path: root/docs/users_guide/bugs.xml

<?xml version="1.0" encoding="iso-8859-1"?>
<chapter id="bugs-and-infelicities">
  <title>Known bugs and infelicities</title>

  <sect1 id="vs-Haskell-defn">
    <title>Haskell&nbsp;standards vs.&nbsp;Glasgow Haskell: language non-compliance
</title>

    <indexterm><primary>GHC vs the Haskell standards</primary></indexterm>
    <indexterm><primary>Haskell standards vs GHC</primary></indexterm>

    <para>
        This section lists Glasgow Haskell infelicities in its
        implementation of Haskell&nbsp;98 and Haskell&nbsp;2010.
        See also the &ldquo;when things go wrong&rdquo; section
        (<xref linkend="wrong"/>) for information about crashes,
        space leaks, and other undesirable phenomena.
    </para>

    <para>
        The limitations here are listed in Haskell Report order
        (roughly).
    </para>

  <sect2 id="haskell-standards-divergence">
    <title>Divergence from Haskell&nbsp;98 and Haskell&nbsp;2010</title>

    <para>
        By default, GHC mainly aims to behave (mostly) like a Haskell&nbsp;2010
        compiler, although you can tell it to try to behave like a
        particular version of the language with the
        <literal>-XHaskell98</literal> and
        <literal>-XHaskell2010</literal> flags. The known deviations
        from the standards are described below. Unless otherwise stated,
        the deviation applies in Haskell&nbsp;98, Haskell&nbsp;2010 and
        the default modes.
    </para>

    <sect3 id="infelicities-lexical">
      <title>Lexical syntax</title>

      <itemizedlist>
	<listitem>
	  <para>Certain lexical rules regarding qualified identifiers
	  are slightly different in GHC compared to the Haskell
	  report.  When you have
	  <replaceable>module</replaceable><literal>.</literal><replaceable>reservedop</replaceable>,
	  such as <literal>M.\</literal>, GHC will interpret it as a
	  single qualified operator rather than the two lexemes
	  <literal>M</literal> and <literal>.\</literal>.</para>
	</listitem>
      </itemizedlist>
    </sect3>

      <sect3 id="infelicities-syntax">
	<title>Context-free syntax</title>

	<itemizedlist>
	  <listitem>
	    <para>In Haskell&nbsp;98 mode and by default (but not in
          Haskell&nbsp;2010 mode), GHC is a little less strict about the
          layout rule when used
	      in <literal>do</literal> expressions.  Specifically, the
	      restriction that "a nested context must be indented further to
	      the right than the enclosing context" is relaxed to allow the
	      nested context to be at the same level as the enclosing context,
	      if the enclosing context is a <literal>do</literal>
	      expression.</para>

	    <para>For example, the following code is accepted by GHC:

<programlisting>
main = do args &lt;- getArgs
	  if null args then return [] else do
          ps &lt;- mapM process args
          mapM print ps</programlisting>

        This behaviour is controlled by the
        <literal>NondecreasingIndentation</literal> extension.
	      </para>
	  </listitem>

	<listitem>
	  <para>GHC doesn't do the fixity resolution in expressions during
	  parsing as required by Haskell&nbsp;98 (but not by Haskell&nbsp;2010).
      For example, according to the Haskell&nbsp;98 report, the
	  following expression is legal:
<programlisting>
    let x = 42 in x == 42 == True</programlisting>
	and parses as:
<programlisting>
    (let x = 42 in x == 42) == True</programlisting>

          because according to the report, the <literal>let</literal>
	  expression <quote>extends as far to the right as
	  possible</quote>.  Since it can't extend past the second
	  equals sign without causing a parse error
	  (<literal>==</literal> is non-fix), the
	  <literal>let</literal>-expression must terminate there.  GHC
	  simply gobbles up the whole expression, parsing like this:
<programlisting>
    (let x = 42 in x == 42 == True)</programlisting></para>
	</listitem>
      </itemizedlist>
    </sect3>

  <sect3 id="infelicities-exprs-pats">
      <title>Expressions and patterns</title>

