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-vect</option>:
154 <indexterm><primary><option>-ddump-vect</option></primary></indexterm>
157 <para>dumps the output of the vectoriser.
164 <option>-ddump-simpl</option>:
165 <indexterm><primary><option>-ddump-simpl</option></primary></indexterm>
168 <para>simplifier output (Core-to-Core passes)</para>
174 <option>-ddump-inlinings</option>:
175 <indexterm><primary><option>-ddump-inlinings</option></primary></indexterm>
178 <para>inlining info from the simplifier</para>
184 <option>-ddump-cpranal</option>:
185 <indexterm><primary><option>-ddump-cpranal</option></primary></indexterm>
188 <para>CPR analyser output</para>
194 <option>-ddump-stranal</option>:
195 <indexterm><primary><option>-ddump-stranal</option></primary></indexterm>
198 <para>strictness analyser output</para>
204 <option>-ddump-cse</option>:
205 <indexterm><primary><option>-ddump-cse</option></primary></indexterm>
208 <para>CSE pass output</para>
214 <option>-ddump-worker-wrapper</option>:
215 <indexterm><primary><option>-ddump-worker-wrapper</option></primary></indexterm>
218 <para>worker/wrapper split output</para>
224 <option>-ddump-occur-anal</option>:
225 <indexterm><primary><option>-ddump-occur-anal</option></primary></indexterm>
228 <para>`occurrence analysis' output</para>
234 <option>-ddump-prep</option>:
235 <indexterm><primary><option>-ddump-prep</option></primary></indexterm>
238 <para>output of core preparation pass</para>
244 <option>-ddump-stg</option>:
245 <indexterm><primary><option>-ddump-stg</option></primary></indexterm>
248 <para>output of STG-to-STG passes</para>
254 <option>-ddump-flatC</option>:
255 <indexterm><primary><option>-ddump-flatC</option></primary></indexterm>
258 <para><emphasis>flattened</emphasis> Abstract C</para>
264 <option>-ddump-cmm</option>:
265 <indexterm><primary><option>-ddump-cmm</option></primary></indexterm>
268 <para>Print the C-- code out.</para>
274 <option>-ddump-opt-cmm</option>:
275 <indexterm><primary><option>-ddump-opt-cmm</option></primary></indexterm>
278 <para>Dump the results of C-- to C-- optimising passes.</para>
284 <option>-ddump-asm</option>:
285 <indexterm><primary><option>-ddump-asm</option></primary></indexterm>
288 <para>assembly language from the native-code generator</para>
294 <option>-ddump-bcos</option>:
295 <indexterm><primary><option>-ddump-bcos</option></primary></indexterm>
298 <para>byte code compiler output</para>
304 <option>-ddump-foreign</option>:
305 <indexterm><primary><option>-ddump-foreign</option></primary></indexterm>
308 <para>dump foreign export stubs</para>
317 <option>-ddump-simpl-phases</option>:
318 <indexterm><primary><option>-ddump-simpl-phases</option></primary></indexterm>
321 <para>Show the output of each run of the simplifier. Used when even
322 <option>-dverbose-core2core</option> doesn't cut it.</para>
328 <option>-ddump-simpl-iterations</option>:
329 <indexterm><primary><option>-ddump-simpl-iterations</option></primary></indexterm>
332 <para>Show the output of each <emphasis>iteration</emphasis>
333 of the simplifier (each run of the simplifier has a maximum
334 number of iterations, normally 4). This outputs even more information
335 than <option>-ddump-simpl-phases</option>.</para>
341 <option>-ddump-simpl-stats</option>
342 <indexterm><primary><option>-ddump-simpl-stats option</option></primary></indexterm>
345 <para>Dump statistics about how many of each kind of
346 transformation too place. If you add
347 <option>-dppr-debug</option> you get more detailed
354 <option>-ddump-if-trace</option>
355 <indexterm><primary><option>-ddump-if-trace</option></primary></indexterm>
358 <para>Make the interface loader be *real* chatty about what it is
365 <option>-ddump-tc-trace</option>
366 <indexterm><primary><option>-ddump-tc-trace</option></primary></indexterm>
369 <para>Make the type checker be *real* chatty about what it is
376 <option>-ddump-vt-trace</option>
