP>
249 c8661ffa 2005-11-29 devnull Encodes the C++ <CODE>bool</CODE> type,
250 c8661ffa 2005-11-29 devnull and the Java <CODE>boolean</CODE> type.
252 c8661ffa 2005-11-29 devnull <DT><SAMP>`c'</SAMP>
254 c8661ffa 2005-11-29 devnull Encodes the C++ <CODE>char</CODE> type, and the Java <CODE>byte</CODE> type.
256 c8661ffa 2005-11-29 devnull <DT><SAMP>`C'</SAMP>
258 c8661ffa 2005-11-29 devnull A modifier to indicate a <CODE>const</CODE> type.
259 c8661ffa 2005-11-29 devnull Also used to indicate a <CODE>const</CODE> member function
260 c8661ffa 2005-11-29 devnull (in which cases it precedes the encoding of the method's class).
262 c8661ffa 2005-11-29 devnull <DT><SAMP>`d'</SAMP>
264 c8661ffa 2005-11-29 devnull Encodes the C++ and Java <CODE>double</CODE> types.
266 c8661ffa 2005-11-29 devnull <DT><SAMP>`e'</SAMP>
268 c8661ffa 2005-11-29 devnull Indicates extra unknown arguments <CODE>...</CODE>.
270 c8661ffa 2005-11-29 devnull <DT><SAMP>`f'</SAMP>
272 c8661ffa 2005-11-29 devnull Encodes the C++ and Java <CODE>float</CODE> types.
274 c8661ffa 2005-11-29 devnull <DT><SAMP>`F'</SAMP>
276 c8661ffa 2005-11-29 devnull Used to indicate a function type.
278 c8661ffa 2005-11-29 devnull <DT><SAMP>`H'</SAMP>
280 c8661ffa 2005-11-29 devnull Used to indicate a template function.
282 c8661ffa 2005-11-29 devnull <DT><SAMP>`i'</SAMP>
284 c8661ffa 2005-11-29 devnull Encodes the C++ and Java <CODE>int</CODE> types.
286 c8661ffa 2005-11-29 devnull <DT><SAMP>`J'</SAMP>
288 c8661ffa 2005-11-29 devnull Indicates a complex type.
290 c8661ffa 2005-11-29 devnull <DT><SAMP>`l'</SAMP>
292 c8661ffa 2005-11-29 devnull Encodes the C++ <CODE>long</CODE> type.
294 c8661ffa 2005-11-29 devnull <DT><SAMP>`P'</SAMP>
296 c8661ffa 2005-11-29 devnull Indicates a pointer type. Followed by the type pointed to.
298 c8661ffa 2005-11-29 devnull <DT><SAMP>`Q'</SAMP>
300 c8661ffa 2005-11-29 devnull Used to mangle qualified names, which arise from nested classes.
301 c8661ffa 2005-11-29 devnull Should also be used for namespaces (?).
302 c8661ffa 2005-11-29 devnull In Java used to mangle package-qualified names, and inner classes.
304 c8661ffa 2005-11-29 devnull <DT><SAMP>`r'</SAMP>
306 c8661ffa 2005-11-29 devnull Encodes the GNU C++ <CODE>long double</CODE> type.
308 c8661ffa 2005-11-29 devnull <DT><SAMP>`R'</SAMP>
310 c8661ffa 2005-11-29 devnull Indicates a reference type. Followed by the referenced type.
312 c8661ffa 2005-11-29 devnull <DT><SAMP>`s'</SAMP>
314 c8661ffa 2005-11-29 devnull Encodes the C++ and java <CODE>short</CODE> types.
316 c8661ffa 2005-11-29 devnull <DT><SAMP>`S'</SAMP>
318 c8661ffa 2005-11-29 devnull A modifier that indicates that the following integer type is signed.
319 c8661ffa 2005-11-29 devnull Only used with <CODE>char</CODE>.
321 c8661ffa 2005-11-29 devnull Also used as a modifier to indicate a static member function.
323 c8661ffa 2005-11-29 devnull <DT><SAMP>`t'</SAMP>
325 c8661ffa 2005-11-29 devnull Indicates a template instantiation.
327 c8661ffa 2005-11-29 devnull <DT><SAMP>`T'</SAMP>
329 c8661ffa 2005-11-29 devnull A back reference to a previously seen type.
331 c8661ffa 2005-11-29 devnull <DT><SAMP>`U'</SAMP>
333 c8661ffa 2005-11-29 devnull A modifier that indicates that the following integer type is unsigned.
334 c8661ffa 2005-11-29 devnull Also used to indicate that the following class or namespace name
335 c8661ffa 2005-11-29 devnull is encoded using Unicode-mangling.
337 c8661ffa 2005-11-29 devnull <DT><SAMP>`v'</SAMP>
339 c8661ffa 2005-11-29 devnull Encodes the C++ and Java <CODE>void</CODE> types.
341 c8661ffa 2005-11-29 devnull <DT><SAMP>`V'</SAMP>
343 c8661ffa 2005-11-29 devnull A modified for a <CODE>const</CODE> type or method.
345 c8661ffa 2005-11-29 devnull <DT><SAMP>`w'</SAMP>
347 c8661ffa 2005-11-29 devnull Encodes the C++ <CODE>wchar_t</CODE> type, and the Java <CODE>char</CODE> types.
349 c8661ffa 2005-11-29 devnull <DT><SAMP>`x'</SAMP>
351 c8661ffa 2005-11-29 devnull Encodes the GNU C++ <CODE>long long</CODE> type, and the Java <CODE>long</CODE> type.
353 c8661ffa 2005-11-29 devnull <DT><SAMP>`X'</SAMP>
355 c8661ffa 2005-11-29 devnull Encodes a template type parameter, when part of a function type.
357 c8661ffa 2005-11-29 devnull <DT><SAMP>`Y'</SAMP>
359 c8661ffa 2005-11-29 devnull Encodes a template constant parameter, when part of a function type.
361 c8661ffa 2005-11-29 devnull <DT><SAMP>`Z'</SAMP>
363 c8661ffa 2005-11-29 devnull Used for template type parameters.
368 c8661ffa 2005-11-29 devnull The letters <SAMP>`G'</SAMP>, <SAMP>`M'</SAMP>, <SAMP>`O'</SAMP>, and <SAMP>`p'</SAMP>
369 c8661ffa 2005-11-29 devnull also seem to be used for obscure purposes ...
372 c8661ffa 2005-11-29 devnull <P><HR><P>
373 c8661ffa 2005-11-29 devnull Go to the <A HREF="gxxint_1.html">first</A>, <A HREF="gxxint_14.html">previous</A>, <A HREF="gxxint_16.html">next</A>, <A HREF="gxxint_16.html">last</A> section, <A HREF="gxxint_toc.html">table of contents</A>.
