1 <!DOCTYPE Article PUBLIC "-//OASIS//DTD DocBook V3.1//EN">
3 <Article id="building-guide">
7 <Title>Building the Glasgow Functional Programming Tools Suite</Title>
8 <Author><OtherName>The GHC Team</OtherName></Author>
9 <Address><Email>glasgow-haskell-{users,bugs}@haskell.org</Email></Address>
10 <PubDate>November 2001</PubDate>
13 <para>The Glasgow fptools suite is a collection of Functional
14 Programming related tools, including the Glasgow Haskell
15 Compiler (GHC). The source code for the whole suite is kept in
16 a single CVS repository and shares a common build and
17 installation system.</para>
19 <para>This guide is intended for people who want to build or
20 modify programs from the Glasgow <Literal>fptools</Literal>
21 suite (as distinct from those who merely want to
22 <Emphasis>run</Emphasis> them). Installation instructions are
23 now provided in the user guide.</para>
25 <para>The bulk of this guide applies to building on Unix
26 systems; see <XRef LinkEnd="winbuild"> for Windows notes.</para>
32 <sect1 id="sec-getting">
33 <title>Getting the sources</title>
35 <para>You can get your hands on the <literal>fptools</literal>
41 <term><indexterm><primary>Source
42 distributions</primary></indexterm>Source distributions</term>
44 <para>You have a supported platform, but (a) you like
45 the warm fuzzy feeling of compiling things yourself;
46 (b) you want to build something ``extra”—e.g., a
47 set of libraries with strictness-analysis turned off; or
48 (c) you want to hack on GHC yourself.</para>
50 <para>A source distribution contains complete sources for
51 one or more projects in the <literal>fptools</literal>
52 suite. Not only that, but the more awkward
53 machine-independent steps are done for you. For example, if
55 <command>happy</command><indexterm><primary>happy</primary></indexterm>
56 you'll find it convenient that the source distribution
57 contains the result of running <command>happy</command> on
58 the parser specifications. If you don't want to alter the
59 parser then this saves you having to find and install
60 <command>happy</command>. You will still need a working
61 version of GHC (preferably version 4.08+) on your machine in
62 order to compile (most of) the sources, however.</para>
67 <term>The CVS repository.</term>
68 <indexterm><primary>CVS repository</primary>
71 <para>We make releases infrequently. If you want more
72 up-to-the minute (but less tested) source code then you need
73 to get access to our CVS repository.</para>
75 <para>All the <literal>fptools</literal> source code is held
76 in a CVS repository. CVS is a pretty good source-code
77 control system, and best of all it works over the
80 <para>The repository holds source code only. It holds no
81 mechanically generated files at all. So if you check out a
82 source tree from CVS you will need to install every utility
83 so that you can build all the derived files from
86 <para>More information about our CVS repository can be found
87 in <xref linkend="sec-cvs">.</para>
92 <para>If you are going to do any building from sources (either
93 from a source distribution or the CVS repository) then you need to
94 read all of this manual in detail.</para>
98 <title>Using the CVS repository</title>
100 <para>We use <ulink url="http://www.cvshome.org/">CVS</ulink> (Concurrent Version System) to keep track of our
101 sources for various software projects. CVS lets several people
102 work on the same software at the same time, allowing changes to be
103 checked in incrementally. </para>
105 <para>This section is a set of guidelines for how to use our CVS
106 repository, and will probably evolve in time. The main thing to
107 remember is that most mistakes can be undone, but if there's
108 anything you're not sure about feel free to bug the local CVS
109 meister (namely Jeff Lewis
110 <email>jlewis@galconn.com</email>). </para>
112 <sect2 id="cvs-access">
113 <title>Getting access to the CVS Repository</title>
115 <para>You can access the repository in one of two ways:
116 read-only (<xref linkend="cvs-read-only">), or read-write (<xref
117 linkend="cvs-read-write">).</para>
119 <sect3 id="cvs-read-only">
120 <title>Remote Read-only CVS Access</title>
122 <para>Read-only access is available to anyone - there's no
123 need to ask us first. With read-only CVS access you can do
124 anything except commit changes to the repository. You can
125 make changes to your local tree, and still use CVS's merge
126 facility to keep your tree up to date, and you can generate
127 patches using 'cvs diff' in order to send to us for
130 <para>To get read-only access to the repository:</para>
134 <para>Make sure that <application>cvs</application> is
135 installed on your machine.</para>
138 <para>Set your <literal>$CVSROOT</literal> environment variable to
139 <literal>:pserver:anoncvs@glass.cse.ogi.edu:/cvs</literal></para>
142 <para>Run the command</para>
146 <para>The password is simply <literal>cvs</literal>. This
147 sets up a file in your home directory called
148 <literal>.cvspass</literal>, which squirrels away the
149 dummy password, so you only need to do this step once.</para>
153 <para>Now go to <xref linkend="cvs-first">.</para>
158 <sect3 id="cvs-read-write">
159 <title>Remote Read-Write CVS Access</title>
161 <para>We generally supply read-write access to folk doing
162 serious development on some part of the source tree, when
163 going through us would be a pain. If you're developing some
164 feature, or think you have the time and inclination to fix
165 bugs in our sources, feel free to ask for read-write
166 access. There is a certain amount of responsibility that goes
167 with commit privileges; we are more likely to grant you access
168 if you've demonstrated your competence by sending us patches
169 via mail in the past.</para>
171 <para>To get remote read-write CVS access, you need to do the
172 following steps.</para>
176 <para>Make sure that <literal>cvs</literal> and
177 <literal>ssh</literal> are both installed on your
182 <para>Generate a DSA private-key/public-key pair, thus:</para>
186 <para>(<literal>ssh-keygen</literal> comes with
187 <literal>ssh</literal>.) Running <literal>ssh-keygen
188 -d</literal> creates the private and public keys in
189 <literal>$HOME/.ssh/id_dsa</literal> and
190 <literal>$HOME/.ssh/id_dsa.pub</literal> respectively
191 (assuming you accept the standard defaults).</para>
193 <para><literal>ssh-keygen -d</literal> will only work if
194 you have Version 2 <literal>ssh</literal> installed; it
195 will fail harmlessly otherwise. If you only have Version
196 1 you can instead generate an RSA key pair using plain</para>
201 <para>Doing so creates the private and public RSA keys in
202 <literal>$HOME/.ssh/identity</literal> and
203 <literal>$HOME/.ssh/identity.pub</literal>
206 <para>[Deprecated.] Incidentally, you can force a Version
207 2 <literal>ssh</literal> to use the Version 1 protocol by
208 creating <literal>$HOME/config</literal> with the
209 following in it:</para>
217 <para>In both cases, <literal>ssh-keygen</literal> will
218 ask for a <firstterm>passphrase</firstterm>. The
219 passphrase is a password that protects your private key.
220 In response to the 'Enter passphrase' question, you can
224 <para>[Recommended.] Enter a passphrase, which you
225 will quote each time you use CVS.
226 <literal>ssh-agent</literal> makes this entirely
230 <para>[Deprecated.] Just hit return (i.e. use an empty
231 passphrase); then you won't need to quote the
232 passphrase when using CVS. The downside is that
233 anyone who can see into your <literal>.ssh</literal>
234 directory, and thereby get your private key, can mess
235 up the repository. So you must keep the
236 <literal>.ssh</literal> directory with draconian
237 no-access permissions.</para>
243 [Windows users.] The programs <command>ssh-keygen1</command>, <command>ssh1</command>, and <command>cvs</command>,
244 seem to lock up <command>bash</command> entirely if they try to get user input (e.g. if
245 they ask for a password). To solve this, start up <filename>cmd.exe</filename>
246 and run it as follows:
248 c:\tmp> set CYGWIN32=tty
249 c:\tmp> c:/user/local/bin/ssh-keygen1
252 <para>[Windows users.] To protect your
253 <literal>.ssh</literal> from access by anyone else,
254 right-click your <literal>.ssh</literal> directory, and
255 select <literal>Properties</literal>. If you are not on
256 the access control list, add yourself, and give yourself
257 full permissions (the second panel). Remove everyone else
258 from the access control list. Don't leave them there but
259 deny them access, because 'they' may be a list that
261 <para>[March 2003] In fact <command>ssh</command> 3.6.1 now seems to <emphasis>require</emphasis>
262 you to have Unix permissions 600 (read/write for owner only)
263 on the <literal>.ssh/identity</literal> file, else it
264 bombs out. For your local C drive, it seems that <literal>chmod 600 identity</literal> works,
265 but on Windows NT/XP, it doesn't work on a network drive (exact dteails obscure).
266 The solution seems to be to set the CYGWIN environment
267 variable to "<literal>ntsec neta</literal>". The CYGWIN environment variable is discussed
268 in <ulink url="http://cygwin.com/cygwin-ug-net/using-cygwinenv.html">the Cygwin User's Guide</ulink>,
269 and there are more details in <ulink url="http://cygwin.com/faq/faq_4.html#SEC44">the Cygwin FAQ</ulink>.
274 <para>Send a message to to the CVS repository
275 administrator (currently Jeff Lewis
276 <email>jeff@galconn.com</email>), containing:</para>
279 <para>Your desired user-name.</para>
282 <para>Your <literal>.ssh/id_dsa.pub</literal> (or
283 <literal>.ssh/identity.pub</literal>).</para>
286 <para>He will set up your account.</para>
290 <para>Set the following environment variables:</para>
294 <constant>$HOME</constant>: points to your home directory. This is where CVS
295 will look for its <filename>.cvsrc</filename> file.
301 <constant>$CVS_RSH</constant> to <filename>ssh</filename>
303 <para>[Windows users.] Setting your <literal>CVS_RSH</literal> to
304 <literal>ssh</literal> assumes that your CVS client
305 understands how to execute shell script
306 ("#!"s,really), which is what
307 <literal>ssh</literal> is. This may not be the case on
308 Win32 platforms, so in that case set <literal>CVS_RSH</literal> to
309 <literal>ssh1</literal>.</para>
313 <para><literal>$CVSROOT</literal> to
314 <literal>:ext:</literal><replaceable>your-username</replaceable>
315 <literal>@cvs.haskell.org:/home/cvs/root</literal>
316 where <replaceable>your-username</replaceable> is your user name on
317 <literal>cvs.haskell.org</literal>.
319 <para>The <literal>CVSROOT</literal> environment variable will
320 be recorded in the checked-out tree, so you don't need to set
321 this every time. </para>
327 <constant>$CVSEDITOR</constant>: <filename>bin/gnuclient.exe</filename>
328 if you want to use an Emacs buffer for typing in those long commit messages.
334 <constant>$SHELL</constant>: To use bash as the shell in Emacs, you need to
335 set this to point to <filename>bash.exe</filename>.
346 Put the following in <filename>$HOME/.cvsrc</filename>:
357 These are the default options for the specified CVS commands,
358 and represent better defaults than the usual ones. (Feel
359 free to change them.)
363 [Windows users.] Filenames starting with <filename>.</filename> were illegal in
364 the 8.3 DOS filesystem, but that restriction should have
365 been lifted by now (i.e., you're using VFAT or later filesystems.) If
366 you're still having problems creating it, don't worry; <filename>.cvsrc</filename> is entirely
374 <para>[Experts.] Once your account is set up, you can get
375 access from other machines without bothering Jeff, thus:</para>
378 <para>Generate a public/private key pair on the new
382 <para>Use ssh to log in to
383 <literal>cvs.haskell.org</literal>, from your old
387 <para>Add the public key for the new machine to the file
388 <literal>$HOME/ssh/authorized_keys</literal> on
389 <literal>cvs.haskell.org</literal>.
390 (<literal>authorized_keys2</literal>, I think, for Version
394 <para>Make sure that the new version of
395 <literal>authorized_keys</literal> still has 600 file
404 <sect2 id="cvs-first">
405 <title>Checking Out a Source Tree</title>
409 <para>Make sure you set your <literal>CVSROOT</literal>
410 environment variable according to either of the remote
411 methods above. The Approved Way to check out a source tree
412 is as follows:</para>
415 $ cvs checkout fpconfig
418 <para>At this point you have a new directory called
419 <literal>fptools</literal> which contains the basic stuff
420 for the fptools suite, including the configuration files and
421 some other junk. </para>
423 <para>[Windows users.] The following messages appear to be harmless:
425 setsockopt IPTOS_LOWDELAY: Invalid argument
426 setsockopt IPTOS_THROUGHPUT: Invalid argument
431 <para>You can call the fptools directory whatever you like,
432 CVS won't mind: </para>
435 $ mv fptools <replaceable>directory</replaceable>
438 <para> NB: after you've read the CVS manual you might be
439 tempted to try</para>
441 $ cvs checkout -d <replaceable>directory</replaceable> fpconfig
444 <para>instead of checking out <literal>fpconfig</literal>
445 and then renaming it. But this doesn't work, and will
446 result in checking out the entire repository instead of just
447 the <literal>fpconfig</literal> bit.</para>
449 $ cd <replaceable>directory</replaceable>
450 $ cvs checkout ghc hslibs libraries
453 <para>The second command here checks out the relevant
454 modules you want to work on. For a GHC build, for instance,
455 you need at least the <literal>ghc</literal>,
456 <literal>hslibs</literal> and <literal>libraries</literal>
457 modules (for a full list of the projects available, see
458 <xref linkend="projects">).</para>
460 <para>Remeber that if you do not have
461 <literal>happy</literal> installed, you need to check it out
467 <sect2 id="cvs-committing">
468 <title>Committing Changes</title>
470 <para>This is only if you have read-write access to the
471 repository. For anoncvs users, CVS will issue a "read-only
472 repository" error if you try to commit changes.</para>
476 <para>Build the software, if necessary. Unless you're just
477 working on documentation, you'll probably want to build the
478 software in order to test any changes you make.</para>
482 <para>Make changes. Preferably small ones first.</para>
486 <para>Test them. You can see exactly what changes you've
487 made by using the <literal>cvs diff</literal> command:</para>
491 <para>lists all the changes (using the
492 <literal>diff</literal> command) in and below the current
493 directory. In emacs, <literal>C-c C-v =</literal> runs
494 <literal>cvs diff</literal> on the current buffer and shows
495 you the results.</para>
499 <para>Before checking in a change, you need to update your
506 <para>This pulls in any changes that other people have made,
507 and merges them with yours. If there are any conflicts, CVS
508 will tell you, and you'll have to resolve them before you
509 can check your changes in. The documentation describes what
510 to do in the event of a conflict.</para>
512 <para>It's not always necessary to do a full cvs update
513 before checking in a change, since CVS will always tell you
514 if you try to check in a file that someone else has changed.
