1 <?xml version="1.0" encoding="iso-8859-1"?>
2 <sect1 id="options-debugging">
3 <title>Debugging the compiler</title>
5 <indexterm><primary>debugging options (for GHC)</primary></indexterm>
7 <para>HACKER TERRITORY. HACKER TERRITORY. (You were warned.)</para>
9 <sect2 id="dumping-output">
10 <title>Dumping out compiler intermediate structures</title>
12 <indexterm><primary>dumping GHC intermediates</primary></indexterm>
13 <indexterm><primary>intermediate passes, output</primary></indexterm>
18 <option>-ddump-</option><replaceable>pass</replaceable>
19 <indexterm><primary><option>-ddump</option> options</primary></indexterm>
22 <para>Make a debugging dump after pass
23 <literal><pass></literal> (may be common enough to need
24 a short form…). You can get all of these at once
25 (<emphasis>lots</emphasis> of output) by using
26 <option>-v5</option>, or most of them with
27 <option>-v4</option>. Some of the most useful ones
33 <option>-ddump-parsed</option>:
34 <indexterm><primary><option>-ddump-parsed</option></primary></indexterm>
37 <para>parser output</para>
43 <option>-ddump-rn</option>:
44 <indexterm><primary><option>-ddump-rn</option></primary></indexterm>
47 <para>renamer output</para>
53 <option>-ddump-tc</option>:
54 <indexterm><primary><option>-ddump-tc</option></primary></indexterm>
57 <para>typechecker output</para>
63 <option>-ddump-splices</option>:
64 <indexterm><primary><option>-ddump-splices</option></primary></indexterm>
67 <para>Dump Template Haskell expressions that we splice in,
68 and what Haskell code the expression evaluates to.</para>
74 <option>-ddump-types</option>:
75 <indexterm><primary><option>-ddump-types</option></primary></indexterm>
78 <para>Dump a type signature for each value defined at
79 the top level of the module. The list is sorted
80 alphabetically. Using <option>-dppr-debug</option>
81 dumps a type signature for all the imported and
82 system-defined things as well; useful for debugging the
89 <option>-ddump-deriv</option>:
90 <indexterm><primary><option>-ddump-deriv</option></primary></indexterm>
93 <para>derived instances</para>
99 <option>-ddump-ds</option>:
100 <indexterm><primary><option>-ddump-ds</option></primary></indexterm>
103 <para>desugarer output</para>
109 <option>-ddump-spec</option>:
110 <indexterm><primary><option>-ddump-spec</option></primary></indexterm>
113 <para>output of specialisation pass</para>
119 <option>-ddump-rules</option>:
120 <indexterm><primary><option>-ddump-rules</option></primary></indexterm>
123 <para>dumps all rewrite rules specified in this module;
124 see <xref linkend="controlling-rules"/>.