    <para>In its default mode, GHC makes some programs slightly more defined 
    than they should be. For example, consider
    <programlisting>
f :: [a] -> b -> b
f [] = error "urk"
f (x:xs) = \v -> v

main = print (f [] `seq` True)
    </programlisting>
This should call <literal>error</literal> but actually prints <literal>True</literal>.
Reason: GHC eta-expands <literal>f</literal> to
    <programlisting>
f :: [a] -> b -> b
f []     v = error "urk"
f (x:xs) v = v
    </programlisting>
This improves efficiency slightly but significantly for most programs, and
is bad for only a few.  To suppress this bogus "optimisation" use <option>-fpedantic-bottoms</option>.
</para>

    </sect3>

    <sect3 id="infelicities-decls">
      <title>Declarations and bindings</title>

      <para>In its default mode, GHC does not accept datatype contexts,
      as it has been decided to remove them from the next version of the
      language standard. This behaviour can be controlled with the
      <option>DatatypeContexts</option> extension.
      See <xref linkend="datatype-contexts" />.</para>
    </sect3>

      <sect3 id="infelicities-Modules">
	<title>Module system and interface files</title>

	<para>GHC requires the use of <literal>hs-boot</literal>
	  files to cut the recursive loops among mutually recursive modules
	  as described in <xref linkend="mutual-recursion"/>.  This more of an infelicity
	    than a bug: the Haskell Report says
	  (<ulink url="http://haskell.org/onlinereport/modules.html#sect5.7">Section 5.7</ulink>) "Depending on the Haskell
	implementation used, separate compilation of mutually
	recursive modules may require that imported modules contain
	additional information so that they may be referenced before
	they are compiled. Explicit type signatures for all exported
	values may be necessary to deal with mutual recursion. The
	precise details of separate compilation are not defined by
	this Report."

	</para>

    </sect3>

    <sect3 id="infelicities-numbers">
      <title>Numbers, basic types, and built-in classes</title>

      <variablelist>
	<varlistentry>
	  <term>Num superclasses</term>
	  <listitem>
	    <para>
              The <literal>Num</literal> class does not have
              <literal>Show</literal> or <literal>Eq</literal>
              superclasses.
            </para>

	    <para>
              You can make code that works with both
              Haskell98/Haskell2010 and GHC by:
              <itemizedlist>
                <listitem>
                  <para>
                    Whenever you make a <literal>Num</literal> instance
                    of a type, also make <literal>Show</literal> and
                    <literal>Eq</literal> instances, and
                  </para>
                </listitem>
                <listitem>
                  <para>
                    Whenever you give a function, instance or class a
                    <literal>Num t</literal> constraint, also give it
                    <literal>Show t</literal> and
                    <literal>Eq t</literal> constraints.
                  </para>
                </listitem>
              </itemizedlist>
            </para>
	    </listitem>
	  </varlistentry>

	<varlistentry>
	  <term>Bits superclasses</term>
	  <listitem>
	    <para>
              The <literal>Bits</literal> class does not have
              a <literal>Num</literal> superclasses. It therefore
              does not have default methods for the
              <literal>bit</literal>,
              <literal>testBit</literal> and
              <literal>popCount</literal> methods.
            </para>

	    <para>
              You can make code that works with both
              Haskell2010 and GHC by:
              <itemizedlist>
                <listitem>
                  <para>
                    Whenever you make a <literal>Bits</literal> instance
                    of a type, also make a <literal>Num</literal>
                    instance, and
                  </para>
                </listitem>
                <listitem>
                  <para>
                    Whenever you give a function, instance or class a
                    <literal>Bits t</literal> constraint, also give it
                    a <literal>Num t</literal> constraint, and
                  </para>
                </listitem>
                <listitem>
                  <para>
                    Always define the <literal>bit</literal>,
                    <literal>testBit</literal> and
                    <literal>popCount</literal> methods in
                    <literal>Bits</literal> instances.
                  </para>
                </listitem>
              </itemizedlist>
            </para>
	    </listitem>
	  </varlistentry>

	<varlistentry>
	  <term>Extra instances</term>
	  <listitem>
	    <para>
              The following extra instances are defined:
            </para>
<programlisting>
instance Functor ((->) r)
instance Monad ((->) r)
instance Functor ((,) a)
instance Functor (Either a)
instance Monad (Either e)
</programlisting>
	    </listitem>
	  </varlistentry>

	<varlistentry>
	  <term>Multiply-defined array elements&mdash;not checked:</term>
	  <listitem>
	    <para>This code fragment should
	    elicit a fatal error, but it does not:

<programlisting>
main = print (array (1,1) [(1,2), (1,3)])</programlisting>
GHC's implementation of <literal>array</literal> takes the value of an
array slot from the last (index,value) pair in the list, and does no
checking for duplicates.  The reason for this is efficiency, pure and simple.
            </para>
	  </listitem>
	</varlistentry>
      </variablelist>

    </sect3>

      <sect3 id="infelicities-Prelude">
	<title>In <literal>Prelude</literal> support</title>

      <variablelist>
	<varlistentry>
	  <term>Arbitrary-sized tuples</term>
	  <listitem>
	    <para>Tuples are currently limited to size 100.  HOWEVER:
            standard instances for tuples (<literal>Eq</literal>,
            <literal>Ord</literal>, <literal>Bounded</literal>,
            <literal>Ix</literal> <literal>Read</literal>, and
            <literal>Show</literal>) are available
            <emphasis>only</emphasis> up to 16-tuples.</para>

	    <para>This limitation is easily subvertible, so please ask
            if you get stuck on it.</para>
	    </listitem>
	  </varlistentry>

	  <varlistentry>
	    <term><literal>Read</literal>ing integers</term>
	    <listitem>
	      <para>GHC's implementation of the
	      <literal>Read</literal> class for integral types accepts
	      hexadecimal and octal literals (the code in the Haskell
	      98 report doesn't).  So, for example,
<programlisting>read "0xf00" :: Int</programlisting>
              works in GHC.</para>
	      <para>A possible reason for this is that <literal>readLitChar</literal> accepts hex and
		octal escapes, so it seems inconsistent not to do so for integers too.</para>
	    </listitem>
	  </varlistentry>

	  <varlistentry>
	    <term><literal>isAlpha</literal></term>
	    <listitem>
	      <para>The Haskell 98 definition of <literal>isAlpha</literal>
              is:</para>

<programlisting>isAlpha c = isUpper c || isLower c</programlisting>

	      <para>GHC's implementation diverges from the Haskell 98
              definition in the sense that Unicode alphabetic characters which
              are neither upper nor lower case will still be identified as
              alphabetic by <literal>isAlpha</literal>.</para>
	    </listitem>
	  </varlistentry>

          <varlistentry>
            <term><literal>hGetContents</literal></term>
            <listitem>
              <para>
                Lazy I/O throws an exception if an error is
                encountered, in contrast to the Haskell 98 spec which
                requires that errors are discarded (see Section 21.2.2
                of the Haskell 98 report).  The exception thrown is
                the usual IO exception that would be thrown if the
                failing IO operation was performed in the IO monad, and can
                be caught by <literal>System.IO.Error.catch</literal>
                or <literal>Control.Exception.catch</literal>.
              </para>
            </listitem>
          </varlistentry>
	</variablelist>
    </sect3>

    <sect3 id="infelicities-ffi">
      <title>The Foreign Function Interface</title>
      <variablelist>
      <varlistentry>
        <term><literal>hs_init()</literal> not allowed
        after <literal>hs_exit()</literal></term>
        <listitem>
          <para>The FFI spec requires the implementation to support
            re-initialising itself after being shut down
            with <literal>hs_exit()</literal>, but GHC does not
            currently support that.</para>
        </listitem>
      </varlistentry>
    </variablelist>
    </sect3>

  </sect2>

  <sect2 id="haskell-98-2010-undefined">
    <title>GHC's interpretation of undefined behaviour in
    Haskell&nbsp;98 and Haskell&nbsp;2010</title>

    <para>This section documents GHC's take on various issues that are
    left undefined or implementation specific in Haskell 98.</para>

    <variablelist>
      <varlistentry>
	<term>
          The <literal>Char</literal> type
          <indexterm><primary><literal>Char</literal></primary><secondary>size of</secondary></indexterm>
        </term>
	<listitem>
	  <para>Following the ISO-10646 standard,
	  <literal>maxBound :: Char</literal> in GHC is
	  <literal>0x10FFFF</literal>.</para>
	</listitem>
      </varlistentry>

      <varlistentry>
	<term>
          Sized integral types
          <indexterm><primary><literal>Int</literal></primary><secondary>size of</secondary></indexterm>
	</term>
	<listitem>
	  <para>In GHC the <literal>Int</literal> type follows the
	  size of an address on the host architecture; in other words
	  it holds 32 bits on a 32-bit machine, and 64-bits on a
	  64-bit machine.</para>