515 However, you should still update at regular intervals to
516 avoid making changes that don't work in conjuction with
517 changes that someone else made. Keeping an eye on what goes
518 by on the mailing list can help here.</para>
522 <para>When you're happy that your change isn't going to
523 break anything, check it in. For a one-file change:</para>
526 $ cvs commit <replaceable>filename</replaceable>
529 <para>CVS will then pop up an editor for you to enter a
530 "commit message", this is just a short description
531 of what your change does, and will be kept in the history of
534 <para>If you're using emacs, simply load up the file into a
535 buffer and type <literal>C-x C-q</literal>, and emacs will
536 prompt for a commit message and then check in the file for
539 <para>For a multiple-file change, things are a bit
540 trickier. There are several ways to do this, but this is the
541 way I find easiest. First type the commit message into a
542 temporary file. Then either</para>
545 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>file_1</replaceable> .... <replaceable>file_n</replaceable>
548 <para>or, if nothing else has changed in this part of the
552 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>directory</replaceable>
555 <para>where <replaceable>directory</replaceable> is a common
556 parent directory for all your changes, and
557 <replaceable>commit-message</replaceable> is the name of the
558 file containing the commit message.</para>
560 <para>Shortly afterwards, you'll get some mail from the
561 relevant mailing list saying which files changed, and giving
562 the commit message. For a multiple-file change, you should
563 still get only <emphasis>one</emphasis> message.</para>
568 <sect2 id="cvs-update">
569 <title>Updating Your Source Tree</title>
571 <para>It can be tempting to cvs update just part of a source
572 tree to bring in some changes that someone else has made, or
573 before committing your own changes. This is NOT RECOMMENDED!
574 Quite often changes in one part of the tree are dependent on
575 changes in another part of the tree (the
576 <literal>mk/*.mk</literal> files are a good example where
577 problems crop up quite often). Having an inconsistent tree is a
578 major cause of headaches. </para>
580 <para>So, to avoid a lot of hassle, follow this recipe for
581 updating your tree: </para>
585 $ cvs update -P 2>&1 | tee log</screen>
587 <para>Look at the log file, and fix any conflicts (denoted by a
588 <quote>C</quote> in the first column). New directories may have
589 appeared in the repository; CVS doesn't check these out by
590 default, so to get new directories you have to explicitly do
592 $ cvs update -d</screen>
593 in each project subdirectory. Don't do this at the top level,
594 because then <emphasis>all</emphasis> the projects will be
597 <para>If you're using multiple build trees, then for every build
598 tree you have pointing at this source tree, you need to update
599 the links in case any new files have appeared: </para>
602 $ cd <replaceable>build-tree</replaceable>
603 $ lndir <replaceable>source-tree</replaceable>
606 <para>Some files might have been removed, so you need to remove
607 the links pointing to these non-existent files:</para>
610 $ find . -xtype l -exec rm '{}' \;
613 <para>To be <emphasis>really</emphasis> safe, you should do
616 <screen>$ gmake all</screen>
618 <para>from the top-level, to update the dependencies and build
619 any changed files. </para>
622 <sect2 id="cvs-tags">
623 <title>GHC Tag Policy</title>
625 <para>If you want to check out a particular version of GHC,
626 you'll need to know how we tag versions in the repository. The
627 policy (as of 4.04) is:</para>
631 <para>The tree is branched before every major release. The
632 branch tag is <literal>ghc-x-xx-branch</literal>, where
633 <literal>x-xx</literal> is the version number of the release
634 with the <literal>'.'</literal> replaced by a
635 <literal>'-'</literal>. For example, the 4.04 release lives
636 on <literal>ghc-4-04-branch</literal>.</para>
640 <para>The release itself is tagged with
641 <literal>ghc-x-xx</literal> (on the branch). eg. 4.06 is
642 called <literal>ghc-4-06</literal>.</para>
646 <para>We didn't always follow these guidelines, so to see
647 what tags there are for previous versions, do <literal>cvs
648 log</literal> on a file that's been around for a while (like
649 <literal>fptools/ghc/README</literal>).</para>
653 <para>So, to check out a fresh GHC 4.06 tree you would
657 $ cvs co -r ghc-4-06 fpconfig
659 $ cvs co -r ghc-4-06 ghc hslibs
663 <sect2 id="cvs-hints">
664 <title>General Hints</title>
668 <para>As a general rule: commit changes in small units,
669 preferably addressing one issue or implementing a single
670 feature. Provide a descriptive log message so that the
671 repository records exactly which changes were required to
672 implement a given feature/fix a bug. I've found this
673 <emphasis>very</emphasis> useful in the past for finding out
674 when a particular bug was introduced: you can just wind back
675 the CVS tree until the bug disappears.</para>
679 <para>Keep the sources at least *buildable* at any given
680 time. No doubt bugs will creep in, but it's quite easy to
681 ensure that any change made at least leaves the tree in a
682 buildable state. We do nightly builds of GHC to keep an eye
683 on what things work/don't work each day and how we're doing
684 in relation to previous verions. This idea is truely wrecked
685 if the compiler won't build in the first place!</para>
689 <para>To check out extra bits into an already-checked-out
690 tree, use the following procedure. Suppose you have a
691 checked-out fptools tree containing just ghc, and you want
692 to add nofib to it:</para>
703 $ cvs update -d nofib
706 <para>(the -d flag tells update to create a new
707 directory). If you just want part of the nofib suite, you
712 $ cvs checkout nofib/spectral
715 <para>This works because <literal>nofib</literal> is a
716 module in its own right, and spectral is a subdirectory of
717 the nofib module. The path argument to checkout must always
718 start with a module name. There's no equivalent form of this
719 command using <literal>update</literal>.</para>
725 <sect1 id="projects">
726 <title>What projects are there?</title>
728 <para>The <literal>fptools</literal> suite consists of several
729 <firstterm>projects</firstterm>, most of which can be downloaded,
730 built and installed individually. Each project corresponds to a
731 subdirectory in the source tree, and if checking out from CVS then
732 each project can be checked out individually by sitting in the top
733 level of your source tree and typing <command>cvs checkout
734 <replaceable>project</replaceable></command>.</para>
736 <para>Here is a list of the projects currently available:</para>
740 <term><literal>ghc</literal></term>
741 <indexterm><primary><literal>ghc</literal></primary>
742 <secondary>project</secondary></indexterm>
744 <para>The <ulink url="http://www.haskell.org/ghc/">Glasgow
745 Haskell Compiler</ulink> (minus libraries). Absolutely
746 required for building GHC.</para>
751 <term><literal>glafp-utils</literal></term>
752 <indexterm><primary><literal>glafp-utils</literal></primary><secondary>project</secondary></indexterm>
754 <para>Utility programs, some of which are used by the
755 build/installation system. Required for pretty much
761 <term><literal>green-card</literal></term>
762 <indexterm><primary><literal>green-card</literal></primary><secondary>project</secondary></indexterm>
765 url="http://www.haskell.org/greencard/">Green Card</ulink>
766 system for generating Haskell foreign function
772 <term><literal>haggis</literal></term>
773 <indexterm><primary><literal>haggis</literal></primary><secondary>project</secondary></indexterm>
776 url="http://www.dcs.gla.ac.uk/fp/software/haggis/">Haggis</ulink>
777 Haskell GUI framework.</para>
782 <term><literal>haddock</literal></term>
783 <indexterm><primary><literal>haddock</literal></primary><secondary>project</secondary></indexterm>
786 url="http://www.haskell.org/haddock/">Haddock</ulink>
787 documentation tool.</para>
792 <term><literal>happy</literal></term>
793 <indexterm><primary><literal>happy</literal></primary><secondary>project</secondary></indexterm>
796 url="http://www.haskell.org/happy/">Happy</ulink> Parser
802 <term><literal>hdirect</literal></term>
803 <indexterm><primary><literal>hdirect</literal></primary><secondary>project</secondary></indexterm>
806 url="http://www.haskell.org/hdirect/">H/Direct</ulink>
807 Haskell interoperability tool.</para>
812 <term><literal>hood</literal></term>
813 <indexterm><primary><literal>hood</literal></primary><secondary>project</secondary></indexterm>
815 <para>The <ulink url="http://www.haskell.org/hood/">Haskell
816 Object Observation Debugger</ulink>.</para>
821 <term><literal>hslibs</literal></term>
822 <indexterm><primary><literal>hslibs</literal></primary><secondary>project</secondary></indexterm>
824 <para>Supplemental libraries for GHC
825 (<emphasis>required</emphasis> for building GHC).</para>
830 <term><literal>libraries</literal></term>
831 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
833 <para>Hierarchical Haskell library suite
834 (<emphasis>required</emphasis> for building GHC).</para>
839 <term><literal>mhms</literal></term>
840 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
842 <para>The Modular Haskell Metric System.</para>
847 <term><literal>nofib</literal></term>
848 <indexterm><primary><literal>nofib</literal></primary><secondary>project</secondary></indexterm>
850 <para>The NoFib suite: A collection of Haskell programs used
851 primarily for benchmarking.</para>
856 <term><literal>testsuite</literal></term>
857 <indexterm><primary><literal>testsuite</literal></primary><secondary>project</secondary></indexterm>
859 <para>A testing framework, including GHC's regression test
865 <para>So, to build GHC you need at least the
866 <literal>ghc</literal>, <literal>libraries</literal> and
867 <literal>hslibs</literal> projects (a GHC source distribution will
868 already include the bits you need).</para>
871 <sect1 id="sec-build-checks">
872 <title>Things to check before you start</title>
874 <para>Here's a list of things to check before you get
880 <indexterm><primary>Disk space needed</primary></indexterm>
881 <para>Disk space needed: from about 100Mb for a basic GHC
882 build, up to probably 500Mb for a GHC build with everything
883 included (libraries built several different ways,
888 <para>Use an appropriate machine / operating system. <xref
889 linkend="sec-port-info"> lists the supported platforms; if
890 yours isn't amongst these then you can try porting GHC (see
891 <xref linkend="sec-porting-ghc">).</para>
895 <para>Be sure that the “pre-supposed” utilities are
896 installed. <Xref LinkEnd="sec-pre-supposed">
901 <para>If you have any problem when building or installing the
902 Glasgow tools, please check the “known pitfalls” (<Xref
903 LinkEnd="sec-build-pitfalls">). Also check the FAQ for the
904 version you're building, which is part of the User's Guide and
905 available on the <ulink URL="http://www.haskell.org/ghc/" >GHC web
908 <indexterm><primary>bugs</primary><secondary>known</secondary></indexterm>
910 <para>If you feel there is still some shortcoming in our
911 procedure or instructions, please report it.</para>
913 <para>For GHC, please see the <ulink
914 url="http://www.haskell.org/ghc/docs/latest/set/bug-reporting.html">bug-reporting
915 section of the GHC Users' Guide</ulink>, to maximise the
916 usefulness of your report.</para>
918 <indexterm><primary>bugs</primary><secondary>seporting</secondary></indexterm>
919 <para>If in doubt, please send a message to
920 <email>glasgow-haskell-bugs@haskell.org</email>.
921 <indexterm><primary>bugs</primary><secondary>mailing
922 list</secondary></indexterm></para>
927 <sect1 id="sec-port-info">
928 <title>What machines the Glasgow tools run on</title>
930 <indexterm><primary>ports</primary><secondary>GHC</secondary></indexterm>
931 <indexterm><primary>GHC</primary><secondary>ports</secondary></indexterm>
932 <indexterm><primary>platforms</primary><secondary>supported</secondary></indexterm>
934 <para>The main question is whether or not the Haskell compiler
935 (GHC) runs on your platform.</para>
937 <para>A “platform” is a
938 architecture/manufacturer/operating-system combination, such as
939 <literal>sparc-sun-solaris2</literal>. Other common ones are
940 <literal>alpha-dec-osf2</literal>,
941 <literal>hppa1.1-hp-hpux9</literal>,
942 <literal>i386-unknown-linux</literal>,
943 <literal>i386-unknown-solaris2</literal>,
944 <literal>i386-unknown-freebsd</literal>,
945 <literal>i386-unknown-cygwin32</literal>,
946 <literal>m68k-sun-sunos4</literal>,
947 <literal>mips-sgi-irix5</literal>,
948 <literal>sparc-sun-sunos4</literal>,
949 <literal>sparc-sun-solaris2</literal>,
950 <literal>powerpc-ibm-aix</literal>.</para>
952 <para>Some libraries may only work on a limited number of
953 platforms; for example, a sockets library is of no use unless the
954 operating system supports the underlying BSDisms.</para>
957 <title>What platforms the Haskell compiler (GHC) runs on</title>
959 <indexterm><primary>fully-supported platforms</primary></indexterm>
960 <indexterm><primary>native-code generator</primary></indexterm>
961 <indexterm><primary>registerised ports</primary></indexterm>
962 <indexterm><primary>unregisterised ports</primary></indexterm>
964 <para>The GHC hierarchy of Porting Goodness: (a) Best is a
965 native-code generator; (b) next best is a
966 “registerised” port; (c) the bare minimum is an
967 “unregisterised” port.
