mingw/info/stabs/Builtin-Type-Descriptors.html

   1 <html lang="en">
   2 <head>
   3 <title>STABS</title>
   4 <meta http-equiv="Content-Type" content="text/html">
   5 <meta name="description" content="STABS">
   6 <meta name="generator" content="makeinfo 4.3">
   7 <link href="http://www.gnu.org/software/texinfo/" rel="generator-home">
   8 </head>
   9 <body>
  10 <div class="node">
  11 <p>
  12 Node:<a name="Builtin%20Type%20Descriptors">Builtin Type Descriptors</a>,
  13 Next:<a rel="next" accesskey="n" href="Negative-Type-Numbers.html#Negative%20Type%20Numbers">Negative Type Numbers</a>,
  14 Previous:<a rel="previous" accesskey="p" href="Traditional-Builtin-Types.html#Traditional%20Builtin%20Types">Traditional Builtin Types</a>,
  15 Up:<a rel="up" accesskey="u" href="Builtin-Types.html#Builtin%20Types">Builtin Types</a>
  16 <hr><br>
  17 </div>
  18
  19 <h4 class="subsection">Defining Builtin Types Using Builtin Type Descriptors</h4>
  20
  21    <p>This is the method used by Sun's <code>acc</code> for defining builtin types.
  22 These are the type descriptors to define builtin types:
  23
  24      <dl>
  25 <dt><code>b </code><var>signed</var><code> </code><var>char-flag</var><code> </code><var>width</var><code> ; </code><var>offset</var><code> ; </code><var>nbits</var><code> ;</code>
  26      <dd>Define an integral type.  <var>signed</var> is <code>u</code> for unsigned or
  27 <code>s</code> for signed.  <var>char-flag</var> is <code>c</code> which indicates this
  28 is a character type, or is omitted.  I assume this is to distinguish an
  29 integral type from a character type of the same size, for example it
  30 might make sense to set it for the C type <code>wchar_t</code> so the debugger
  31 can print such variables differently (Solaris does not do this).  Sun
  32 sets it on the C types <code>signed char</code> and <code>unsigned char</code> which
  33 arguably is wrong.  <var>width</var> and <var>offset</var> appear to be for small
  34 objects stored in larger ones, for example a <code>short</code> in an
  35 <code>int</code> register.  <var>width</var> is normally the number of bytes in the
  36 type.  <var>offset</var> seems to always be zero.  <var>nbits</var> is the number
  37 of bits in the type.
  38
  39      <p>Note that type descriptor <code>b</code> used for builtin types conflicts with
  40 its use for Pascal space types (see <a href="Miscellaneous-Types.html#Miscellaneous%20Types">Miscellaneous Types</a>); they can
  41 be distinguished because the character following the type descriptor
  42 will be a digit, <code>(</code>, or <code>-</code> for a Pascal space type, or
  43 <code>u</code> or <code>s</code> for a builtin type.
  44
  45      <br><dt><code>w</code>
  46      <dd>Documented by AIX to define a wide character type, but their compiler
  47 actually uses negative type numbers (see <a href="Negative-Type-Numbers.html#Negative%20Type%20Numbers">Negative Type Numbers</a>).
  48
  49      <br><dt><code>R </code><var>fp-type</var><code> ; </code><var>bytes</var><code> ;</code>
  50      <dd>Define a floating point type.  <var>fp-type</var> has one of the following values:
  51
  52           <dl>
  53 <dt><code>1 (NF_SINGLE)</code>
  54           <dd>IEEE 32-bit (single precision) floating point format.
  55
  56           <br><dt><code>2 (NF_DOUBLE)</code>
  57           <dd>IEEE 64-bit (double precision) floating point format.
  58
  59           <br><dt><code>3 (NF_COMPLEX)</code>
  60           <dd><br><dt><code>4 (NF_COMPLEX16)</code>
  61           <dd><br><dt><code>5 (NF_COMPLEX32)</code>
  62           <dd>These are for complex numbers.  A comment in the GDB source describes
  63 them as Fortran <code>complex</code>, <code>double complex</code>, and
  64 <code>complex*16</code>, respectively, but what does that mean?  (i.e., Single
  65 precision?  Double precision?).
  66
  67           <br><dt><code>6 (NF_LDOUBLE)</code>
  68           <dd>Long double.  This should probably only be used for Sun format
  69 <code>long double</code>, and new codes should be used for other floating
  70 point formats (<code>NF_DOUBLE</code> can be used if a <code>long double</code> is
  71 really just an IEEE double, of course).
  72 </dl>
  73
  74      <p><var>bytes</var> is the number of bytes occupied by the type.  This allows a
  75 debugger to perform some operations with the type even if it doesn't
  76 understand <var>fp-type</var>.
  77
  78      <br><dt><code>g </code><var>type-information</var><code> ; </code><var>nbits</var><code></code>
  79      <dd>Documented by AIX to define a floating type, but their compiler actually
  80 uses negative type numbers (see <a href="Negative-Type-Numbers.html#Negative%20Type%20Numbers">Negative Type Numbers</a>).
  81
  82      <br><dt><code>c </code><var>type-information</var><code> ; </code><var>nbits</var><code></code>
  83      <dd>Documented by AIX to define a complex type, but their compiler actually
  84 uses negative type numbers (see <a href="Negative-Type-Numbers.html#Negative%20Type%20Numbers">Negative Type Numbers</a>).
  85 </dl>
  86
  87    <p>The C <code>void</code> type is defined as a signed integral type 0 bits long:
  88 <pre class="example">     .stabs "void:t19=bs0;0;0",128,0,0,0
  89      </pre>
  90    The Solaris compiler seems to omit the trailing semicolon in this case.
  91 Getting sloppy in this way is not a swift move because if a type is
  92 embedded in a more complex expression it is necessary to be able to tell
  93 where it ends.
  94
  95    <p>I'm not sure how a boolean type is represented.
  96
  97    </body></html>
  98