	  <para>Arithmetic on <literal>Int</literal> is unchecked for
	  overflow<indexterm><primary>overflow</primary><secondary><literal>Int</literal></secondary>
	    </indexterm>, so all operations on <literal>Int</literal> happen
	  modulo
	  2<superscript><replaceable>n</replaceable></superscript>
	  where <replaceable>n</replaceable> is the size in bits of
	  the <literal>Int</literal> type.</para>

	  <para>The <literal>fromInteger</literal><indexterm><primary><literal>fromInteger</literal></primary>
	    </indexterm> function (and hence
	  also <literal>fromIntegral</literal><indexterm><primary><literal>fromIntegral</literal></primary>
	    </indexterm>) is a special case when
	  converting to <literal>Int</literal>.  The value of
	  <literal>fromIntegral x :: Int</literal> is given by taking
	  the lower <replaceable>n</replaceable> bits of <literal>(abs
	  x)</literal>, multiplied by the sign of <literal>x</literal>
	  (in 2's complement <replaceable>n</replaceable>-bit
	  arithmetic).  This behaviour was chosen so that for example
	  writing <literal>0xffffffff :: Int</literal> preserves the
	  bit-pattern in the resulting <literal>Int</literal>.</para>


	   <para>Negative literals, such as <literal>-3</literal>, are
             specified by (a careful reading of) the Haskell Report as
             meaning <literal>Prelude.negate (Prelude.fromInteger 3)</literal>.
	     So <literal>-2147483648</literal> means <literal>negate (fromInteger 2147483648)</literal>.
	     Since <literal>fromInteger</literal> takes the lower 32 bits of the representation,
	     <literal>fromInteger (2147483648::Integer)</literal>, computed at type <literal>Int</literal> is
	     <literal>-2147483648::Int</literal>.  The <literal>negate</literal> operation then
	     overflows, but it is unchecked, so <literal>negate (-2147483648::Int)</literal> is just
	     <literal>-2147483648</literal>.  In short, one can write <literal>minBound::Int</literal> as
	     a literal with the expected meaning (but that is not in general guaranteed).
             </para>

	  <para>The <literal>fromIntegral</literal> function also
	  preserves bit-patterns when converting between the sized
	  integral types (<literal>Int8</literal>,
	  <literal>Int16</literal>, <literal>Int32</literal>,
	  <literal>Int64</literal> and the unsigned
	  <literal>Word</literal> variants), see the modules
	  <literal>Data.Int</literal> and <literal>Data.Word</literal>
	  in the library documentation.</para>
	</listitem>
      </varlistentry>

      <varlistentry>
	<term>Unchecked float arithmetic</term>
	<listitem>
	  <para>Operations on <literal>Float</literal> and
          <literal>Double</literal> numbers are
          <emphasis>unchecked</emphasis> for overflow, underflow, and
          other sad occurrences.  (note, however, that some
          architectures trap floating-point overflow and
          loss-of-precision and report a floating-point exception,
          probably terminating the
          program)<indexterm><primary>floating-point
          exceptions</primary></indexterm>.</para>
	</listitem>
      </varlistentry>
    </variablelist>
  </sect2>

  </sect1>


  <sect1 id="bugs">
    <title>Known bugs or infelicities</title>

    <para>The bug tracker lists bugs that have been reported in GHC but not
      yet fixed: see the <ulink url="http://ghc.haskell.org/trac/ghc/">GHC Trac</ulink>.  In addition to those, GHC also has the following known bugs
      or  infelicities.  These bugs are more permanent; it is unlikely that
      any of them will be fixed in the short term.</para>

  <sect2 id="bugs-ghc">
    <title>Bugs in GHC</title>

    <itemizedlist>
      <listitem>
	<para> GHC can warn about non-exhaustive or overlapping
        patterns (see <xref linkend="options-sanity"/>), and usually
        does so correctly.  But not always.  It gets confused by
        string patterns, and by guards, and can then emit bogus
        warnings.  The entire overlap-check code needs an overhaul
        really.</para>
      </listitem>