968 (“Unregisterised” is so terrible that we won't say
969 more about it).</para>
971 <para>We use Sparcs running Solaris 2.7 and x86 boxes running
972 FreeBSD and Linux, so those are the best supported platforms,
973 unsurprisingly.</para>
975 <para>Here's everything that's known about GHC ports. We
976 identify platforms by their “canonical”
977 CPU/Manufacturer/OS triple.</para>
981 <term>alpha-dec-{osf,linux,freebsd,openbsd,netbsd}:</term>
982 <indexterm><primary>alpha-dec-osf</primary></indexterm>
983 <indexterm><primary>alpha-dec-linux</primary></indexterm>
984 <indexterm><primary>alpha-dec-freebsd</primary></indexterm>
985 <indexterm><primary>alpha-dec-openbsd</primary></indexterm>
986 <indexterm><primary>alpha-dec-netbsd</primary></indexterm>
989 <para>The OSF port is currently working (as of GHC version
990 5.02.1) and well supported. The native code generator is
991 currently non-working. Other operating systems will
992 require some minor porting.</para>
997 <term>sparc-sun-sunos4</term>
998 <indexterm><primary>sparc-sun-sunos4</primary></indexterm>
1000 <para>Probably works with minor tweaks, hasn't been tested
1006 <term>sparc-sun-solaris2</term>
1007 <indexterm><primary>sparc-sun-solaris2</primary></indexterm>
1009 <para>Fully supported (at least for Solaris 2.7),
1010 including native-code generator.</para>
1015 <term>hppa1.1-hp-hpux (HP-PA boxes running HPUX 9.x)</term>
1016 <indexterm><primary>hppa1.1-hp-hpux</primary></indexterm>
1018 <para>A registerised port is available for version 4.08,
1019 but GHC hasn't been built on that platform since (as far
1020 as we know). No native-code generator.</para>
1025 <term>i386-unknown-linux (PCs running Linux, ELF binary format)</term>
1026 <indexterm><primary>i386-*-linux</primary></indexterm>
1028 <para>GHC works registerised and has a native code
1029 generator. You <Emphasis>must</Emphasis> have GCC 2.7.x
1030 or later. NOTE about <literal>glibc</literal> versions:
1031 GHC binaries built on a system running <literal>glibc
1032 2.0</literal> won't work on a system running
1033 <literal>glibc 2.1</literal>, and vice versa. In general,
1034 don't expect compatibility between
1035 <literal>glibc</literal> versions, even if the shared
1036 library version hasn't changed.</para>
1041 <term>i386-unknown-freebsd (PCs running FreeBSD 2.2 or
1043 <indexterm><primary>i386-unknown-freebsd</primary></indexterm>
1045 <para>GHC works registerised. Pre-built packages are
1046 available in the native package format, so if you just
1047 need binaries you're better off just installing the
1048 package (it might even be on your installation
1054 <term>i386-unknown-openbsd (PCs running OpenBSD)</term>
1055 <indexterm><primary>i386-unknown-openbsd</primary></indexterm>
1057 <para>Supported, with native code generator. Packages are
1058 available through the ports system in the native package
1064 <term>i386-unknown-netbsd (PCs running NetBSD and
1066 <indexterm><primary>i386-unknown-netbsd</primary></indexterm>
1068 <para>Will require some minor porting effort, but should
1069 work registerised.</para>
1074 <term>i386-unknown-mingw32 (PCs running Windows)</term>
1075 <indexterm><primary>i386-unknown-mingw32</primary></indexterm>
1077 <para>Fully supported under Win9x, WinNT, Win2k, and
1078 WinXP. Includes a native code generator. Building from
1079 source requires a recent <ulink
1080 url="http://www.cygwin.com/">Cygwin</ulink> distribution
1081 to be installed.</para>
1086 <term>ia64-unknown-linux</term>
1087 <indexterm><primary>ia64-unknown-linux</primary></indexterm>
1089 <para>GHC currently works unregisterised. A registerised
1090 port is in progress.</para>
1095 <term>mips-sgi-irix5</term>
1096 <indexterm><primary>mips-sgi-irix[5-6]</primary></indexterm>
1098 <para>Port has worked in the past, but hasn't been tested
1099 for some time (and will certainly have rotted in various
1100 ways). As usual, we don't have access to machines and
1101 there hasn't been an overwhelming demand for this port,
1102 but feel free to get in touch.</para>
1107 <term>powerpc-ibm-aix</term>
1108 <indexterm><primary>powerpc-ibm-aix</primary></indexterm>
1110 <para>Port currently doesn't work, needs some minimal
1111 porting effort. As usual, we don't have access to
1112 machines and there hasn't been an overwhelming demand for
1113 this port, but feel free to get in touch.</para>
1118 <term>powerpc-apple-darwin</term>
1119 <indexterm><primary>powerpc-apple-darwin</primary></indexterm>
1121 <para>Supported registerised. No native code
1127 <term>powerpc-apple-linux</term>
1128 <indexterm><primary>powerpc-apple-linux</primary></indexterm>
1130 <para>Not supported (yet).</para>
1135 <para>Various other systems have had GHC ported to them in the
1136 distant past, including various Motorola 68k boxes. The 68k
1137 support still remains, but porting to one of these systems will
1138 certainly be a non-trivial task.</para>
1142 <title>What machines the other tools run on</title>
1144 <para>Unless you hear otherwise, the other tools work if GHC
1150 <sect1 id="sec-pre-supposed">
1151 <title>Installing pre-supposed utilities</title>
1153 <indexterm><primary>pre-supposed utilities</primary></indexterm>
1154 <indexterm><primary>utilities, pre-supposed</primary></indexterm>
1156 <para>Here are the gory details about some utility programs you
1157 may need; <command>perl</command>, <command>gcc</command> and
1158 <command>happy</command> are the only important
1159 ones. (PVM<indexterm><primary>PVM</primary></indexterm> is
1160 important if you're going for Parallel Haskell.) The
1161 <command>configure</command><indexterm><primary>configure</primary></indexterm>
1162 script will tell you if you are missing something.</para>
1168 <indexterm><primary>pre-supposed: GHC</primary></indexterm>
1169 <indexterm><primary>GHC, pre-supposed</primary></indexterm>
1171 <para>GHC is required to build many of the tools, including
1172 GHC itself. If you need to port GHC to your platform
1173 because there isn't a binary distribution of GHC available,
1174 then see <xref linkend="sec-porting-ghc">.</para>
1176 <para>Which version of GHC you need will depend on the
1177 packages you intend to build. GHC itself will normally
1178 build using one of several older versions of itself - check
1179 the announcement or release notes for details.</para>
1185 <indexterm><primary>pre-supposed: Perl</primary></indexterm>
1186 <indexterm><primary>Perl, pre-supposed</primary></indexterm>
1188 <para><emphasis>You have to have Perl to proceed!</emphasis>
1189 Perl version 5 at least is required. GHC has been known to
1190 tickle bugs in Perl, so if you find that Perl crashes when
1191 running GHC try updating (or downgrading) your Perl
1192 installation. Versions of Perl that we use and are known to
1193 be fairly stable are 5.005 and 5.6.1.</para>
1195 <para>For Win32 platforms, you should use the binary
1196 supplied in the InstallShield (copy it to
1197 <filename>/bin</filename>). The Cygwin-supplied Perl seems
1200 <para>Perl should be put somewhere so that it can be invoked
1201 by the <literal>#!</literal> script-invoking
1202 mechanism. The full pathname may need to be less than 32
1203 characters long on some systems.</para>
1208 <term>GNU C (<command>gcc</command>)</term>
1209 <indexterm><primary>pre-supposed: GCC (GNU C
1210 compiler)</primary></indexterm> <indexterm><primary>GCC (GNU C
1211 compiler), pre-supposed</primary></indexterm>
1213 <para>We recommend using GCC version 2.95.2 on all
1214 platforms. Failing that, version 2.7.2 is stable on most
1215 platforms. Earlier versions of GCC can be assumed not to
1216 work, and versions in between 2.7.2 and 2.95.2 (including
1217 <command>egcs</command>) have varying degrees of stability
1218 depending on the platform.</para>
1220 <para>GCC 3.2 is currently known to have problems building
1221 GHC on Sparc, but is stable on x86.</para>
1223 <para>GCC 3.3 currnetly cannot be used to build GHC, due to
1224 some problems with the new C preprocessor.</para>
1226 <para>If your GCC dies with “internal error” on
1227 some GHC source file, please let us know, so we can report
1228 it and get things improved. (Exception: on iX86
1229 boxes—you may need to fiddle with GHC's
1230 <option>-monly-N-regs</option> option; see the User's
1236 <term>GNU Make</term>
1237 <indexterm><primary>make</primary><secondary>GNU</secondary>
1240 <para>The fptools build system makes heavy use of features
1241 specific to GNU <command>make</command>, so you must have
1242 this installed in order to build any of the fptools
1249 <indexterm><primary>Happy</primary></indexterm>
1251 <para>Happy is a parser generator tool for Haskell, and is
1252 used to generate GHC's parsers. Happy is written in
1253 Haskell, and is a project in the CVS repository
1254 (<literal>fptools/happy</literal>). It can be built from
1255 source, but bear in mind that you'll need GHC installed in
1256 order to build it. To avoid the chicken/egg problem,
1257 install a binary distribtion of either Happy or GHC to get
1258 started. Happy distributions are available from <ulink
1259 url="http://www.haskell.org/happy/">Happy's Web
1260 Page</ulink>.</para>
1265 <term>Autoconf</term>
1266 <indexterm><primary>pre-supposed: Autoconf</primary></indexterm>
1267 <indexterm><primary>Autoconf, pre-supposed</primary></indexterm>
1269 <para>GNU Autoconf is needed if you intend to build from the
1270 CVS sources, it is <emphasis>not</emphasis> needed if you
1271 just intend to build a standard source distribution.</para>
1273 <para>Version 2.52 or later of autoconf is required.
1274 NB. vesrion 2.13 will no longer work, as of GHC version
1277 <para>Autoconf builds the <command>configure</command>
1278 script from <filename>configure.in</filename> and
1279 <filename>aclocal.m4</filename>. If you modify either of
1280 these files, you'll need <command>autoconf</command> to
1281 rebuild <filename>configure</filename>.</para>
1286 <term><command>sed</command></term>
1287 <indexterm><primary>pre-supposed: sed</primary></indexterm>
1288 <indexterm><primary>sed, pre-supposed</primary></indexterm>
1290 <para>You need a working <command>sed</command> if you are
1291 going to build from sources. The build-configuration stuff
1292 needs it. GNU sed version 2.0.4 is no good! It has a bug
1293 in it that is tickled by the build-configuration. 2.0.5 is
1294 OK. Others are probably OK too (assuming we don't create too
1295 elaborate configure scripts.)</para>
1300 <para>One <literal>fptools</literal> project is worth a quick note
1301 at this point, because it is useful for all the others:
1302 <literal>glafp-utils</literal> contains several utilities which
1303 aren't particularly Glasgow-ish, but Occasionally Indispensable.
1304 Like <command>lndir</command> for creating symbolic link
1307 <sect2 id="pre-supposed-gph-tools">
1308 <title>Tools for building parallel GHC (GPH)</title>
1312 <term>PVM version 3:</term>
1313 <indexterm><primary>pre-supposed: PVM3 (Parallel Virtual Machine)</primary></indexterm>
1314 <indexterm><primary>PVM3 (Parallel Virtual Machine), pre-supposed</primary></indexterm>
1316 <para>PVM is the Parallel Virtual Machine on which
1317 Parallel Haskell programs run. (You only need this if you
1318 plan to run Parallel Haskell. Concurent Haskell, which
1319 runs concurrent threads on a uniprocessor doesn't need
1320 it.) Underneath PVM, you can have (for example) a network
1321 of workstations (slow) or a multiprocessor box
1324 <para>The current version of PVM is 3.3.11; we use 3.3.7.
1325 It is readily available on the net; I think I got it from
1326 <literal>research.att.com</literal>, in
1327 <filename>netlib</filename>.</para>
1329 <para>A PVM installation is slightly quirky, but easy to
1330 do. Just follow the <filename>Readme</filename>
1331 instructions.</para>
1336 <term><command>bash</command>:</term>
1337 <indexterm><primary>bash, presupposed (Parallel Haskell only)</primary></indexterm>
1339 <para>Sadly, the <command>gr2ps</command> script, used to
1340 convert “parallelism profiles” to PostScript,
1341 is written in Bash (GNU's Bourne Again shell). This bug
1342 will be fixed (someday).</para>
1348 <sect2 id="pre-supposed-other-tools">
1349 <title>Other useful tools</title>
1354 <indexterm><primary>pre-supposed: flex</primary></indexterm>
1355 <indexterm><primary>flex, pre-supposed</primary></indexterm>
1357 <para>This is a quite-a-bit-better-than-Lex lexer. Used
1358 to build a couple of utilities in
1359 <literal>glafp-utils</literal>. Depending on your
1360 operating system, the supplied <command>lex</command> may
1361 or may not work; you should get the GNU version.</para>
1366 <para>More tools are required if you want to format the documentation
1367 that comes with GHC and other fptools projects. See <xref
1368 linkend="building-docs">.</para>
1372 <sect1 id="sec-building-from-source">
1373 <title>Building from source</title>
1375 <indexterm><primary>Building from source</primary></indexterm>
1376 <indexterm><primary>Source, building from</primary></indexterm>
1378 <para>You've been rash enough to want to build some of the Glasgow
1379 Functional Programming tools (GHC, Happy, nofib, etc.) from
1380 source. You've slurped the source, from the CVS repository or
1381 from a source distribution, and now you're sitting looking at a
1382 huge mound of bits, wondering what to do next.</para>
1384 <para>Gingerly, you type <command>make</command>. Wrong
1387 <para>This rest of this guide is intended for duffers like me, who
1388 aren't really interested in Makefiles and systems configurations,
1389 but who need a mental model of the interlocking pieces so that
1390 they can make them work, extend them consistently when adding new
1391 software, and lay hands on them gently when they don't
1394 <sect2 id="quick-start">
1395 <title>Quick Start</title>
1397 <para>If you are starting from a source distribution, and just
1398 want a completely standard build, then the following should
1401 <screen>$ ./configure
1406 <para>For GHC, this will do a 2-stage bootstrap build of the
1407 compiler, with profiling libraries, and install the
1410 <para>If you want to do anything at all non-standard, or you
1411 want to do some development, read on...</para>
1414 <sect2 id="sec-source-tree">
1415 <title>Your source tree</title>
1417 <para>The source code is held in your <emphasis>source
1418 tree</emphasis>. The root directory of your source tree
1419 <emphasis>must</emphasis> contain the following directories and
1424 <para><filename>Makefile</filename>: the root
1429 <para><filename>mk/</filename>: the directory that contains
1430 the main Makefile code, shared by all the
1431 <literal>fptools</literal> software.</para>
1435 <para><filename>configure.in</filename>,
1436 <filename>config.sub</filename>,
1437 <filename>config.guess</filename>: these files support the
1438 configuration process.</para>
1442 <para><filename>install-sh</filename>.</para>
1446 <para>All the other directories are individual
1447 <emphasis>projects</emphasis> of the <literal>fptools</literal>
1448 system—for example, the Glasgow Haskell Compiler
1449 (<literal>ghc</literal>), the Happy parser generator
1450 (<literal>happy</literal>), the <literal>nofib</literal>
1451 benchmark suite, and so on. You can have zero or more of these.
