1 <Chapter id="bugs-and-infelicities">
2 <title>Known bugs and infelicities
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>
45 <para>When <option>-fglasgow-exts</option> is on, GHC
46 reserves several keywords beginning with two underscores.
47 This is due to the fact that GHC uses the same lexical
48 analyser for interface file parsing as it does for source
49 file parsing, and these keywords are used in interface
50 files. Do not use any identifiers beginning with a double
51 underscore in <option>-fglasgow-exts</option> mode.</para>
56 <sect3 id="infelicities-syntax">
57 <title>Context-free syntax</title>
61 <para>GHC doesn't do fixity resolution in expressions during
62 parsing. For example, according to the Haskell report, the
63 following expression is legal Haskell:
65 let x = 42 in x == 42 == True</programlisting>
68 (let x = 42 in x == 42) == True</programlisting>
70 because according to the report, the <literal>let</literal>
71 expression <quote>extends as far to the right as
72 possible</quote>. Since it can't extend past the second
73 equals sign without causing a parse error
74 (<literal>==</literal> is non-fix), the
75 <literal>let</literal>-expression must terminate there. GHC
76 simply gobbles up the whole expression, parsing like this:
78 (let x = 42 in x == 42 == True)</programlisting>
80 The Haskell report is arguably wrong here, but nevertheless
81 it's a difference between GHC & Haskell 98.</para>
86 <sect3 id="infelicities-exprs-pats">
87 <title>Expressions and patterns</title>
91 <term>Very long <literal>String</literal> constants:</term>
93 <para>May not go through. If you add a “string
94 gap” every few thousand characters, then the strings
95 can be as long as you like.</para>
97 <para>Bear in mind that string gaps and the
98 <option>-cpp</option><indexterm><primary><option>-cpp</option>
99 </primary></indexterm> option don't mix very well (see
100 <xref linkend="c-pre-processor">).</para>
107 <sect3 id="infelicities-decls">
108 <title>Declarations and bindings</title>
110 <para>None known.</para>
114 <sect3 id="infelicities-Modules">
115 <title>Module system and interface files</title>
120 <term> Namespace pollution </term>
122 <para>Several modules internal to GHC are visible in the
123 standard namespace. All of these modules begin with
124 <literal>Prel</literal>, so the rule is: don't use any
125 modules beginning with <literal>Prel</literal> in your
126 program, or you may be comprehensively screwed.</para>
133 <sect3 id="infelicities-numbers">
134 <title>Numbers, basic types, and built-in classes</title>
138 <term>Multiply-defined array elements—not checked:</term>
140 <para>This code fragment <emphasis>should</emphasis>
141 elicit a fatal error, but it does not:
144 main = print (array (1,1) [(1,2), (1,3)])</programlisting>
153 <sect3 id="infelicities-Prelude">
154 <title>In Prelude support</title>
158 <term>The <literal>Char</literal> type</term>
159 <indexterm><primary><literal>Char</literal></primary><secondary>size
160 of</secondary></indexterm>
162 <para>The Haskell report says that the
163 <literal>Char</literal> type holds 16 bits. GHC follows
164 the ISO-10646 standard a little more closely:
165 <literal>maxBound :: Char</literal> in GHC is
166 <literal>0x10FFFF</literal>.</para>
171 <term>Arbitrary-sized tuples</term>
173 <para>Tuples are currently limited to size 61. HOWEVER:
174 standard instances for tuples (<literal>Eq</literal>,
175 <literal>Ord</literal>, <literal>Bounded</literal>,
176 <literal>Ix</literal> <literal>Read</literal>, and
177 <literal>Show</literal>) are available
178 <emphasis>only</emphasis> up to 5-tuples.</para>
180 <para>This limitation is easily subvertible, so please ask
181 if you get stuck on it.</para>
186 <term><literal>Read</literal>ing integers</term>
188 <para>GHC's implementation of the
189 <literal>Read</literal> class for integral types accepts
190 hexadeciaml and octal literals (the code in the Haskell
191 98 report doesn't). So, for example,
192 <programlisting>read "0xf00" :: Int</programlisting>
194 <para>A possible reason for this is that <literal>readLitChar</literal> accepts hex and
195 octal escapes, so it seems inconsistent not to do so for integers too.</para>
202 <sect2 id="haskell98-undefined">
203 <title>GHC's interpretation of undefined behaviour in
204 Haskell 98</title>
206 <para>This section documents GHC's take on various issues that are
207 left undefined or implementation specific in Haskell 98.</para>
211 <term>Sized integral types</term>
212 <indexterm><primary><literal>Int</literal></primary><secondary>size of</secondary>
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>
288 <title>Known bugs or infelicities</title>
290 <para>GHC has the following known bugs or infelicities:
294 GHC only provides tuples up to size 62, and derived tuple instances (for
295 Eq, Ord, etc) up to size 15.
299 GHC can warn about non-exhaustive or overlapping patterns, and usually does so correctly.
300 But not always. It gets confused by string patterns, and by guards, and can then
301 emit bogus warnings. The entire overlap-check code needs an overhaul really.
306 <listitem><para>Dangers with multiple Main modules.</para>
309 GHC does not insist that module <literal>Main</literal> lives in a file called <filename>Main.hs</filename>.
310 This is useful if you want multiple versions of <literal>Main</literal>. But there's a danger: when
311 compiling module <literal>Main</literal> (regardless of what file it comes from), GHC looks for
312 the interface <filename>Main.hi</filename>; it uses this to get version information from the last
313 time it recompiled <literal>Main</literal>. The trouble is that this <filename>Main.hi</filename>
314 may not correspond to the source file being compiled.
317 Solution: remove <filename>Main.hi</filename> first. A better solution would be for GHC to
318 record the source-file filename in the interface file, or even an MD5 checksum.
324 GHCi does not respect the <literal>default</literal> declaration in the module whose
325 scope you are in. Instead, for expressions typed at the command line, you always
326 get the default default-type behaviour; that is, <literal>default(Int,Double)</literal>.
329 It would be better for GHCi to record what the default settings in each module are, and
330 use those of the 'current' module (whatever that is).
334 GHCi does not keep careful track of what instance declarations are 'in scope' if they
335 come from other packages.
336 Instead, all instance declarations that GHC has seen in other packages are all in scope
337 everywhere, whether or not the module from that package is used by the command-line expression.
341 GHC's inliner can be persuaded into non-termination using the standard way to encode
342 recursion via a data type:
344 data U = MkU (U -> Bool)
347 russel u@(MkU p) = not $ p u
350 x = russel (MkU russel)
352 We have never found another class of programs, other than this contrived one, that makes GHC
353 diverge, and fixing the problem would impose an extra overhead on every compilation. So the
354 bug remains un-fixed. There is more background in
356 url="http://research.microsoft.com/~simonpj/Papers/inlining">
357 Secrets of the GHC inliner</ulink>.
359 </itemizedlist></para>
365 ;;; Local Variables: ***
367 ;;; sgml-parent-document: ("users_guide.sgml" "book" "chapter" "sect1") ***