gcc/doc/cpp/implementation-defined-behavior.rst

   1 ..
   2   Copyright 1988-2022 Free Software Foundation, Inc.
   3   This is part of the GCC manual.
   4   For copying conditions, see the copyright.rst file.
   5
   6 .. _implementation-defined-behavior:
   7
   8 .. _identifier-characters:
   9
  10 Implementation-defined behavior
  11 *******************************
  12
  13 This is how CPP behaves in all the cases which the C standard
  14 describes as :dfn:`implementation-defined`.  This term means that the
  15 implementation is free to do what it likes, but must document its choice
  16 and stick to it.
  17
  18 .. todo:: Check the C++ standard for more implementation-defined stuff.
  19
  20 * The mapping of physical source file multi-byte characters to the
  21   execution character set.
  22
  23   The input character set can be specified using the
  24   :option:`-finput-charset` option, while the execution character set may
  25   be controlled using the :option:`-fexec-charset` and
  26   :option:`-fwide-exec-charset` options.
  27
  28 * Identifier characters.
  29
  30   The C and C++ standards allow identifiers to be composed of :samp:`_`
  31   and the alphanumeric characters.  C++ also allows universal character
  32   names.  C99 and later C standards permit both universal character
  33   names and implementation-defined characters.  In both C and C++ modes,
  34   GCC accepts in identifiers exactly those extended characters that
  35   correspond to universal character names permitted by the chosen
  36   standard.
  37
  38   GCC allows the :samp:`$` character in identifiers as an extension for
  39   most targets.  This is true regardless of the std= switch,
  40   since this extension cannot conflict with standards-conforming
  41   programs.  When preprocessing assembler, however, dollars are not
  42   identifier characters by default.
  43
  44   Currently the targets that by default do not permit :samp:`$` are AVR,
  45   IP2K, MMIX, MIPS Irix 3, ARM aout, and PowerPC targets for the AIX
  46   operating system.
  47
  48   You can override the default with :option:`-fdollars-in-identifiers` or
  49   :option:`-fno-dollars-in-identifiers`.  See :option:`-fdollars-in-identifiers`.
  50
  51 * Non-empty sequences of whitespace characters.
  52
  53   In textual output, each whitespace sequence is collapsed to a single
  54   space.  For aesthetic reasons, the first token on each non-directive
  55   line of output is preceded with sufficient spaces that it appears in the
  56   same column as it did in the original source file.
  57
  58 * The numeric value of character constants in preprocessor expressions.
  59
  60   The preprocessor and compiler interpret character constants in the
  61   same way; i.e. escape sequences such as :samp:`\\a` are given the
  62   values they would have on the target machine.
  63
  64   The compiler evaluates a multi-character character constant a character
  65   at a time, shifting the previous value left by the number of bits per
  66   target character, and then or-ing in the bit-pattern of the new
  67   character truncated to the width of a target character.  The final
  68   bit-pattern is given type ``int``, and is therefore signed,
  69   regardless of whether single characters are signed or not.
  70   If there are more
  71   characters in the constant than would fit in the target ``int`` the
  72   compiler issues a warning, and the excess leading characters are
  73   ignored.
  74
  75   For example, ``'ab'`` for a target with an 8-bit ``char`` would be
  76   interpreted as :samp:`(int) ((unsigned char) 'a' * 256 + (unsigned char)
  77   'b')`, and ``'\234a'`` as :samp:`(int) ((unsigned char) '\\234' *
  78   256 + (unsigned char) 'a')`.
  79
  80 * Source file inclusion.
  81
  82   For a discussion on how the preprocessor locates header files,
  83   :ref:`include-operation`.
  84
  85 * Interpretation of the filename resulting from a macro-expanded
  86   :samp:`#include` directive.
  87
  88   See :ref:`computed-includes`.
  89
  90 * Treatment of a :samp:`#pragma` directive that after macro-expansion
  91   results in a standard pragma.
  92
  93   No macro expansion occurs on any :samp:`#pragma` directive line, so the
  94   question does not arise.
  95
  96   Note that GCC does not yet implement any of the standard
  97   pragmas.