1452 Needless to say, some of them are needed to build others.</para>
1454 <para>The important thing to remember is that even if you want
1455 only one project (<literal>happy</literal>, say), you must have
1456 a source tree whose root directory contains
1457 <filename>Makefile</filename>, <filename>mk/</filename>,
1458 <filename>configure.in</filename>, and the project(s) you want
1459 (<filename>happy/</filename> in this case). You cannot get by
1460 with just the <filename>happy/</filename> directory.</para>
1464 <title>Build trees</title>
1465 <indexterm><primary>build trees</primary></indexterm>
1466 <indexterm><primary>link trees, for building</primary></indexterm>
1468 <para>If you just want to build the software once on a single
1469 platform, then your source tree can also be your build tree, and
1470 you can skip the rest of this section.</para>
1472 <para>We often want to build multiple versions of our software
1473 for different architectures, or with different options
1474 (e.g. profiling). It's very desirable to share a single copy of
1475 the source code among all these builds.</para>
1477 <para>So for every source tree we have zero or more
1478 <emphasis>build trees</emphasis>. Each build tree is initially
1479 an exact copy of the source tree, except that each file is a
1480 symbolic link to the source file, rather than being a copy of
1481 the source file. There are “standard” Unix
1482 utilities that make such copies, so standard that they go by
1484 <command>lndir</command><indexterm><primary>lndir</primary></indexterm>,
1485 <command>mkshadowdir</command><indexterm><primary>mkshadowdir</primary></indexterm>
1486 are two (If you don't have either, the source distribution
1487 includes sources for the X11
1488 <command>lndir</command>—check out
1489 <filename>fptools/glafp-utils/lndir</filename>). See <Xref
1490 LinkEnd="sec-storysofar"> for a typical invocation.</para>
1492 <para>The build tree does not need to be anywhere near the
1493 source tree in the file system. Indeed, one advantage of
1494 separating the build tree from the source is that the build tree
1495 can be placed in a non-backed-up partition, saving your systems
1496 support people from backing up untold megabytes of
1497 easily-regenerated, and rapidly-changing, gubbins. The golden
1498 rule is that (with a single exception—<XRef
1499 LinkEnd="sec-build-config">) <emphasis>absolutely everything in
1500 the build tree is either a symbolic link to the source tree, or
1501 else is mechanically generated</emphasis>. It should be
1502 perfectly OK for your build tree to vanish overnight; an hour or
1503 two compiling and you're on the road again.</para>
1505 <para>You need to be a bit careful, though, that any new files
1506 you create (if you do any development work) are in the source
1507 tree, not a build tree!</para>
1509 <para>Remember, that the source files in the build tree are
1510 <emphasis>symbolic links</emphasis> to the files in the source
1511 tree. (The build tree soon accumulates lots of built files like
1512 <filename>Foo.o</filename>, as well.) You can
1513 <emphasis>delete</emphasis> a source file from the build tree
1514 without affecting the source tree (though it's an odd thing to
1515 do). On the other hand, if you <emphasis>edit</emphasis> a
1516 source file from the build tree, you'll edit the source-tree
1517 file directly. (You can set up Emacs so that if you edit a
1518 source file from the build tree, Emacs will silently create an
1519 edited copy of the source file in the build tree, leaving the
1520 source file unchanged; but the danger is that you think you've
1521 edited the source file whereas actually all you've done is edit
1522 the build-tree copy. More commonly you do want to edit the
1523 source file.)</para>
1525 <para>Like the source tree, the top level of your build tree
1526 must be (a linked copy of) the root directory of the
1527 <literal>fptools</literal> suite. Inside Makefiles, the root of
1528 your build tree is called
1529 <constant>$(FPTOOLS_TOP)</constant><indexterm><primary>FPTOOLS_TOP</primary></indexterm>.
1530 In the rest of this document path names are relative to
1531 <constant>$(FPTOOLS_TOP)</constant> unless
1532 otherwise stated. For example, the file
1533 <filename>ghc/mk/target.mk</filename> is actually
1534 <filename><constant>$(FPTOOLS_TOP)</constant>/ghc/mk/target.mk</filename>.</para>
1537 <sect2 id="sec-build-config">
1538 <title>Getting the build you want</title>
1540 <para>When you build <literal>fptools</literal> you will be
1541 compiling code on a particular <emphasis>host
1542 platform</emphasis>, to run on a particular <emphasis>target
1543 platform</emphasis> (usually the same as the host
1544 platform)<indexterm><primary>platform</primary></indexterm>.
1545 The difficulty is that there are minor differences between
1546 different platforms; minor, but enough that the code needs to be
1547 a bit different for each. There are some big differences too:
1548 for a different architecture we need to build GHC with a
1549 different native-code generator.</para>
1551 <para>There are also knobs you can turn to control how the
1552 <literal>fptools</literal> software is built. For example, you
1553 might want to build GHC optimised (so that it runs fast) or
1554 unoptimised (so that you can compile it fast after you've
1555 modified it. Or, you might want to compile it with debugging on
1556 (so that extra consistency-checking code gets included) or off.
1559 <para>All of this stuff is called the
1560 <emphasis>configuration</emphasis> of your build. You set the
1561 configuration using a three-step process.</para>
1565 <term>Step 1: get ready for configuration.</term>
1567 <para>NOTE: if you're starting from a source distribution,
1568 rather than CVS sources, you can skip this step.</para>
1570 <para>Change directory to
1571 <constant>$(FPTOOLS_TOP)</constant> and
1573 <command>autoconf</command><indexterm><primary>autoconf</primary></indexterm>
1574 (with no arguments). This GNU program converts
1575 <filename><constant>$(FPTOOLS_TOP)</constant>/configure.in</filename>
1576 to a shell script called
1577 <filename><constant>$(FPTOOLS_TOP)</constant>/configure</filename>.
1580 <para>Some projects, including GHC, have their own
1581 configure script. If there's an
1582 <constant>$(FPTOOLS_TOP)/<project>/configure.in</constant>,
1583 then you need to run <command>autoconf</command> in that
1584 directory too.</para>
1586 <para>Both these steps are completely
1587 platform-independent; they just mean that the
1588 human-written file (<filename>configure.in</filename>) can
1589 be short, although the resulting shell script,
1590 <command>configure</command>, and
1591 <filename>mk/config.h.in</filename>, are long.</para>
1596 <term>Step 2: system configuration.</term>
1598 <para>Runs the newly-created <command>configure</command>
1599 script, thus:</para>
1602 ./configure <optional><parameter>args</parameter></optional>
1605 <para><command>configure</command>'s mission is to scurry
1606 round your computer working out what architecture it has,
1607 what operating system, whether it has the
1608 <Function>vfork</Function> system call, where
1609 <command>yacc</command> is kept, whether
1610 <command>gcc</command> is available, where various obscure
1611 <literal>#include</literal> files are, whether it's a
1612 leap year, and what the systems manager had for lunch. It
1613 communicates these snippets of information in two
1620 <filename>mk/config.mk.in</filename><indexterm><primary>config.mk.in</primary></indexterm>
1622 <filename>mk/config.mk</filename><indexterm><primary>config.mk</primary></indexterm>,
1623 substituting for things between
1624 “<literal>@</literal>” brackets. So,
1625 “<literal>@HaveGcc@</literal>” will be
1626 replaced by “<literal>YES</literal>” or
1627 “<literal>NO</literal>” depending on what
1628 <command>configure</command> finds.
1629 <filename>mk/config.mk</filename> is included by every
1630 Makefile (directly or indirectly), so the
1631 configuration information is thereby communicated to
1632 all Makefiles.</para>
1636 <para> It translates
1637 <filename>mk/config.h.in</filename><indexterm><primary>config.h.in</primary></indexterm>
1639 <filename>mk/config.h</filename><indexterm><primary>config.h</primary></indexterm>.
1640 The latter is <literal>#include</literal>d by
1641 various C programs, which can thereby make use of
1642 configuration information.</para>
1646 <para><command>configure</command> takes some optional
1647 arguments. Use <literal>./configure --help</literal> to
1648 get a list of the available arguments. Here are some of
1649 the ones you might need:</para>
1653 <term><literal>--with-ghc=<parameter>path</parameter></literal></term>
1654 <indexterm><primary><literal>--with-ghc</literal></primary>
1657 <para>Specifies the path to an installed GHC which
1658 you would like to use. This compiler will be used
1659 for compiling GHC-specific code (eg. GHC itself).
1660 This option <emphasis>cannot</emphasis> be specified
1661 using <filename>build.mk</filename> (see later),
1662 because <command>configure</command> needs to
1663 auto-detect the version of GHC you're using. The
1664 default is to look for a compiler named
1665 <literal>ghc</literal> in your path.</para>
1670 <term><literal>--with-hc=<parameter>path</parameter></literal></term>
1671 <indexterm><primary><literal>--with-hc</literal></primary>
1674 <para>Specifies the path to any installed Haskell
1675 compiler. This compiler will be used for compiling
1676 generic Haskell code. The default is to use
1677 <literal>ghc</literal>.</para>
1682 <term><literal>--with-gcc=<parameter>path</parameter></literal></term>
1683 <indexterm><primary><literal>--with-gcc</literal></primary>
1686 <para>Specifies the path to the installed GCC. This
1687 compiler will be used to compile all C files,
1688 <emphasis>except</emphasis> any generated by the
1689 installed Haskell compiler, which will have its own
1690 idea of which C compiler (if any) to use. The
1691 default is to use <literal>gcc</literal>.</para>
1696 <para><command>configure</command> caches the results of
1697 its run in <filename>config.cache</filename>. Quite often
1698 you don't want that; you're running
1699 <command>configure</command> a second time because
1700 something has changed. In that case, simply delete
1701 <filename>config.cache</filename>.</para>
1706 <term>Step 3: build configuration.</term>
1708 <para>Next, you say how this build of
1709 <literal>fptools</literal> is to differ from the standard
1710 defaults by creating a new file
1711 <filename>mk/build.mk</filename><indexterm><primary>build.mk</primary></indexterm>
1712 <emphasis>in the build tree</emphasis>. This file is the
1713 one and only file you edit in the build tree, precisely
1714 because it says how this build differs from the source.
1715 (Just in case your build tree does die, you might want to
1716 keep a private directory of <filename>build.mk</filename>
1717 files, and use a symbolic link in each build tree to point
1718 to the appropriate one.) So
1719 <filename>mk/build.mk</filename> never exists in the
1720 source tree—you create one in each build tree from
1721 the template. We'll discuss what to put in it
1727 <para>And that's it for configuration. Simple, eh?</para>
1729 <para>What do you put in your build-specific configuration file
1730 <filename>mk/build.mk</filename>? <emphasis>For almost all
1731 purposes all you will do is put make variable definitions that
1732 override those in</emphasis>
1733 <filename>mk/config.mk.in</filename>. The whole point of
1734 <filename>mk/config.mk.in</filename>—and its derived
1735 counterpart <filename>mk/config.mk</filename>—is to define
1736 the build configuration. It is heavily commented, as you will
1737 see if you look at it. So generally, what you do is look at
1738 <filename>mk/config.mk.in</filename>, and add definitions in
1739 <filename>mk/build.mk</filename> that override any of the
1740 <filename>config.mk</filename> definitions that you want to
1741 change. (The override occurs because the main boilerplate file,
1742 <filename>mk/boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>,
1743 includes <filename>build.mk</filename> after
1744 <filename>config.mk</filename>.)</para>
1746 <para>For your convenience, there's a file called <filename>build.mk.sample</filename>
1747 that can serve as a starting point for your <filename>build.mk</filename>.</para>
1749 <para>For example, <filename>config.mk.in</filename> contains
1750 the definition:</para>
1753 GhcHcOpts=-O -Rghc-timing
1756 <para>The accompanying comment explains that this is the list of
1757 flags passed to GHC when building GHC itself. For doing
1758 development, it is wise to add <literal>-DDEBUG</literal>, to
1759 enable debugging code. So you would add the following to
1760 <filename>build.mk</filename>:</para>
1762 <para>or, if you prefer,</para>
1765 GhcHcOpts += -DDEBUG
1768 <para>GNU <command>make</command> allows existing definitions to
1769 have new text appended using the “<literal>+=</literal>”
1770 operator, which is quite a convenient feature.)</para>
1772 <para>If you want to remove the <literal>-O</literal> as well (a
1773 good idea when developing, because the turn-around cycle gets a
1774 lot quicker), you can just override
1775 <literal>GhcLibHcOpts</literal> altogether:</para>
1778 GhcHcOpts=-DDEBUG -Rghc-timing
1781 <para>When reading <filename>config.mk.in</filename>, remember
1782 that anything between “@...@” signs is going to be substituted
1783 by <command>configure</command> later. You
1784 <emphasis>can</emphasis> override the resulting definition if
1785 you want, but you need to be a bit surer what you are doing.