377 <indexterm><primary><option>-ddump-tv-trace</option></primary></indexterm>
380 <para>Make the vectoriser be *real* chatty about what it is
387 <option>-ddump-rn-trace</option>
388 <indexterm><primary><option>-ddump-rn-trace</option></primary></indexterm>
391 <para>Make the renamer be *real* chatty about what it is
398 <option>-ddump-rn-stats</option>
399 <indexterm><primary><option>-dshow-rn-stats</option></primary></indexterm>
402 <para>Print out summary of what kind of information the renamer
403 had to bring in.</para>
409 <option>-dverbose-core2core</option>
410 <indexterm><primary><option>-dverbose-core2core</option></primary></indexterm>
413 <option>-dverbose-stg2stg</option>
414 <indexterm><primary><option>-dverbose-stg2stg</option></primary></indexterm>
417 <para>Show the output of the intermediate Core-to-Core and
418 STG-to-STG passes, respectively. (<emphasis>Lots</emphasis>
419 of output!) So: when we're really desperate:</para>
422 % ghc -noC -O -ddump-simpl -dverbose-core2core -dcore-lint Foo.hs
430 <option>-dshow-passes</option>
431 <indexterm><primary><option>-dshow-passes</option></primary></indexterm>
434 <para>Print out each pass name as it happens.</para>
440 <option>-ddump-core-stats</option>
441 <indexterm><primary><option>-ddump-core-stats</option></primary></indexterm>
444 <para>Print a one-line summary of the size of the Core program
445 at the end of the optimisation pipeline.</para>
451 <option>-dfaststring-stats</option>
452 <indexterm><primary><option>-dfaststring-stats</option></primary></indexterm>
455 <para>Show statistics for the usage of fast strings by the
462 <option>-dppr-debug</option>
463 <indexterm><primary><option>-dppr-debug</option></primary></indexterm>
466 <para>Debugging output is in one of several
467 “styles.” Take the printing of types, for
468 example. In the “user” style (the default), the
469 compiler's internal ideas about types are presented in
470 Haskell source-level syntax, insofar as possible. In the
471 “debug” style (which is the default for
472 debugging output), the types are printed in with explicit
473 foralls, and variables have their unique-id attached (so you
474 can check for things that look the same but aren't). This
475 flag makes debugging output appear in the more verbose debug
482 <option>-dsuppress-uniques</option>
483 <indexterm><primary><option>-dsuppress-uniques</option></primary></indexterm>
486 <para>Suppress the printing of uniques in debugging output. This may make
487 the printout ambiguous (e.g. unclear where an occurrence of 'x' is bound), but
488 it makes the output of two compiler runs have many fewer gratuitous differences,
489 so you can realistically apply <command>diff</command>. Once <command>diff</command>
490 has shown you where to look, you can try again without <option>-dsuppress-uniques</option></para>
496 <option>-dsuppress-coercions</option>
497 <indexterm><primary><option>-dsuppress-coercions</option></primary></indexterm>
500 <para>Suppress the printing of coercions in Core dumps to make them
507 <option>-dsuppress-module-prefixes</option>
508 <indexterm><primary><option>-dsuppress-module-prefixes</option></primary></indexterm>
511 <para>Suppress the printing of module qualification prefixes in Core dumps to make them easier to read.</para>
517 <option>-dppr-user-length</option>
518 <indexterm><primary><option>-dppr-user-length</option></primary></indexterm>
521 <para>In error messages, expressions are printed to a
522 certain “depth”, with subexpressions beyond the
523 depth replaced by ellipses. This flag sets the
524 depth. Its default value is 5.</para>
530 <option>-dno-debug-output</option>
531 <indexterm><primary><option>-dno-debug-output</option></primary></indexterm>
534 <para>Suppress any unsolicited debugging output. When GHC
535 has been built with the <literal>DEBUG</literal> option it
536 occasionally emits debug output of interest to developers.
