1 <?xml version="1.0" encoding="iso-8859-1"?>
2 <chapter id="bugs-and-infelicities">
3 <title>Known bugs and infelicities</title>
5 <sect1 id="vs-Haskell-defn">
6 <title>Haskell 98 vs. Glasgow Haskell: language non-compliance
9 <indexterm><primary>GHC vs the Haskell 98 language</primary></indexterm>
10 <indexterm><primary>Haskell 98 language vs GHC</primary></indexterm>
12 <para>This section lists Glasgow Haskell infelicities in its
13 implementation of Haskell 98. See also the “when things
14 go wrong” section (<xref linkend="wrong"/>) for information
15 about crashes, space leaks, and other undesirable phenomena.</para>
17 <para>The limitations here are listed in Haskell Report order
20 <sect2 id="haskell98-divergence">
21 <title>Divergence from Haskell 98</title>
24 <sect3 id="infelicities-lexical">
25 <title>Lexical syntax</title>
29 <para>The Haskell report specifies that programs may be
30 written using Unicode. GHC only accepts the ISO-8859-1
31 character set at the moment.</para>
35 <para>Certain lexical rules regarding qualified identifiers
36 are slightly different in GHC compared to the Haskell
38 <replaceable>module</replaceable><literal>.</literal><replaceable>reservedop</replaceable>,
39 such as <literal>M.\</literal>, GHC will interpret it as a
40 single qualified operator rather than the two lexemes
41 <literal>M</literal> and <literal>.\</literal>.</para>
46 <sect3 id="infelicities-syntax">
47 <title>Context-free syntax</title>
51 <para>GHC is a little less strict about the layout rule when used
52 in <literal>do</literal> expressions. Specifically, the
53 restriction that "a nested context must be indented further to
54 the right than the enclosing context" is relaxed to allow the
55 nested context to be at the same level as the enclosing context,
56 if the enclosing context is a <literal>do</literal>
59 <para>For example, the following code is accepted by GHC:
62 main = do args <- getArgs
63 if null args then return [] else do
64 ps <- mapM process args
65 mapM print ps</programlisting>
71 <para>GHC doesn't do fixity resolution in expressions during
72 parsing. For example, according to the Haskell report, the
73 following expression is legal Haskell:
75 let x = 42 in x == 42 == True</programlisting>
78 (let x = 42 in x == 42) == True</programlisting>
80 because according to the report, the <literal>let</literal>
81 expression <quote>extends as far to the right as
82 possible</quote>. Since it can't extend past the second
83 equals sign without causing a parse error
84 (<literal>==</literal> is non-fix), the
85 <literal>let</literal>-expression must terminate there. GHC
86 simply gobbles up the whole expression, parsing like this:
88 (let x = 42 in x == 42 == True)</programlisting>
90 The Haskell report is arguably wrong here, but nevertheless
91 it's a difference between GHC & Haskell 98.</para>
96 <sect3 id="infelicities-exprs-pats">
97 <title>Expressions and patterns</title>
99 <para>None known.</para>
102 <sect3 id="infelicities-decls">
103 <title>Declarations and bindings</title>
105 <para>GHC's typechecker makes all pattern bindings monomorphic
106 by default; this behaviour can be disabled with
107 <option>-fno-mono-pat-binds</option>. See <xref
108 linkend="options-language" />.</para>
111 <sect3 id="infelicities-Modules">
112 <title>Module system and interface files</title>
114 <para>None known.</para>
117 <sect3 id="infelicities-numbers">
118 <title>Numbers, basic types, and built-in classes</title>
122 <term>Multiply-defined array elements—not checked:</term>
124 <para>This code fragment should
125 elicit a fatal error, but it does not:
128 main = print (array (1,1) [(1,2), (1,3)])</programlisting>
129 GHC's implementation of <literal>array</literal> takes the value of an
130 array slot from the last (index,value) pair in the list, and does no
131 checking for duplicates. The reason for this is efficiency, pure and simple.