1786 For example, there's a line that says:</para>
1792 <para>This defines the Make variables <constant>YACC</constant>
1793 to the pathname for a <command>yacc</command> that
1794 <command>configure</command> finds somewhere. If you have your
1795 own pet <command>yacc</command> you want to use instead, that's
1796 fine. Just add this line to <filename>mk/build.mk</filename>:</para>
1802 <para>You do not <emphasis>have</emphasis> to have a
1803 <filename>mk/build.mk</filename> file at all; if you don't,
1804 you'll get all the default settings from
1805 <filename>mk/config.mk.in</filename>.</para>
1807 <para>You can also use <filename>build.mk</filename> to override
1808 anything that <command>configure</command> got wrong. One place
1809 where this happens often is with the definition of
1810 <constant>FPTOOLS_TOP_ABS</constant>: this
1811 variable is supposed to be the canonical path to the top of your
1812 source tree, but if your system uses an automounter then the
1813 correct directory is hard to find automatically. If you find
1814 that <command>configure</command> has got it wrong, just put the
1815 correct definition in <filename>build.mk</filename>.</para>
1819 <sect2 id="sec-storysofar">
1820 <title>The story so far</title>
1822 <para>Let's summarise the steps you need to carry to get
1823 yourself a fully-configured build tree from scratch.</para>
1827 <para> Get your source tree from somewhere (CVS repository
1828 or source distribution). Say you call the root directory
1829 <filename>myfptools</filename> (it does not have to be
1830 called <filename>fptools</filename>). Make sure that you
1831 have the essential files (see <XRef
1832 LinkEnd="sec-source-tree">).</para>
1837 <para>(Optional) Use <command>lndir</command> or
1838 <command>mkshadowdir</command> to create a build tree.</para>
1842 $ mkshadowdir . /scratch/joe-bloggs/myfptools-sun4
1845 <para>(N.B. <command>mkshadowdir</command>'s first argument
1846 is taken relative to its second.) You probably want to give
1847 the build tree a name that suggests its main defining
1848 characteristic (in your mind at least), in case you later
1853 <para>Change directory to the build tree. Everything is
1854 going to happen there now.</para>
1857 $ cd /scratch/joe-bloggs/myfptools-sun4
1863 <para>Prepare for system configuration:</para>
1869 <para>(You can skip this step if you are starting from a
1870 source distribution, and you already have
1871 <filename>configure</filename> and
1872 <filename>mk/config.h.in</filename>.)</para>
1874 <para>Some projects, including GHC itself, have their own
1875 configure scripts, so it is necessary to run autoconf again
1876 in the appropriate subdirectories. eg:</para>
1879 $ (cd ghc; autoconf)
1884 <para>Do system configuration:</para>
1890 <para>Don't forget to check whether you need to add any
1891 arguments to <literal>configure</literal>; for example, a
1892 common requirement is to specify which GHC to use with
1893 <option>--with-ghc=<replaceable>ghc</replaceable></option>.</para>
1897 <para>Create the file <filename>mk/build.mk</filename>,
1898 adding definitions for your desired configuration
1907 <para>You can make subsequent changes to
1908 <filename>mk/build.mk</filename> as often as you like. You do
1909 not have to run any further configuration programs to make these
1910 changes take effect. In theory you should, however, say
1911 <command>gmake clean</command>, <command>gmake all</command>,
1912 because configuration option changes could affect
1913 anything—but in practice you are likely to know what's
1918 <title>Making things</title>
1920 <para>At this point you have made yourself a fully-configured
1921 build tree, so you are ready to start building real
1924 <para>The first thing you need to know is that <emphasis>you
1925 must use GNU <command>make</command>, usually called
1926 <command>gmake</command>, not standard Unix
1927 <command>make</command></emphasis>. If you use standard Unix
1928 <command>make</command> you will get all sorts of error messages
1929 (but no damage) because the <literal>fptools</literal>
1930 <command>Makefiles</command> use GNU <command>make</command>'s
1931 facilities extensively.</para>
1933 <para>To just build the whole thing, <command>cd</command> to
1934 the top of your <literal>fptools</literal> tree and type
1935 <command>gmake</command>. This will prepare the tree and build
1936 the various projects in the correct order.</para>
1939 <sect2 id="sec-bootstrapping">
1940 <title>Bootstrapping GHC</title>
1942 <para>GHC requires a 2-stage bootstrap in order to provide
1943 full functionality, including GHCi. By a 2-stage bootstrap, we
1944 mean that the compiler is built once using the installed GHC,
1945 and then again using the compiler built in the first stage. You
1946 can also build a stage 3 compiler, but this normally isn't
1947 necessary except to verify that the stage 2 compiler is working
1950 <para>Note that when doing a bootstrap, the stage 1 compiler
1951 must be built, followed by the runtime system and libraries, and
1952 then the stage 2 compiler. The correct ordering is implemented
1953 by the top-level fptools <filename>Makefile</filename>, so if
1954 you want everything to work automatically it's best to start
1955 <command>make</command> from the top of the tree. When building
1956 GHC, the top-level fptools <filename>Makefile</filename> is set
1957 up to do a 2-stage bootstrap by default (when you say
1958 <command>make</command>). Some other targets it supports
1965 <para>Build everything as normal, including the stage 1
1973 <para>Build the stage 2 compiler only.</para>
1980 <para>Build the stage 3 compiler only.</para>
1985 <term>bootstrap</term> <term>bootstrap2</term>
1987 <para>Build stage 1 followed by stage 2.</para>
1992 <term>bootstrap3</term>
1994 <para>Build stages 1, 2 and 3.</para>
1999 <term>install</term>
2001 <para>Install everything, including the compiler built in
2002 stage 2. To override the stage, say <literal>make install
2003 stage=<replaceable>n</replaceable></literal> where
2004 <replaceable>n</replaceable> is the stage to install.</para>
2009 <para>The top-level <filename>Makefile</filename> also arranges
2010 to do the appropriate <literal>make boot</literal> steps (see
2011 below) before actually building anything.</para>
2013 <para>The <literal>stage1</literal>, <literal>stage2</literal>
2014 and <literal>stage3</literal> targets also work in the
2015 <literal>ghc/compiler</literal> directory, but don't forget that
2016 each stage requires its own <literal>make boot</literal> step:
2017 for example, you must do</para>
2019 <screen>$ make boot stage=2</screen>
2021 <para>before <literal>make stage2</literal> in
2022 <literal>ghc/compiler</literal>.</para>
2025 <sect2 id="sec-standard-targets">
2026 <title>Standard Targets</title>
2027 <indexterm><primary>targets, standard makefile</primary></indexterm>
2028 <indexterm><primary>makefile targets</primary></indexterm>
2030 <para>In any directory you should be able to make the following:</para>
2034 <term><literal>boot</literal></term>
2036 <para>does the one-off preparation required to get ready
2037 for the real work. Notably, it does <command>gmake
2038 depend</command> in all directories that contain programs.
2039 It also builds the necessary tools for compilation to
2042 <para>Invoking the <literal>boot</literal> target
2043 explicitly is not normally necessary. From the top-level
2044 <literal>fptools</literal> directory, invoking
2045 <literal>gmake</literal> causes <literal>gmake boot
2046 all</literal> to be invoked in each of the project
2047 subdirectories, in the order specified by
2048 <literal>$(AllTargets)</literal> in
2049 <literal>config.mk</literal>.</para>
2051 <para>If you're working in a subdirectory somewhere and
2052 need to update the dependencies, <literal>gmake
2053 boot</literal> is a good way to do it.</para>
2058 <term><literal>all</literal></term>
2060 <para>makes all the final target(s) for this Makefile.
2061 Depending on which directory you are in a “final
2062 target” may be an executable program, a library
2063 archive, a shell script, or a Postscript file. Typing
2064 <command>gmake</command> alone is generally the same as
2065 typing <command>gmake all</command>.</para>
2070 <term><literal>install</literal></term>
2072 <para>installs the things built by <literal>all</literal>
2073 (except for the documentation). Where does it install
2074 them? That is specified by
2075 <filename>mk/config.mk.in</filename>; you can override it
2076 in <filename>mk/build.mk</filename>, or by running
2077 <command>configure</command> with command-line arguments
2078 like <literal>--bindir=/home/simonpj/bin</literal>; see
2079 <literal>./configure --help</literal> for the full
2085 <term><literal>install-docs</literal></term>
2087 <para>installs the documentation. Otherwise behaves just
2088 like <literal>install</literal>.</para>
2093 <term><literal>uninstall</literal></term>
2095 <para>reverses the effect of
2096 <literal>install</literal>.</para>
2101 <term><literal>clean</literal></term>
2103 <para>Delete all files from the current directory that are
2104 normally created by building the program. Don't delete
2105 the files that record the configuration, or files
2106 generated by <command>gmake boot</command>. Also preserve
2107 files that could be made by building, but normally aren't
2108 because the distribution comes with them.</para>
2113 <term><literal>distclean</literal></term>
2115 <para>Delete all files from the current directory that are
2116 created by configuring or building the program. If you
2117 have unpacked the source and built the program without
2118 creating any other files, <literal>make
2119 distclean</literal> should leave only the files that were
2120 in the distribution.</para>
2125 <term><literal>mostlyclean</literal></term>
2127 <para>Like <literal>clean</literal>, but may refrain from
2128 deleting a few files that people normally don't want to
2134 <term><literal>maintainer-clean</literal></term>
2136 <para>Delete everything from the current directory that
2137 can be reconstructed with this Makefile. This typically
2138 includes everything deleted by
2139 <literal>distclean</literal>, plus more: C source files
2140 produced by Bison, tags tables, Info files, and so
2143 <para>One exception, however: <literal>make
2144 maintainer-clean</literal> should not delete
2145 <filename>configure</filename> even if
2146 <filename>configure</filename> can be remade using a rule
2147 in the <filename>Makefile</filename>. More generally,
2148 <literal>make maintainer-clean</literal> should not delete
2149 anything that needs to exist in order to run
2150 <filename>configure</filename> and then begin to build the
2156 <term><literal>check</literal></term>
2158 <para>run the test suite.</para>
2163 <para>All of these standard targets automatically recurse into
2164 sub-directories. Certain other standard targets do not:</para>
2168 <term><literal>configure</literal></term>
2170 <para>is only available in the root directory
2171 <constant>$(FPTOOLS_TOP)</constant>; it has
2172 been discussed in <XRef
2173 LinkEnd="sec-build-config">.</para>
2178 <term><literal>depend</literal></term>
2180 <para>make a <filename>.depend</filename> file in each
2181 directory that needs it. This <filename>.depend</filename>
2182 file contains mechanically-generated dependency
2183 information; for example, suppose a directory contains a
2184 Haskell source module <filename>Foo.lhs</filename> which
2185 imports another module <literal>Baz</literal>. Then the
2186 generated <filename>.depend</filename> file will contain
2187 the dependency:</para>
2193 <para>which says that the object file
2194 <filename>Foo.o</filename> depends on the interface file
2195 <filename>Baz.hi</filename> generated by compiling module
2196 <literal>Baz</literal>. The <filename>.depend</filename>
2197 file is automatically included by every Makefile.</para>
2202 <term><literal>binary-dist</literal></term>
2204 <para>make a binary distribution. This is the target we
2205 use to build the binary distributions of GHC and
2211 <term><literal>dist</literal></term>
2213 <para>make a source distribution. Note that this target
2214 does “make distclean” as part of its work;
2215 don't use it if you want to keep what you've built.</para>
2220 <para>Most <filename>Makefile</filename>s have targets other
2221 than these. You can discover them by looking in the
2222 <filename>Makefile</filename> itself.</para>
2226 <title>Using a project from the build tree</title>
2228 <para>If you want to build GHC (say) and just use it direct from
2229 the build tree without doing <literal>make install</literal>
2230 first, you can run the in-place driver script:
2231 <filename>ghc/compiler/ghc-inplace</filename>.</para>
2233 <para> Do <emphasis>NOT</emphasis> use
2234 <filename>ghc/compiler/ghc</filename>, or
2235 <filename>ghc/compiler/ghc-6.xx</filename>, as these are the
2236 scripts intended for installation, and contain hard-wired paths
2237 to the installed libraries, rather than the libraries in the
2240 <para>Happy can similarly be run from the build tree, using
2241 <filename>happy/src/happy-inplace</filename>.</para>
2245 <title>Fast Making</title>
2247 <indexterm><primary>fastmake</primary></indexterm>
2248 <indexterm><primary>dependencies, omitting</primary></indexterm>
2249 <indexterm><primary>FAST, makefile variable</primary></indexterm>
2251 <para>Sometimes the dependencies get in the way: if you've made
2252 a small change to one file, and you're absolutely sure that it
2253 won't affect anything else, but you know that
2254 <command>make</command> is going to rebuild everything anyway,
2255 the following hack may be useful:</para>
2261 <para>This tells the make system to ignore dependencies and just
2262 build what you tell it to. In other words, it's equivalent to
2263 temporarily removing the <filename>.depend</filename> file in
2264 the current directory (where <command>mkdependHS</command> and
2265 friends store their dependency information).</para>
2267 <para>A bit of history: GHC used to come with a
2268 <command>fastmake</command> script that did the above job, but
2269 GNU make provides the features we need to do it without
2270 resorting to a script. Also, we've found that fastmaking is
2271 less useful since the advent of GHC's recompilation checker (see
2272 the User's Guide section on "Separate Compilation").</para>
2276 <sect1 id="sec-makefile-arch">
2277 <title>The <filename>Makefile</filename> architecture</title>
2278 <indexterm><primary>makefile architecture</primary></indexterm>
2280 <para><command>make</command> is great if everything
2281 works—you type <command>gmake install</command> and lo! the
2282 right things get compiled and installed in the right places. Our
2283 goal is to make this happen often, but somehow it often doesn't;
2284 instead some weird error message eventually emerges from the
2285 bowels of a directory you didn't know existed.</para>
2287 <para>The purpose of this section is to give you a road-map to
2288 help you figure out what is going right and what is going
2292 <title>Debugging</title>
2294 <para>Debugging <filename>Makefile</filename>s is something of a
2295 black art, but here's a couple of tricks that we find
2296 particularly useful. The following command allows you to see
2297 the contents of any make variable in the context of the current
2298 <filename>Makefile</filename>:</para>
2300 <screen>$ make show VALUE=HS_SRCS</screen>
2302 <para>where you can replace <literal>HS_SRCS</literal> with the
2303 name of any variable you wish to see the value of.</para>
2305 <para>GNU make has a <option>-d</option> option which generates
2306 a dump of the decision procedure used to arrive at a conclusion
2307 about which files should be recompiled. Sometimes useful for
2308 tracking down problems with superfluous or missing
2309 recompilations.</para>
2313 <title>A small project</title>
2315 <para>To get started, let us look at the
2316 <filename>Makefile</filename> for an imaginary small
2317 <literal>fptools</literal> project, <literal>small</literal>.