131 <option>-ddump-rule-firings</option>:
132 <indexterm><primary><option>-ddump-rule-firings</option></primary></indexterm>
135 <para>dumps the names of all rules that fired in this module</para>
141 <option>-ddump-rule-rewrites</option>:
142 <indexterm><primary><option>-ddump-rule-rewrites</option></primary></indexterm>
145 <para>dumps detailed information about all rules that fired in
153 <option>-ddump-simpl</option>:
154 <indexterm><primary><option>-ddump-simpl</option></primary></indexterm>
157 <para>simplifier output (Core-to-Core passes)</para>
163 <option>-ddump-inlinings</option>:
164 <indexterm><primary><option>-ddump-inlinings</option></primary></indexterm>
167 <para>inlining info from the simplifier</para>
173 <option>-ddump-cpranal</option>:
174 <indexterm><primary><option>-ddump-cpranal</option></primary></indexterm>
177 <para>CPR analyser output</para>
183 <option>-ddump-stranal</option>:
184 <indexterm><primary><option>-ddump-stranal</option></primary></indexterm>
187 <para>strictness analyser output</para>
193 <option>-ddump-cse</option>:
194 <indexterm><primary><option>-ddump-cse</option></primary></indexterm>
197 <para>CSE pass output</para>
203 <option>-ddump-worker-wrapper</option>:
204 <indexterm><primary><option>-ddump-worker-wrapper</option></primary></indexterm>
207 <para>worker/wrapper split output</para>
213 <option>-ddump-occur-anal</option>:
214 <indexterm><primary><option>-ddump-occur-anal</option></primary></indexterm>
217 <para>`occurrence analysis' output</para>
223 <option>-ddump-prep</option>:
224 <indexterm><primary><option>-ddump-prep</option></primary></indexterm>
227 <para>output of core preparation pass</para>
233 <option>-ddump-stg</option>:
234 <indexterm><primary><option>-ddump-stg</option></primary></indexterm>
237 <para>output of STG-to-STG passes</para>
243 <option>-ddump-flatC</option>:
244 <indexterm><primary><option>-ddump-flatC</option></primary></indexterm>
247 <para><emphasis>flattened</emphasis> Abstract C</para>
253 <option>-ddump-cmm</option>:
254 <indexterm><primary><option>-ddump-cmm</option></primary></indexterm>
257 <para>Print the C-- code out.</para>
263 <option>-ddump-opt-cmm</option>:
264 <indexterm><primary><option>-ddump-opt-cmm</option></primary></indexterm>
267 <para>Dump the results of C-- to C-- optimising passes.</para>
273 <option>-ddump-asm</option>:
274 <indexterm><primary><option>-ddump-asm</option></primary></indexterm>
277 <para>assembly language from the native-code generator</para>
283 <option>-ddump-bcos</option>:
284 <indexterm><primary><option>-ddump-bcos</option></primary></indexterm>
287 <para>byte code compiler output</para>
293 <option>-ddump-foreign</option>:
294 <indexterm><primary><option>-ddump-foreign</option></primary></indexterm>
297 <para>dump foreign export stubs</para>
306 <option>-ddump-simpl-phases</option>:
307 <indexterm><primary><option>-ddump-simpl-phases</option></primary></indexterm>
310 <para>Show the output of each run of the simplifier. Used when even
311 <option>-dverbose-core2core</option> doesn't cut it.</para>
317 <option>-ddump-simpl-iterations</option>:
318 <indexterm><primary><option>-ddump-simpl-iterations</option></primary></indexterm>
321 <para>Show the output of each <emphasis>iteration</emphasis>
322 of the simplifier (each run of the simplifier has a maximum
323 number of iterations, normally 4). This outputs even more information
324 than <option>-ddump-simpl-phases</option>.</para>
330 <option>-ddump-simpl-stats</option>
331 <indexterm><primary><option>-ddump-simpl-stats option</option></primary></indexterm>
334 <para>Dump statistics about how many of each kind of
335 transformation too place. If you add
336 <option>-dppr-debug</option> you get more detailed
343 <option>-ddump-if-trace</option>
344 <indexterm><primary><option>-ddump-if-trace</option></primary></indexterm>