537 The extra output can confuse the testing framework and
538 cause bogus test failures, so this flag is provided to
545 <sect2 id="checking-consistency">
546 <title>Checking for consistency</title>
548 <indexterm><primary>consistency checks</primary></indexterm>
549 <indexterm><primary>lint</primary></indexterm>
555 <option>-dcore-lint</option>
556 <indexterm><primary><option>-dcore-lint</option></primary></indexterm>
559 <para>Turn on heavyweight intra-pass sanity-checking within
560 GHC, at Core level. (It checks GHC's sanity, not yours.)</para>
566 <option>-dstg-lint</option>:
567 <indexterm><primary><option>-dstg-lint</option></primary></indexterm>
570 <para>Ditto for STG level. (NOTE: currently doesn't work).</para>
576 <option>-dcmm-lint</option>:
577 <indexterm><primary><option>-dcmm-lint</option></primary></indexterm>
580 <para>Ditto for C-- level.</para>
588 <title>How to read Core syntax (from some <option>-ddump</option>
591 <indexterm><primary>reading Core syntax</primary></indexterm>
592 <indexterm><primary>Core syntax, how to read</primary></indexterm>
594 <para>Let's do this by commenting an example. It's from doing
595 <option>-ddump-ds</option> on this code:
598 skip2 m = m : skip2 (m+2)
601 Before we jump in, a word about names of things. Within GHC,
602 variables, type constructors, etc., are identified by their
603 “Uniques.” These are of the form `letter' plus
604 `number' (both loosely interpreted). The `letter' gives some idea
605 of where the Unique came from; e.g., <literal>_</literal>
606 means “built-in type variable”; <literal>t</literal>
607 means “from the typechecker”; <literal>s</literal>
608 means “from the simplifier”; and so on. The `number'
609 is printed fairly compactly in a `base-62' format, which everyone
610 hates except me (WDP).</para>
612 <para>Remember, everything has a “Unique” and it is
613 usually printed out when debugging, in some form or another. So
614 here we go…</para>
618 Main.skip2{-r1L6-} :: _forall_ a$_4 =>{{Num a$_4}} -> a$_4 -> [a$_4]
620 --# `r1L6' is the Unique for Main.skip2;
621 --# `_4' is the Unique for the type-variable (template) `a'
622 --# `{{Num a$_4}}' is a dictionary argument
626 --# `_NI_' means "no (pragmatic) information" yet; it will later
627 --# evolve into the GHC_PRAGMA info that goes into interface files.
630 /\ _4 -> \ d.Num.t4Gt ->
633 +.t4Hg :: _4 -> _4 -> _4
635 +.t4Hg = (+{-r3JH-} _4) d.Num.t4Gt
637 fromInt.t4GS :: Int{-2i-} -> _4
639 fromInt.t4GS = (fromInt{-r3JX-} _4) d.Num.t4Gt
641 --# The `+' class method (Unique: r3JH) selects the addition code
642 --# from a `Num' dictionary (now an explicit lambda'd argument).
643 --# Because Core is 2nd-order lambda-calculus, type applications
644 --# and lambdas (/\) are explicit. So `+' is first applied to a
645 --# type (`_4'), then to a dictionary, yielding the actual addition
646 --# function that we will use subsequently...
648 --# We play the exact same game with the (non-standard) class method
649 --# `fromInt'. Unsurprisingly, the type `Int' is wired into the
659 } in fromInt.t4GS ds.d4Qz
661 --# `I# 2#' is just the literal Int `2'; it reflects the fact that
662 --# GHC defines `data Int = I# Int#', where Int# is the primitive
663 --# unboxed type. (see relevant info about unboxed types elsewhere...)
665 --# The `!' after `I#' indicates that this is a *saturated*
666 --# application of the `I#' data constructor (i.e., not partially
669 skip2.t3Ja :: _4 -> [_4]
673 let { ds.d4QQ :: [_4]
679 ds.d4QY = +.t4Hg m.r1H4 lit.t4Hb
680 } in skip2.t3Ja ds.d4QY
688 <para>(“It's just a simple functional language” is an
689 unregisterised trademark of Peyton Jones Enterprises, plc.)</para>
694 <title>Unregisterised compilation</title>
695 <indexterm><primary>unregisterised compilation</primary></indexterm>
697 <para>The term "unregisterised" really means "compile via vanilla
698 C", disabling some of the platform-specific tricks that GHC
699 normally uses to make programs go faster. When compiling
700 unregisterised, GHC simply generates a C file which is compiled
703 <para>Unregisterised compilation can be useful when porting GHC to
704 a new machine, since it reduces the prerequisite tools to
705 <command>gcc</command>, <command>as</command>, and
706 <command>ld</command> and nothing more, and furthermore the amount
707 of platform-specific code that needs to be written in order to get
708 unregisterised compilation going is usually fairly small.</para>
710 <para>Unregisterised compilation cannot be selected at
711 compile-time; you have to build GHC with the appropriate options
712 set. Consult the GHC Building Guide for details.</para>
718 ;;; Local Variables: ***
719 ;;; sgml-parent-document: ("users_guide.xml" "book" "chapter" "sect1") ***