139 <sect3 id="infelicities-Prelude">
140 <title>In <literal>Prelude</literal> support</title>
144 <term>Arbitrary-sized tuples</term>
146 <para>Tuples are currently limited to size 100. HOWEVER:
147 standard instances for tuples (<literal>Eq</literal>,
148 <literal>Ord</literal>, <literal>Bounded</literal>,
149 <literal>Ix</literal> <literal>Read</literal>, and
150 <literal>Show</literal>) are available
151 <emphasis>only</emphasis> up to 16-tuples.</para>
153 <para>This limitation is easily subvertible, so please ask
154 if you get stuck on it.</para>
159 <term><literal>Read</literal>ing integers</term>
161 <para>GHC's implementation of the
162 <literal>Read</literal> class for integral types accepts
163 hexadecimal and octal literals (the code in the Haskell
164 98 report doesn't). So, for example,
165 <programlisting>read "0xf00" :: Int</programlisting>
167 <para>A possible reason for this is that <literal>readLitChar</literal> accepts hex and
168 octal escapes, so it seems inconsistent not to do so for integers too.</para>
173 <term><literal>isAlpha</literal></term>
175 <para>The Haskell 98 definition of <literal>isAlpha</literal>
178 <programlisting>isAlpha c = isUpper c || isLower c</programlisting>
180 <para>GHC's implementation diverges from the Haskell 98
181 definition in the sense that Unicode alphabetic characters which
182 are neither upper nor lower case will still be identified as
183 alphabetic by <literal>isAlpha</literal>.</para>
190 <sect2 id="haskell98-undefined">
191 <title>GHC's interpretation of undefined behaviour in
192 Haskell 98</title>
194 <para>This section documents GHC's take on various issues that are
195 left undefined or implementation specific in Haskell 98.</para>
200 The <literal>Char</literal> type
201 <indexterm><primary><literal>Char</literal></primary><secondary>size of</secondary></indexterm>
204 <para>Following the ISO-10646 standard,
205 <literal>maxBound :: Char</literal> in GHC is
206 <literal>0x10FFFF</literal>.</para>
213 <indexterm><primary><literal>Int</literal></primary><secondary>size of</secondary></indexterm>
216 <para>In GHC the <literal>Int</literal> type follows the
217 size of an address on the host architecture; in other words
218 it holds 32 bits on a 32-bit machine, and 64-bits on a
219 64-bit machine.</para>
221 <para>Arithmetic on <literal>Int</literal> is unchecked for
222 overflow<indexterm><primary>overflow</primary><secondary><literal>Int</literal></secondary>
223 </indexterm>, so all operations on <literal>Int</literal> happen
225 2<superscript><replaceable>n</replaceable></superscript>
226 where <replaceable>n</replaceable> is the size in bits of
227 the <literal>Int</literal> type.</para>
229 <para>The <literal>fromInteger</literal><indexterm><primary><literal>fromInteger</literal></primary>
230 </indexterm>function (and hence
231 also <literal>fromIntegral</literal><indexterm><primary><literal>fromIntegral</literal></primary>
232 </indexterm>) is a special case when
233 converting to <literal>Int</literal>. The value of
234 <literal>fromIntegral x :: Int</literal> is given by taking
235 the lower <replaceable>n</replaceable> bits of <literal>(abs
236 x)</literal>, multiplied by the sign of <literal>x</literal>
237 (in 2's complement <replaceable>n</replaceable>-bit
238 arithmetic). This behaviour was chosen so that for example
239 writing <literal>0xffffffff :: Int</literal> preserves the
240 bit-pattern in the resulting <literal>Int</literal>.</para>
243 <para>Negative literals, such as <literal>-3</literal>, are
244 specified by (a careful reading of) the Haskell Report as
245 meaning <literal>Prelude.negate (Prelude.fromInteger 3)</literal>.
246 So <literal>-2147483648</literal> means <literal>negate (fromInteger 2147483648)</literal>.
247 Since <literal>fromInteger</literal> takes the lower 32 bits of the representation,
248 <literal>fromInteger (2147483648::Integer)</literal>, computed at type <literal>Int</literal> is
249 <literal>-2147483648::Int</literal>. The <literal>negate</literal> operation then
250 overflows, but it is unchecked, so <literal>negate (-2147483648::Int)</literal> is just
251 <literal>-2147483648</literal>. In short, one can write <literal>minBound::Int</literal> as
252 a literal with the expected meaning (but that is not in general guaranteed.