347 <para>Make the interface loader be *real* chatty about what it is
354 <option>-ddump-tc-trace</option>
355 <indexterm><primary><option>-ddump-tc-trace</option></primary></indexterm>
358 <para>Make the type checker be *real* chatty about what it is
365 <option>-ddump-rn-trace</option>
366 <indexterm><primary><option>-ddump-rn-trace</option></primary></indexterm>
369 <para>Make the renamer be *real* chatty about what it is
376 <option>-ddump-rn-stats</option>
377 <indexterm><primary><option>-dshow-rn-stats</option></primary></indexterm>
380 <para>Print out summary of what kind of information the renamer
381 had to bring in.</para>
387 <option>-dverbose-core2core</option>
388 <indexterm><primary><option>-dverbose-core2core</option></primary></indexterm>
391 <option>-dverbose-stg2stg</option>
392 <indexterm><primary><option>-dverbose-stg2stg</option></primary></indexterm>
395 <para>Show the output of the intermediate Core-to-Core and
396 STG-to-STG passes, respectively. (<emphasis>Lots</emphasis>
397 of output!) So: when we're really desperate:</para>
400 % ghc -noC -O -ddump-simpl -dverbose-core2core -dcore-lint Foo.hs
408 <option>-dshow-passes</option>
409 <indexterm><primary><option>-dshow-passes</option></primary></indexterm>
412 <para>Print out each pass name as it happens.</para>
418 <option>-dfaststring-stats</option>
419 <indexterm><primary><option>-dfaststring-stats</option></primary></indexterm>
422 <para>Show statistics for the usage of fast strings by the
429 <option>-dppr-debug</option>
430 <indexterm><primary><option>-dppr-debug</option></primary></indexterm>
433 <para>Debugging output is in one of several
434 “styles.” Take the printing of types, for
435 example. In the “user” style (the default), the
436 compiler's internal ideas about types are presented in
437 Haskell source-level syntax, insofar as possible. In the
438 “debug” style (which is the default for
439 debugging output), the types are printed in with explicit
440 foralls, and variables have their unique-id attached (so you
441 can check for things that look the same but aren't). This
442 flag makes debugging output appear in the more verbose debug
449 <option>-dsuppress-uniques</option>
450 <indexterm><primary><option>-dsuppress-uniques</option></primary></indexterm>
453 <para>Suppress the printing of uniques in debugging output. This may make
454 the printout ambiguous (e.g. unclear where an occurrence of 'x' is bound), but
455 it makes the output of two compiler runs have many fewer gratuitous differences,
456 so you can realistically apply <command>diff</command>. Once <command>diff</command>
457 has shown you where to look, you can try again without <option>-dsuppress-uniques</option></para>
463 <option>-dsuppress-coercions</option>
464 <indexterm><primary><option>-dsuppress-coercions</option></primary></indexterm>
467 <para>Suppress the printing of coercions in Core dumps to make them
474 <option>-dsuppress-module-prefixes</option>
475 <indexterm><primary><option>-dsuppress-module-prefixes</option></primary></indexterm>
478 <para>Suppress the printing of module qualification prefixes in Core dumps to make them easier to read.</para>
484 <option>-dppr-user-length</option>
485 <indexterm><primary><option>-dppr-user-length</option></primary></indexterm>
488 <para>In error messages, expressions are printed to a
489 certain “depth”, with subexpressions beyond the
490 depth replaced by ellipses. This flag sets the
491 depth. Its default value is 5.</para>
497 <option>-dno-debug-output</option>
498 <indexterm><primary><option>-dno-debug-output</option></primary></indexterm>
501 <para>Suppress any unsolicited debugging output. When GHC
502 has been built with the <literal>DEBUG</literal> option it
503 occasionally emits debug output of interest to developers.