2318 Each project in <literal>fptools</literal> has its own directory
2319 in <constant>FPTOOLS_TOP</constant>, so the
2320 <literal>small</literal> project will have its own directory
2321 <constant>FPOOLS_TOP/small/</constant>. Inside the
2322 <filename>small/</filename> directory there will be a
2323 <filename>Makefile</filename>, looking something like
2326 <indexterm><primary>Makefile, minimal</primary></indexterm>
2329 # Makefile for fptools project "small"
2332 include $(TOP)/mk/boilerplate.mk
2334 SRCS = $(wildcard *.lhs) $(wildcard *.c)
2337 include $(TOP)/target.mk
2340 <para>this <filename>Makefile</filename> has three
2345 <para>The first section includes
2348 One of the most important
2349 features of GNU <command>make</command> that we use is the ability for a <filename>Makefile</filename> to
2350 include another named file, very like <command>cpp</command>'s <literal>#include</literal>
2355 a file of “boilerplate” code from the level
2356 above (which in this case will be
2357 <filename><constant>FPTOOLS_TOP</constant>/mk/boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>).
2358 As its name suggests, <filename>boilerplate.mk</filename>
2359 consists of a large quantity of standard
2360 <filename>Makefile</filename> code. We discuss this
2361 boilerplate in more detail in <XRef LinkEnd="sec-boiler">.
2362 <indexterm><primary>include, directive in
2363 Makefiles</primary></indexterm> <indexterm><primary>Makefile
2364 inclusion</primary></indexterm></para>
2366 <para>Before the <literal>include</literal> statement, you
2367 must define the <command>make</command> variable
2368 <constant>TOP</constant><indexterm><primary>TOP</primary></indexterm>
2369 to be the directory containing the <filename>mk</filename>
2370 directory in which the <filename>boilerplate.mk</filename>
2371 file is. It is <emphasis>not</emphasis> OK to simply say</para>
2374 include ../mk/boilerplate.mk # NO NO NO
2378 <para>Why? Because the <filename>boilerplate.mk</filename>
2379 file needs to know where it is, so that it can, in turn,
2380 <literal>include</literal> other files. (Unfortunately,
2381 when an <literal>include</literal>d file does an
2382 <literal>include</literal>, the filename is treated relative
2383 to the directory in which <command>gmake</command> is being
2384 run, not the directory in which the
2385 <literal>include</literal>d sits.) In general,
2386 <emphasis>every file <filename>foo.mk</filename> assumes
2388 <filename><constant>$(TOP)</constant>/mk/foo.mk</filename>
2389 refers to itself.</emphasis> It is up to the
2390 <filename>Makefile</filename> doing the
2391 <literal>include</literal> to ensure this is the case.</para>
2393 <para>Files intended for inclusion in other
2394 <filename>Makefile</filename>s are written to have the
2395 following property: <emphasis>after
2396 <filename>foo.mk</filename> is <literal>include</literal>d,
2397 it leaves <constant>TOP</constant> containing the same value
2398 as it had just before the <literal>include</literal>
2399 statement</emphasis>. In our example, this invariant
2400 guarantees that the <literal>include</literal> for
2401 <filename>target.mk</filename> will look in the same
2402 directory as that for <filename>boilerplate.mk</filename>.</para>
2406 <para> The second section defines the following standard
2407 <command>make</command> variables:
2408 <constant>SRCS</constant><indexterm><primary>SRCS</primary></indexterm>
2409 (the source files from which is to be built), and
2410 <constant>HS_PROG</constant><indexterm><primary>HS_PROG</primary></indexterm>
2411 (the executable binary to be built). We will discuss in
2412 more detail what the “standard variables” are,
2413 and how they affect what happens, in <XRef
2414 LinkEnd="sec-targets">.</para>
2416 <para>The definition for <constant>SRCS</constant> uses the
2417 useful GNU <command>make</command> construct
2418 <literal>$(wildcard $pat$)</literal><indexterm><primary>wildcard</primary></indexterm>,
2419 which expands to a list of all the files matching the
2420 pattern <literal>pat</literal> in the current directory. In
2421 this example, <constant>SRCS</constant> is set to the list
2422 of all the <filename>.lhs</filename> and
2423 <filename>.c</filename> files in the directory. (Let's
2424 suppose there is one of each, <filename>Foo.lhs</filename>
2425 and <filename>Baz.c</filename>.)</para>
2429 <para>The last section includes a second file of standard
2431 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>.
2432 It contains the rules that tell <command>gmake</command> how
2433 to make the standard targets (<Xref
2434 LinkEnd="sec-standard-targets">). Why, you ask, can't this
2435 standard code be part of
2436 <filename>boilerplate.mk</filename>? Good question. We
2437 discuss the reason later, in <Xref
2438 LinkEnd="sec-boiler-arch">.</para>
2440 <para>You do not <emphasis>have</emphasis> to
2441 <literal>include</literal> the
2442 <filename>target.mk</filename> file. Instead, you can write
2443 rules of your own for all the standard targets. Usually,
2444 though, you will find quite a big payoff from using the
2445 canned rules in <filename>target.mk</filename>; the price
2446 tag is that you have to understand what canned rules get
2447 enabled, and what they do (<Xref
2448 LinkEnd="sec-targets">).</para>
2452 <para>In our example <filename>Makefile</filename>, most of the
2453 work is done by the two <literal>include</literal>d files. When
2454 you say <command>gmake all</command>, the following things
2459 <para><command>gmake</command> figures out that the object
2460 files are <filename>Foo.o</filename> and
2461 <filename>Baz.o</filename>.</para>
2465 <para>It uses a boilerplate pattern rule to compile
2466 <filename>Foo.lhs</filename> to <filename>Foo.o</filename>
2467 using a Haskell compiler. (Which one? That is set in the
2468 build configuration.)</para>
2472 <para>It uses another standard pattern rule to compile
2473 <filename>Baz.c</filename> to <filename>Baz.o</filename>,
2474 using a C compiler. (Ditto.)</para>
2478 <para>It links the resulting <filename>.o</filename> files
2479 together to make <literal>small</literal>, using the Haskell
2480 compiler to do the link step. (Why not use
2481 <command>ld</command>? Because the Haskell compiler knows
2482 what standard libraries to link in. How did
2483 <command>gmake</command> know to use the Haskell compiler to
2484 do the link, rather than the C compiler? Because we set the
2485 variable <constant>HS_PROG</constant> rather than
2486 <constant>C_PROG</constant>.)</para>
2490 <para>All <filename>Makefile</filename>s should follow the above
2491 three-section format.</para>
2495 <title>A larger project</title>
2497 <para>Larger projects are usually structured into a number of
2498 sub-directories, each of which has its own
2499 <filename>Makefile</filename>. (In very large projects, this
2500 sub-structure might be iterated recursively, though that is
2501 rare.) To give you the idea, here's part of the directory
2502 structure for the (rather large) GHC project:</para>
2512 ...source files for documentation...
2515 ...source files for driver...
2518 parser/...source files for parser...
2519 renamer/...source files for renamer...
2523 <para>The sub-directories <filename>docs</filename>,
2524 <filename>driver</filename>, <filename>compiler</filename>, and
2525 so on, each contains a sub-component of GHC, and each has its
2526 own <filename>Makefile</filename>. There must also be a
2527 <filename>Makefile</filename> in
2528 <filename><constant>$(FPTOOLS_TOP)</constant>/ghc</filename>.
2529 It does most of its work by recursively invoking
2530 <command>gmake</command> on the <filename>Makefile</filename>s
2531 in the sub-directories. We say that
2532 <filename>ghc/Makefile</filename> is a <emphasis>non-leaf
2533 <filename>Makefile</filename></emphasis>, because it does little
2534 except organise its children, while the
2535 <filename>Makefile</filename>s in the sub-directories are all
2536 <emphasis>leaf <filename>Makefile</filename>s</emphasis>. (In
2537 principle the sub-directories might themselves contain a
2538 non-leaf <filename>Makefile</filename> and several
2539 sub-sub-directories, but that does not happen in GHC.)</para>
2541 <para>The <filename>Makefile</filename> in
2542 <filename>ghc/compiler</filename> is considered a leaf
2543 <filename>Makefile</filename> even though the
2544 <filename>ghc/compiler</filename> has sub-directories, because
2545 these sub-directories do not themselves have
2546 <filename>Makefile</filename>s in them. They are just used to
2547 structure the collection of modules that make up GHC, but all
2548 are managed by the single <filename>Makefile</filename> in
2549 <filename>ghc/compiler</filename>.</para>
2551 <para>You will notice that <filename>ghc/</filename> also
2552 contains a directory <filename>ghc/mk/</filename>. It contains
2553 GHC-specific <filename>Makefile</filename> boilerplate code.
2554 More precisely:</para>
2558 <para><filename>ghc/mk/boilerplate.mk</filename> is included
2559 at the top of <filename>ghc/Makefile</filename>, and of all
2560 the leaf <filename>Makefile</filename>s in the
2561 sub-directories. It in turn <literal>include</literal>s the
2562 main boilerplate file
2563 <filename>mk/boilerplate.mk</filename>.</para>
2567 <para><filename>ghc/mk/target.mk</filename> is
2568 <literal>include</literal>d at the bottom of
2569 <filename>ghc/Makefile</filename>, and of all the leaf
2570 <filename>Makefile</filename>s in the sub-directories. It
2571 in turn <literal>include</literal>s the file
2572 <filename>mk/target.mk</filename>.</para>
2576 <para>So these two files are the place to look for GHC-wide
2577 customisation of the standard boilerplate.</para>
2580 <sect2 id="sec-boiler-arch">
2581 <title>Boilerplate architecture</title>
2582 <indexterm><primary>boilerplate architecture</primary></indexterm>
2584 <para>Every <filename>Makefile</filename> includes a
2585 <filename>boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>
2586 file at the top, and
2587 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>
2588 file at the bottom. In this section we discuss what is in these
2589 files, and why there have to be two of them. In general:</para>
2593 <para><filename>boilerplate.mk</filename> consists of:</para>
2597 <para><emphasis>Definitions of millions of
2598 <command>make</command> variables</emphasis> that
2599 collectively specify the build configuration. Examples:
2600 <constant>HC_OPTS</constant><indexterm><primary>HC_OPTS</primary></indexterm>,
2601 the options to feed to the Haskell compiler;
2602 <constant>NoFibSubDirs</constant><indexterm><primary>NoFibSubDirs</primary></indexterm>,
2603 the sub-directories to enable within the
2604 <literal>nofib</literal> project;
2605 <constant>GhcWithHc</constant><indexterm><primary>GhcWithHc</primary></indexterm>,
2606 the name of the Haskell compiler to use when compiling
2607 GHC in the <literal>ghc</literal> project.</para>
2611 <para><emphasis>Standard pattern rules</emphasis> that
2612 tell <command>gmake</command> how to construct one file
2613 from another.</para>
2617 <para><filename>boilerplate.mk</filename> needs to be
2618 <literal>include</literal>d at the <emphasis>top</emphasis>
2619 of each <filename>Makefile</filename>, so that the user can
2620 replace the boilerplate definitions or pattern rules by
2621 simply giving a new definition or pattern rule in the
2622 <filename>Makefile</filename>. <command>gmake</command>
2623 simply takes the last definition as the definitive one.</para>
2625 <para>Instead of <emphasis>replacing</emphasis> boilerplate
2626 definitions, it is also quite common to
2627 <emphasis>augment</emphasis> them. For example, a
2628 <filename>Makefile</filename> might say:</para>
2634 <para>thereby adding “<option>-O</option>” to
2636 <constant>SRC_HC_OPTS</constant><indexterm><primary>SRC_HC_OPTS</primary></indexterm>.</para>
2640 <para><filename>target.mk</filename> contains
2641 <command>make</command> rules for the standard targets
2642 described in <Xref LinkEnd="sec-standard-targets">. These
2643 rules are selectively included, depending on the setting of
2644 certain <command>make</command> variables. These variables
2645 are usually set in the middle section of the
2646 <filename>Makefile</filename> between the two
2647 <literal>include</literal>s.</para>
2649 <para><filename>target.mk</filename> must be included at the
2650 end (rather than being part of
2651 <filename>boilerplate.mk</filename>) for several tiresome
2657 <para><command>gmake</command> commits target and
2658 dependency lists earlier than it should. For example,
2659 <FIlename>target.mk</FIlename> has a rule that looks
2663 $(HS_PROG) : $(OBJS)
2664 $(HC) $(LD_OPTS) $< -o $@
2667 <para>If this rule was in
2668 <filename>boilerplate.mk</filename> then
2669 <constant>$(HS_PROG)</constant><indexterm><primary>HS_PROG</primary></indexterm>
2671 <constant>$(OBJS)</constant><indexterm><primary>OBJS</primary></indexterm>
2672 would not have their final values at the moment
2673 <command>gmake</command> encountered the rule. Alas,
2674 <command>gmake</command> takes a snapshot of their
2675 current values, and wires that snapshot into the rule.
2676 (In contrast, the commands executed when the rule
2677 “fires” are only substituted at the moment
2678 of firing.) So, the rule must follow the definitions
2679 given in the <filename>Makefile</filename> itself.</para>
2683 <para>Unlike pattern rules, ordinary rules cannot be
2684 overriden or replaced by subsequent rules for the same
2685 target (at least, not without an error message).