      <listitem>
	<para>GHC does not allow you to have a data type with a context
	   that mentions type variables that are not data type parameters.
	  For example:
<programlisting>
  data C a b => T a = MkT a
</programlisting>
	  so that <literal>MkT</literal>'s type is
<programlisting>
  MkT :: forall a b. C a b => a -> T a
</programlisting>
        In principle, with a suitable class declaration with a functional dependency,
	 it's possible that this type is not ambiguous; but GHC nevertheless rejects
	  it.  The type variables mentioned in the context of the data type declaration must
	be among the type parameters of the data type.</para>
      </listitem>

      <listitem>
	<para>GHC's inliner can be persuaded into non-termination
        using the standard way to encode recursion via a data type:</para>
<programlisting>
  data U = MkU (U -> Bool)

  russel :: U -> Bool
  russel u@(MkU p) = not $ p u

  x :: Bool
  x = russel (MkU russel)
</programlisting>

        <para>We have never found another class of programs, other
        than this contrived one, that makes GHC diverge, and fixing
        the problem would impose an extra overhead on every
        compilation.  So the bug remains un-fixed.  There is more
        background in <ulink
        url="http://research.microsoft.com/~simonpj/Papers/inlining/">
        Secrets of the GHC inliner</ulink>.</para>
      </listitem>

      <listitem>
        <para>On 32-bit x86 platforms when using the native code
        generator, the
        <option>-fexcess-precision</option><indexterm><primary><option>-fexcess-precision</option></primary></indexterm> option
        is always on.  This means that floating-point calculations are
        non-deterministic, because depending on how the program is
        compiled (optimisation settings, for example), certain
        calculations might be done at 80-bit precision instead of the
        intended 32-bit or 64-bit precision.  Floating-point results
        may differ when optimisation is turned on.  In the worst case,
        referential transparency is violated, because for example
        <literal>let x = E1 in E2</literal> can evaluate to a
        different value than <literal>E2[E1/x]</literal>.</para>

        <para>
          One workaround is to use the
          <option>-msse2</option><indexterm><primary><option>-msse2</option></primary></indexterm>
          option (see <xref linkend="options-platform" />, which
          generates code to use the SSE2 instruction set instead of
          the x87 instruction set.  SSE2 code uses the correct
          precision for all floating-point operations, and so gives
          deterministic results.  However, note that this only works
          with processors that support SSE2 (Intel Pentium 4 or AMD
          Athlon 64 and later), which is why the option is not enabled
          by default.  The libraries that come with GHC are probably
          built without this option, unless you built GHC yourself.
        </para>
      </listitem>

    </itemizedlist>
  </sect2>

  <sect2 id="bugs-ghci">
    <title>Bugs in GHCi (the interactive GHC)</title>
    <itemizedlist>
      <listitem>
	<para>GHCi does not respect the <literal>default</literal>
        declaration in the module whose scope you are in.  Instead,
        for expressions typed at the command line, you always get the
        default default-type behaviour; that is,
        <literal>default(Int,Double)</literal>.</para>

	<para>It would be better for GHCi to record what the default
        settings in each module are, and use those of the 'current'
        module (whatever that is).</para>
      </listitem>

      <listitem>
      <para>On Windows, there's a GNU ld/BFD bug
      whereby it emits bogus PE object files that have more than
      0xffff relocations. When GHCi tries to load a package affected by this
      bug, you get an error message of the form
<screen>
Loading package javavm ... linking ... WARNING: Overflown relocation field (# relocs found: 30765)
</screen>
      The last time we looked, this bug still
      wasn't fixed in the BFD codebase, and there wasn't any
      noticeable interest in fixing it when we reported the bug
      back in 2001 or so.
      </para>
      <para>The workaround is to split up the .o files that make up
      your package into two or more .o's, along the lines of
      how the "base" package does it.</para>
      </listitem>

      <listitem>
        <para>
          GHCi does not keep careful track of what instance
          declarations are 'in scope' if they come from other
          packages.  Instead, all instance declarations that GHC has
          seen in other packages are all available at the prompt,
          whether or not the instance really ought to be in visible
          given the current set of modules in scope.
        </para>
      </listitem>
    </itemizedlist>
  </sect2>
  </sect1>

</chapter>

<!-- Emacs stuff:
     ;;; Local Variables: ***
     ;;; sgml-parent-document: ("users_guide.xml" "book" "chapter") ***
     ;;; End: ***
 -->