504 The extra output can confuse the testing framework and
505 cause bogus test failures, so this flag is provided to
512 <sect2 id="checking-consistency">
513 <title>Checking for consistency</title>
515 <indexterm><primary>consistency checks</primary></indexterm>
516 <indexterm><primary>lint</primary></indexterm>
522 <option>-dcore-lint</option>
523 <indexterm><primary><option>-dcore-lint</option></primary></indexterm>
526 <para>Turn on heavyweight intra-pass sanity-checking within
527 GHC, at Core level. (It checks GHC's sanity, not yours.)</para>
533 <option>-dstg-lint</option>:
534 <indexterm><primary><option>-dstg-lint</option></primary></indexterm>
537 <para>Ditto for STG level. (NOTE: currently doesn't work).</para>
543 <option>-dcmm-lint</option>:
544 <indexterm><primary><option>-dcmm-lint</option></primary></indexterm>
547 <para>Ditto for C-- level.</para>
555 <title>How to read Core syntax (from some <option>-ddump</option>
558 <indexterm><primary>reading Core syntax</primary></indexterm>
559 <indexterm><primary>Core syntax, how to read</primary></indexterm>
561 <para>Let's do this by commenting an example. It's from doing
562 <option>-ddump-ds</option> on this code:
565 skip2 m = m : skip2 (m+2)
568 Before we jump in, a word about names of things. Within GHC,
569 variables, type constructors, etc., are identified by their
570 “Uniques.” These are of the form `letter' plus
571 `number' (both loosely interpreted). The `letter' gives some idea
572 of where the Unique came from; e.g., <literal>_</literal>
573 means “built-in type variable”; <literal>t</literal>
574 means “from the typechecker”; <literal>s</literal>
575 means “from the simplifier”; and so on. The `number'
576 is printed fairly compactly in a `base-62' format, which everyone
577 hates except me (WDP).</para>
579 <para>Remember, everything has a “Unique” and it is
580 usually printed out when debugging, in some form or another. So
581 here we go…</para>
585 Main.skip2{-r1L6-} :: _forall_ a$_4 =>{{Num a$_4}} -> a$_4 -> [a$_4]
587 --# `r1L6' is the Unique for Main.skip2;
588 --# `_4' is the Unique for the type-variable (template) `a'
589 --# `{{Num a$_4}}' is a dictionary argument
593 --# `_NI_' means "no (pragmatic) information" yet; it will later
594 --# evolve into the GHC_PRAGMA info that goes into interface files.
597 /\ _4 -> \ d.Num.t4Gt ->
600 +.t4Hg :: _4 -> _4 -> _4
602 +.t4Hg = (+{-r3JH-} _4) d.Num.t4Gt
604 fromInt.t4GS :: Int{-2i-} -> _4
606 fromInt.t4GS = (fromInt{-r3JX-} _4) d.Num.t4Gt
608 --# The `+' class method (Unique: r3JH) selects the addition code
609 --# from a `Num' dictionary (now an explicit lambda'd argument).
610 --# Because Core is 2nd-order lambda-calculus, type applications
611 --# and lambdas (/\) are explicit. So `+' is first applied to a
612 --# type (`_4'), then to a dictionary, yielding the actual addition
613 --# function that we will use subsequently...
615 --# We play the exact same game with the (non-standard) class method
616 --# `fromInt'. Unsurprisingly, the type `Int' is wired into the
626 } in fromInt.t4GS ds.d4Qz
628 --# `I# 2#' is just the literal Int `2'; it reflects the fact that
629 --# GHC defines `data Int = I# Int#', where Int# is the primitive
630 --# unboxed type. (see relevant info about unboxed types elsewhere...)
632 --# The `!' after `I#' indicates that this is a *saturated*
633 --# application of the `I#' data constructor (i.e., not partially
636 skip2.t3Ja :: _4 -> [_4]
640 let { ds.d4QQ :: [_4]
646 ds.d4QY = +.t4Hg m.r1H4 lit.t4Hb
647 } in skip2.t3Ja ds.d4QY
655 <para>(“It's just a simple functional language” is an
656 unregisterised trademark of Peyton Jones Enterprises, plc.)</para>
661 <title>Unregisterised compilation</title>
662 <indexterm><primary>unregisterised compilation</primary></indexterm>
664 <para>The term "unregisterised" really means "compile via vanilla
665 C", disabling some of the platform-specific tricks that GHC
666 normally uses to make programs go faster. When compiling
667 unregisterised, GHC simply generates a C file which is compiled
670 <para>Unregisterised compilation can be useful when porting GHC to
671 a new machine, since it reduces the prerequisite tools to
672 <command>gcc</command>, <command>as</command>, and
673 <command>ld</command> and nothing more, and furthermore the amount
674 of platform-specific code that needs to be written in order to get
675 unregisterised compilation going is usually fairly small.</para>
677 <para>Unregisterised compilation cannot be selected at
678 compile-time; you have to build GHC with the appropriate options
679 set. Consult the GHC Building Guide for details.</para>
685 ;;; Local Variables: ***
686 ;;; sgml-parent-document: ("users_guide.xml" "book" "chapter" "sect1") ***