255 <para>The <literal>fromIntegral</literal> function also
256 preserves bit-patterns when converting between the sized
257 integral types (<literal>Int8</literal>,
258 <literal>Int16</literal>, <literal>Int32</literal>,
259 <literal>Int64</literal> and the unsigned
260 <literal>Word</literal> variants), see the modules
261 <literal>Data.Int</literal> and <literal>Data.Word</literal>
262 in the library documentation.</para>
267 <term>Unchecked float arithmetic</term>
269 <para>Operations on <literal>Float</literal> and
270 <literal>Double</literal> numbers are
271 <emphasis>unchecked</emphasis> for overflow, underflow, and
272 other sad occurrences. (note, however that some
273 architectures trap floating-point overflow and
274 loss-of-precision and report a floating-point exception,
275 probably terminating the
276 program)<indexterm><primary>floating-point
277 exceptions</primary></indexterm>.</para>
287 <title>Known bugs or infelicities</title>
289 <para>The bug tracker lists bugs that have been reported in GHC but not
290 yet fixed: see the <ulink url="http://sourceforge.net/projects/ghc/">SourceForge GHC
291 page</ulink>. In addition to those, GHC also has the following known bugs
292 or infelicities. These bugs are more permanent; it is unlikely that
293 any of them will be fixed in the short term.</para>
295 <sect2 id="bugs-ghc">
296 <title>Bugs in GHC</title>
300 <para> GHC can warn about non-exhaustive or overlapping
301 patterns (see <xref linkend="options-sanity"/>), and usually
302 does so correctly. But not always. It gets confused by
303 string patterns, and by guards, and can then emit bogus
304 warnings. The entire overlap-check code needs an overhaul
309 <para>GHC does not allow you to have a data type with a context
310 that mentions type variables that are not data type parameters.
313 data C a b => T a = MkT a
315 so that <literal>MkT</literal>'s type is
317 MkT :: forall a b. C a b => a -> T a
319 In principle, with a suitable class declaration with a functional dependency,
320 it's possible that this type is not ambiguous; but GHC nevertheless rejects
321 it. The type variables mentioned in the context of the data type declaration must
322 be among the type parameters of the data type.</para>
326 <para>GHC's inliner can be persuaded into non-termination
327 using the standard way to encode recursion via a data type:</para>
329 data U = MkU (U -> Bool)
332 russel u@(MkU p) = not $ p u
335 x = russel (MkU russel)
338 <para>We have never found another class of programs, other
339 than this contrived one, that makes GHC diverge, and fixing
340 the problem would impose an extra overhead on every
341 compilation. So the bug remains un-fixed. There is more
343 url="http://research.microsoft.com/~simonpj/Papers/inlining">
344 Secrets of the GHC inliner</ulink>.</para>
348 <para>GHC does not keep careful track of
349 what instance declarations are 'in scope' if they come from other packages.
350 Instead, all instance declarations that GHC has seen in other
351 packages are all in scope everywhere, whether or not the
352 module from that package is used by the command-line
353 expression. This bug affects only the <option>--make</option> mode and
360 <sect2 id="bugs-ghci">
361 <title>Bugs in GHCi (the interactive GHC)</title>
364 <para>GHCi does not respect the <literal>default</literal>
365 declaration in the module whose scope you are in. Instead,
366 for expressions typed at the command line, you always get the
367 default default-type behaviour; that is,
368 <literal>default(Int,Double)</literal>.</para>
370 <para>It would be better for GHCi to record what the default
371 settings in each module are, and use those of the 'current'
372 module (whatever that is).</para>
376 <para>On Windows, there's a GNU ld/BFD bug
377 whereby it emits bogus PE object files that have more than
378 0xffff relocations. When GHCi tries to load a package affected by this
379 bug, you get an error message of the form
381 Loading package javavm ... linking ... WARNING: Overflown relocation field (# relocs found: 30765)
383 The last time we looked, this bug still
384 wasn't fixed in the BFD codebase, and there wasn't any
385 noticeable interest in fixing it when we reported the bug
388 <para>The workaround is to split up the .o files that make up
389 your package into two or more .o's, along the lines of
390 how the "base" package does it.</para>
399 ;;; Local Variables: ***
401 ;;; sgml-parent-document: ("users_guide.xml" "book" "chapter") ***