2686 Including ordinary rules in
2687 <filename>boilerplate.mk</filename> would prevent the
2688 user from writing rules for specific targets in specific
2693 <para>There are a couple of other reasons I've
2694 forgotten, but it doesn't matter too much.</para>
2701 <sect2 id="sec-boiler">
2702 <title>The main <filename>mk/boilerplate.mk</filename> file</title>
2703 <indexterm><primary>boilerplate.mk</primary></indexterm>
2705 <para>If you look at
2706 <filename><constant>$(FPTOOLS_TOP)</constant>/mk/boilerplate.mk</filename>
2707 you will find that it consists of the following sections, each
2708 held in a separate file:</para>
2712 <term><filename>config.mk</filename></term>
2713 <indexterm><primary>config.mk</primary></indexterm>
2715 <para>is the build configuration file we discussed at
2716 length in <Xref LinkEnd="sec-build-config">.</para>
2721 <term><filename>paths.mk</filename></term>
2722 <indexterm><primary>paths.mk</primary></indexterm>
2724 <para>defines <command>make</command> variables for
2725 pathnames and file lists. This file contains code for
2726 automatically compiling lists of source files and deriving
2727 lists of object files from those. The results can be
2728 overriden in the <filename>Makefile</filename>, but in
2729 most cases the automatic setup should do the right
2732 <para>The following variables may be set in the
2733 <filename>Makefile</filename> to affect how the automatic
2734 source file search is done:</para>
2738 <term><literal>ALL_DIRS</literal></term>
2739 <indexterm><primary><literal>ALL_DIRS</literal></primary>
2742 <para>Set to a list of directories to search in
2743 addition to the current directory for source
2749 <term><literal>EXCLUDE_SRCS</literal></term>
2750 <indexterm><primary><literal>EXCLUDE_SRCS</literal></primary>
2753 <para>Set to a list of source files (relative to the
2754 current directory) to omit from the automatic
2755 search. The source searching machinery is clever
2756 enough to know that if you exclude a source file
2757 from which other sources are derived, then the
2758 derived sources should also be excluded. For
2759 example, if you set <literal>EXCLUDED_SRCS</literal>
2760 to include <filename>Foo.y</filename>, then
2761 <filename>Foo.hs</filename> will also be
2767 <term><literal>EXTRA_SRCS</literal></term>
2768 <indexterm><primary><literal>EXCLUDE_SRCS</literal></primary>
2771 <para>Set to a list of extra source files (perhaps
2772 in directories not listed in
2773 <literal>ALL_DIRS</literal>) that should be
2779 <para>The results of the automatic source file search are
2780 placed in the following make variables:</para>
2784 <term><literal>SRCS</literal></term>
2785 <indexterm><primary><literal>SRCS</literal></primary></indexterm>
2787 <para>All source files found, sorted and without
2788 duplicates, including those which might not exist
2789 yet but will be derived from other existing sources.
2790 <literal>SRCS</literal> <emphasis>can</emphasis> be
2791 overriden if necessary, in which case the variables
2792 below will follow suit.</para>
2797 <term><literal>HS_SRCS</literal></term>
2798 <indexterm><primary><literal>HS_SRCS</literal></primary></indexterm>
2800 <para>all Haskell source files in the current
2801 directory, including those derived from other source
2802 files (eg. Happy sources also give rise to Haskell
2808 <term><literal>HS_OBJS</literal></term>
2809 <indexterm><primary><literal>HS_OBJS</literal></primary></indexterm>
2811 <para>Object files derived from
2812 <literal>HS_SRCS</literal>.</para>
2817 <term><literal>HS_IFACES</literal></term>
2818 <indexterm><primary><literal>HS_IFACES</literal></primary></indexterm>
2820 <para>Interface files (<literal>.hi</literal> files)
2821 derived from <literal>HS_SRCS</literal>.</para>
2826 <term><literal>C_SRCS</literal></term>
2827 <indexterm><primary><literal>C_SRCS</literal></primary></indexterm>
2829 <para>All C source files found.</para>
2834 <term><literal>C_OBJS</literal></term>
2835 <indexterm><primary><literal>C_OBJS</literal></primary></indexterm>
2837 <para>Object files derived from
2838 <literal>C_SRCS</literal>.</para>
2843 <term><literal>SCRIPT_SRCS</literal></term>
2844 <indexterm><primary><literal>SCRIPT_SRCS</literal></primary></indexterm>
2846 <para>All script source files found
2847 (<literal>.lprl</literal> files).</para>
2852 <term><literal>SCRIPT_OBJS</literal></term>
2853 <indexterm><primary><literal>SCRIPT_OBJS</literal></primary></indexterm>
2855 <para><quote>object</quote> files derived from
2856 <literal>SCRIPT_SRCS</literal>
2857 (<literal>.prl</literal> files).</para>
2862 <term><literal>HSC_SRCS</literal></term>
2863 <indexterm><primary><literal>HSC_SRCS</literal></primary></indexterm>
2865 <para>All <literal>hsc2hs</literal> source files
2866 (<literal>.hsc</literal> files).</para>
2871 <term><literal>HAPPY_SRCS</literal></term>
2872 <indexterm><primary><literal>HAPPY_SRCS</literal></primary></indexterm>
2874 <para>All <literal>happy</literal> source files
2875 (<literal>.y</literal> or <literal>.hy</literal> files).</para>
2880 <term><literal>OBJS</literal></term>
2881 <indexterm><primary>OBJS</primary></indexterm>
2883 <para>the concatenation of
2884 <literal>$(HS_OBJS)</literal>,
2885 <literal>$(C_OBJS)</literal>, and
2886 <literal>$(SCRIPT_OBJS)</literal>.</para>
2891 <para>Any or all of these definitions can easily be
2892 overriden by giving new definitions in your
2893 <filename>Makefile</filename>.</para>
2895 <para>What, exactly, does <filename>paths.mk</filename>
2896 consider a <quote>source file</quote> to be? It's based
2897 on the file's suffix (e.g. <filename>.hs</filename>,
2898 <filename>.lhs</filename>, <filename>.c</filename>,
2899 <filename>.hy</filename>, etc), but this is the kind of
2900 detail that changes, so rather than enumerate the source
2901 suffices here the best thing to do is to look in
2902 <filename>paths.mk</filename>.</para>
2907 <term><filename>opts.mk</filename></term>
2908 <indexterm><primary>opts.mk</primary></indexterm>
2910 <para>defines <command>make</command> variables for option
2911 strings to pass to each program. For example, it defines
2912 <constant>HC_OPTS</constant><indexterm><primary>HC_OPTS</primary></indexterm>,
2913 the option strings to pass to the Haskell compiler. See
2914 <Xref LinkEnd="sec-suffix">.</para>
2919 <term><filename>suffix.mk</filename></term>
2920 <indexterm><primary>suffix.mk</primary></indexterm>
2922 <para>defines standard pattern rules—see <Xref
2923 LinkEnd="sec-suffix">.</para>
2928 <para>Any of the variables and pattern rules defined by the
2929 boilerplate file can easily be overridden in any particular
2930 <filename>Makefile</filename>, because the boilerplate
2931 <literal>include</literal> comes first. Definitions after this
2932 <literal>include</literal> directive simply override the default
2933 ones in <filename>boilerplate.mk</filename>.</para>
2936 <sect2 id="sec-suffix">
2937 <title>Pattern rules and options</title>
2938 <indexterm><primary>Pattern rules</primary></indexterm>
2941 <filename>suffix.mk</filename><indexterm><primary>suffix.mk</primary></indexterm>
2942 defines standard <emphasis>pattern rules</emphasis> that say how
2943 to build one kind of file from another, for example, how to
2944 build a <filename>.o</filename> file from a
2945 <filename>.c</filename> file. (GNU <command>make</command>'s
2946 <emphasis>pattern rules</emphasis> are more powerful and easier
2947 to use than Unix <command>make</command>'s <emphasis>suffix
2948 rules</emphasis>.)</para>
2950 <para>Almost all the rules look something like this:</para>
2955 $(CC) $(CC_OPTS) -c $< -o $@
2958 <para>Here's how to understand the rule. It says that
2959 <emphasis>something</emphasis><filename>.o</filename> (say
2960 <filename>Foo.o</filename>) can be built from
2961 <emphasis>something</emphasis><filename>.c</filename>
2962 (<filename>Foo.c</filename>), by invoking the C compiler (path
2963 name held in <constant>$(CC)</constant>), passing to it
2964 the options <constant>$(CC_OPTS)</constant> and
2965 the rule's dependent file of the rule
2966 <literal>$<</literal> (<filename>Foo.c</filename> in
2967 this case), and putting the result in the rule's target
2968 <literal>$@</literal> (<filename>Foo.o</filename> in this
2971 <para>Every program is held in a <command>make</command>
2972 variable defined in <filename>mk/config.mk</filename>—look
2973 in <filename>mk/config.mk</filename> for the complete list. One
2974 important one is the Haskell compiler, which is called
2975 <constant>$(HC)</constant>.</para>
2977 <para>Every program's options are are held in a
2978 <command>make</command> variables called
2979 <constant><prog>_OPTS</constant>. the
2980 <constant><prog>_OPTS</constant> variables are
2981 defined in <filename>mk/opts.mk</filename>. Almost all of them
2982 are defined like this:</para>
2985 CC_OPTS = $(SRC_CC_OPTS) $(WAY$(_way)_CC_OPTS) $($*_CC_OPTS) $(EXTRA_CC_OPTS)
2988 <para>The four variables from which
2989 <constant>CC_OPTS</constant> is built have the following
2994 <term><constant>SRC_CC_OPTS</constant><indexterm><primary>SRC_CC_OPTS</primary></indexterm>:</term>
2996 <para>options passed to all C compilations.</para>
3001 <term><constant>WAY_<way>_CC_OPTS</constant>:</term>
3003 <para>options passed to C compilations for way
3004 <literal><way></literal>. For example,
3005 <constant>WAY_mp_CC_OPTS</constant>
3006 gives options to pass to the C compiler when compiling way
3007 <literal>mp</literal>. The variable
3008 <constant>WAY_CC_OPTS</constant> holds
3009 options to pass to the C compiler when compiling the
3010 standard way. (<Xref LinkEnd="sec-ways"> dicusses
3011 multi-way compilation.)</para>
3016 <term><constant><module>_CC_OPTS</constant>:</term>
3018 <para>options to pass to the C compiler that are specific
3019 to module <literal><module></literal>. For example,
3020 <constant>SMap_CC_OPTS</constant> gives the
3021 specific options to pass to the C compiler when compiling
3022 <filename>SMap.c</filename>.</para>
3027 <term><constant>EXTRA_CC_OPTS</constant><indexterm><primary>EXTRA_CC_OPTS</primary></indexterm>:</term>
3029 <para>extra options to pass to all C compilations. This
3030 is intended for command line use, thus:</para>
3033 gmake libHS.a EXTRA_CC_OPTS="-v"
3040 <sect2 id="sec-targets">
3041 <title>The main <filename>mk/target.mk</filename> file</title>
3042 <indexterm><primary>target.mk</primary></indexterm>
3044 <para><filename>target.mk</filename> contains canned rules for
3045 all the standard targets described in <Xref
3046 LinkEnd="sec-standard-targets">. It is complicated by the fact
3047 that you don't want all of these rules to be active in every
3048 <filename>Makefile</filename>. Rather than have a plethora of
3049 tiny files which you can include selectively, there is a single
3050 file, <filename>target.mk</filename>, which selectively includes
3051 rules based on whether you have defined certain variables in
3052 your <filename>Makefile</filename>. This section explains what
3053 rules you get, what variables control them, and what the rules
3054 do. Hopefully, you will also get enough of an idea of what is
3055 supposed to happen that you can read and understand any weird
3056 special cases yourself.</para>
3060 <term><constant>HS_PROG</constant><indexterm><primary>HS_PROG</primary></indexterm>.</term>
3062 <para>If <constant>HS_PROG</constant> is defined,
3063 you get rules with the following targets:</para>
3067 <term><filename>HS_PROG</filename><indexterm><primary>HS_PROG</primary></indexterm></term>
3069 <para>itself. This rule links
3070 <constant>$(OBJS)</constant> with the Haskell
3071 runtime system to get an executable called
3072 <constant>$(HS_PROG)</constant>.</para>
3077 <term><literal>install</literal><indexterm><primary>install</primary></indexterm></term>
3080 <constant>$(HS_PROG)</constant> in
3081 <constant>$(bindir)</constant>.</para>
3090 <term><constant>C_PROG</constant><indexterm><primary>C_PROG</primary></indexterm></term>
3092 <para>is similar to <constant>HS_PROG</constant>,
3093 except that the link step links
3094 <constant>$(C_OBJS)</constant> with the C
3095 runtime system.</para>
3100 <term><constant>LIBRARY</constant><indexterm><primary>LIBRARY</primary></indexterm></term>
3102 <para>is similar to <constant>HS_PROG</constant>,
3103 except that it links
3104 <constant>$(LIB_OBJS)</constant> to make the
3105 library archive <constant>$(LIBRARY)</constant>,
3106 and <literal>install</literal> installs it in
3107 <constant>$(libdir)</constant>.</para>
3112 <term><constant>LIB_DATA</constant><indexterm><primary>LIB_DATA</primary></indexterm></term>
3114 <para>…</para>
3119 <term><constant>LIB_EXEC</constant><indexterm><primary>LIB_EXEC</primary></indexterm></term>
3121 <para>…</para>
3126 <term><constant>HS_SRCS</constant><indexterm><primary>HS_SRCS</primary></indexterm>, <constant>C_SRCS</constant><indexterm><primary>C_SRCS</primary></indexterm>.</term>
3128 <para>If <constant>HS_SRCS</constant> is defined
3129 and non-empty, a rule for the target
3130 <literal>depend</literal> is included, which generates
3131 dependency information for Haskell programs. Similarly
3132 for <constant>C_SRCS</constant>.</para>
3137 <para>All of these rules are “double-colon” rules,
3141 install :: $(HS_PROG)
3142 ...how to install it...
3145 <para>GNU <command>make</command> treats double-colon rules as
3146 separate entities. If there are several double-colon rules for
3147 the same target it takes each in turn and fires it if its
3148 dependencies say to do so. This means that you can, for
3149 example, define both <constant>HS_PROG</constant> and
3150 <constant>LIBRARY</constant>, which will generate two rules for
3151 <literal>install</literal>. When you type <command>gmake
3152 install</command> both rules will be fired, and both the program
3153 and the library will be installed, just as you wanted.</para>
3156 <sect2 id="sec-subdirs">
3157 <title>Recursion</title>
3158 <indexterm><primary>recursion, in makefiles</primary></indexterm>
3159 <indexterm><primary>Makefile, recursing into subdirectories</primary></indexterm>
3161 <para>In leaf <filename>Makefile</filename>s the variable
3162 <constant>SUBDIRS</constant><indexterm><primary>SUBDIRS</primary></indexterm>
3163 is undefined. In non-leaf <filename>Makefile</filename>s,
3164 <constant>SUBDIRS</constant> is set to the list of
3165 sub-directories that contain subordinate
3166 <filename>Makefile</filename>s. <emphasis>It is up to you to
3167 set <constant>SUBDIRS</constant> in the
3168 <filename>Makefile</filename>.</emphasis> There is no automation
3169 here—<constant>SUBDIRS</constant> is too important to
3172 <para>When <constant>SUBDIRS</constant> is defined,
3173 <filename>target.mk</filename> includes a rather neat rule for
3174 the standard targets (<Xref LinkEnd="sec-standard-targets"> that
3175 simply invokes <command>make</command> recursively in each of
3176 the sub-directories.</para>
3178 <para><emphasis>These recursive invocations are guaranteed to
3179 occur in the order in which the list of directories is specified
3180 in <constant>SUBDIRS</constant>. </emphasis>This guarantee can
3181 be important. For example, when you say <command>gmake
3182 boot</command> it can be important that the recursive invocation
3183 of <command>make boot</command> is done in one sub-directory
3184 (the include files, say) before another (the source files).
3185 Generally, put the most independent sub-directory first, and the
3186 most dependent last.</para>
3189 <sect2 id="sec-ways">
3190 <title>Way management</title>
3191 <indexterm><primary>way management</primary></indexterm>
3193 <para>We sometimes want to build essentially the same system in
3194 several different “ways”. For example, we want to build GHC's
3195 <literal>Prelude</literal> libraries with and without profiling,
3196 so that there is an appropriately-built library archive to link
3197 with when the user compiles his program. It would be possible
3198 to have a completely separate build tree for each such “way”,
3199 but it would be horribly bureaucratic, especially since often
3200 only parts of the build tree need to be constructed in multiple
3204 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>
3205 contains some clever magic to allow you to build several
3206 versions of a system; and to control locally how many versions
3207 are built and how they differ. This section explains the
3210 <para>The files for a particular way are distinguished by
3211 munging the suffix. The <quote>normal way</quote> is always
3212 built, and its files have the standard suffices
3213 <filename>.o</filename>, <filename>.hi</filename>, and so on.
3214 In addition, you can build one or more extra ways, each
3215 distinguished by a <emphasis>way tag</emphasis>. The object
3216 files and interface files for one of these extra ways are
3217 distinguished by their suffix. For example, way
3218 <literal>mp</literal> has files
3219 <filename>.mp_o</filename> and
3220 <filename>.mp_hi</filename>. Library archives have their
3221 way tag the other side of the dot, for boring reasons; thus,
3222 <filename>libHS_mp.a</filename>.</para>
3224 <para>A <command>make</command> variable called
3225 <constant>way</constant> holds the current way tag.
3226 <emphasis><constant>way</constant> is only ever set on the
3227 command line of <command>gmake</command></emphasis> (usually in
3228 a recursive invocation of <command>gmake</command> by the
3229 system). It is never set inside a
3230 <filename>Makefile</filename>. So it is a global constant for
3231 any one invocation of <command>gmake</command>. Two other
3232 <command>make</command> variables,
3233 <constant>way_</constant> and
3234 <constant>_way</constant> are immediately derived from
3235 <constant>$(way)</constant> and never altered. If
3236 <constant>way</constant> is not set, then neither are
3237 <constant>way_</constant> and
3238 <constant>_way</constant>, and the invocation of
3239 <command>make</command> will build the <quote>normal
3240 way</quote>. If <constant>way</constant> is set, then the other
3241 two variables are set in sympathy. For example, if
3242 <constant>$(way)</constant> is “<literal>mp</literal>”,
3243 then <constant>way_</constant> is set to
3244 “<literal>mp_</literal>” and
3245 <constant>_way</constant> is set to
3246 “<literal>_mp</literal>”. These three variables are
3247 then used when constructing file names.</para>
3249 <para>So how does <command>make</command> ever get recursively
3250 invoked with <constant>way</constant> set? There are two ways
3251 in which this happens:</para>
3255 <para>For some (but not all) of the standard targets, when
3256 in a leaf sub-directory, <command>make</command> is
3257 recursively invoked for each way tag in
3258 <constant>$(WAYS)</constant>. You set
3259 <constant>WAYS</constant> in the
3260 <filename>Makefile</filename> to the list of way tags you
3261 want these targets built for. The mechanism here is very
3262 much like the recursive invocation of
3263 <command>make</command> in sub-directories (<Xref
3264 LinkEnd="sec-subdirs">). It is up to you to set
3265 <constant>WAYS</constant> in your
3266 <filename>Makefile</filename>; this is how you control what
3267 ways will get built.</para>
3271 <para>For a useful collection of targets (such as
3272 <filename>libHS_mp.a</filename>,
3273 <filename>Foo.mp_o</filename>) there is a rule which
3274 recursively invokes <command>make</command> to make the
3275 specified target, setting the <constant>way</constant>
3276 variable. So if you say <command>gmake
3277 Foo.mp_o</command> you should see a recursive
3278 invocation <command>gmake Foo.mp_o way=mp</command>,
3279 and <emphasis>in this recursive invocation the pattern rule
3280 for compiling a Haskell file into a <filename>.o</filename>
3281 file will match</emphasis>. The key pattern rules (in
3282 <filename>suffix.mk</filename>) look like this:
3286 $(HC) $(HC_OPTS) $< -o $@
3293 <para>You can invoke <command>make</command> with a
3294 particular <literal>way</literal> setting yourself, in order
3295 to build files related to a particular
3296 <literal>way</literal> in the current directory. eg.
3302 will build files for the profiling way only in the current
3309 <title>When the canned rule isn't right</title>
3311 <para>Sometimes the canned rule just doesn't do the right thing.
3312 For example, in the <literal>nofib</literal> suite we want the
3313 link step to print out timing information. The thing to do here
3314 is <emphasis>not</emphasis> to define
3315 <constant>HS_PROG</constant> or
3316 <constant>C_PROG</constant>, and instead define a special
3317 purpose rule in your own <filename>Makefile</filename>. By
3318 using different variable names you will avoid the canned rules
3319 being included, and conflicting with yours.</para>
3323 <sect1 id="building-docs">
3324 <title>Building the documentation</title>
3326 <sect2 id="pre-supposed-doc-tools">
3327 <title>Tools for building the Documentation</title>
3329 <para>The following additional tools are required if you want to
3330 format the documentation that comes with the
3331 <literal>fptools</literal> projects:</para>
3335 <term>DocBook</term>
3336 <indexterm><primary>pre-supposed: DocBook</primary></indexterm>
3337 <indexterm><primary>DocBook, pre-supposed</primary></indexterm>
3339 <para>Much of our documentation is written in SGML, using
3340 the DocBook DTD. Instructions on installing and
3341 configuring the DocBook tools are below.</para>
3347 <indexterm><primary>pre-supposed: TeX</primary></indexterm>
3348 <indexterm><primary>TeX, pre-supposed</primary></indexterm>
3350 <para>A decent TeX distribution is required if you want to
3351 produce printable documentation. We recomment teTeX,
3352 which includes just about everything you need.</para>
3357 <term>Haddock</term>
3358 <indexterm><primary>Haddock</primary>
3361 <para>Haddock is a Haskell documentation tool that we use
3362 for automatically generating documentation from the
3363 library source code. It is an <literal>fptools</literal>
3364 project in itself. To build documentation for the
3365 libraries (<literal>fptools/libraries</literal>) you
3366 should check out and build Haddock in
3367 <literal>fptools/haddock</literal>. Haddock requires GHC
3375 <title>Installing the DocBook tools</title>
3378 <title>Installing the DocBook tools on Linux</title>
3380 <para>If you're on a recent RedHat system (7.0+), you probably
3381 have working DocBook tools already installed. The configure
3382 script should detect your setup and you're away.</para>
3384 <para>If you don't have DocBook tools installed, and you are
3385 using a system that can handle RedHat RPM packages, you can
3386 probably use the <ULink
3387 URL="http://sourceware.cygnus.com/docbook-tools/">Cygnus
3388 DocBook tools</ULink>, which is the most shrink-wrapped SGML
3389 suite that we could find. You need all the RPMs except for
3390 psgml (i.e. <Filename>docbook</Filename>,
3391 <Filename>jade</Filename>, <Filename>jadetex</Filename>,
3392 <Filename>sgmlcommon</Filename> and
3393 <Filename>stylesheets</Filename>). Note that most of these
3394 RPMs are architecture neutral, so are likely to be found in a
3395 <Filename>noarch</Filename> directory. The SuSE RPMs also
3396 work; the RedHat ones <Emphasis>don't</Emphasis> in RedHat 6.2
3397 (7.0 and later should be OK), but they are easy to fix: just
3399 <Filename>/usr/lib/sgml/stylesheets/nwalsh-modular/lib/dblib.dsl</Filename>
3400 to <Filename>/usr/lib/sgml/lib/dblib.dsl</Filename>. </para>
3404 <title>Installing DocBook on FreeBSD</title>
3406 <para>On FreeBSD systems, the easiest way to get DocBook up
3407 and running is to install it from the ports tree or a
3408 pre-compiled package (packages are available from your local
3409 FreeBSD mirror site).</para>
3411 <para>To use the ports tree, do this:
3413 $ cd /usr/ports/textproc/docproj
3416 This installs the FreeBSD documentation project tools, which
3417 includes everything needed to format the GHC
3418 documentation.</para>
3422 <title>Installing from binaries on Windows</title>
3424 <Para>It's a good idea to use Norman Walsh's <ULink
3425 URL="http://nwalsh.com/docbook/dsssl/doc/install.html">installation
3426 notes</ULink> as a guide. You should get version 3.1 of
3427 DocBook, and note that his file <Filename>test.sgm</Filename>
3428 won't work, as it needs version 3.0. You should unpack Jade
3429 into <Filename>\Jade</Filename>, along with the entities,
3430 DocBook into <Filename>\docbook</Filename>, and the DocBook
3431 stylesheets into <Filename>\docbook\stylesheets</Filename> (so
3432 they actually end up in
3433 <Filename>\docbook\stylesheets\docbook</Filename>).</para>
3438 <title>Installing the DocBook tools from source</title>
3443 <para>Install <ULink
3444 URL="http://openjade.sourceforge.net/">OpenJade</ULink>
3445 (Windows binaries are available as well as sources). If you
3446 want DVI, PS, or PDF then install JadeTeX from the
3447 <Filename>dsssl</Filename> subdirectory. (If you get the
3451 ! LaTeX Error: Unknown option implicit=false' for package hyperref'.
3454 your version of <Command>hyperref</Command> is out of date;
3455 download it from CTAN
3456 (<Filename>macros/latex/contrib/supported/hyperref</Filename>),
3457 and make it, ensuring that you have first removed or renamed
3458 your old copy. If you start getting file not found errors
3459 when making the test for <Command>hyperref</Command>, you
3460 can abort at that point and proceed straight to
3461 <Command>make install</Command>, or enter them as
3462 <Filename>../</Filename><Emphasis>filename</Emphasis>.)</para>
3464 <para>Make links from <Filename>virtex</Filename> to
3465 <Filename>jadetex</Filename> and
3466 <Filename>pdfvirtex</Filename> to
3467 <Filename>pdfjadetex</Filename> (otherwise DVI, PostScript
3468 and PDF output will not work). Copy
3469 <Filename>dsssl/*.{dtd,dsl}</Filename> and
3470 <Filename>catalog</Filename> to
3471 <Filename>/usr/[local/]lib/sgml</Filename>.</para>
3475 <title>DocBook and the DocBook stylesheets</title>
3477 <para>Get a Zip of <ULink
3478 URL="http://www.oasis-open.org/docbook/sgml/3.1/index.html">DocBook</ULink>
3479 and install the contents in
3480 <Filename>/usr/[local/]/lib/sgml</Filename>.</para>
3482 <para>Get the <ULink
3483 URL="http://nwalsh.com/docbook/dsssl/">DocBook
3484 stylesheets</ULink> and install in
3485 <Filename>/usr/[local/]lib/sgml/stylesheets</Filename>
3486 (thereby creating a subdirectory docbook). For indexing,
3487 copy or link <Filename>collateindex.pl</Filename> from the
3488 DocBook stylesheets archive in <Filename>bin</Filename> into
3489 a directory on your <Constant>PATH</Constant>.</para>
3491 <para>Download the <ULink
3492 URL="http://www.oasis-open.org/cover/ISOEnts.zip">ISO
3493 entities</ULink> into
3494 <Filename>/usr/[local/]lib/sgml</Filename>.</para>
3500 <title>Configuring the DocBook tools</title>
3502 <Para>Once the DocBook tools are installed, the configure script
3503 will detect them and set up the build system accordingly. If you
3504 have a system that isn't supported, let us know, and we'll try
3509 <title>Remaining problems</title>
3511 <para>If you install from source, you'll get a pile of warnings
3514 <Screen>DTDDECL catalog entries are not supported</Screen>
3516 every time you build anything. These can safely be ignored, but
3517 if you find them tedious you can get rid of them by removing all
3518 the <Constant>DTDDECL</Constant> entries from
3519 <Filename>docbook.cat</Filename>.</para>
3523 <title>Building the documentation</title>
3525 <para>To build documentation in a certain format, you can
3526 say, for example,</para>
3532 <para>to build HTML documentation below the current directory.
3533 The available formats are: <literal>dvi</literal>,
3534 <literal>ps</literal>, <literal>pdf</literal>,
3535 <literal>html</literal>, and <literal>rtf</literal>. Note that
3536 not all documentation can be built in all of these formats: HTML
3537 documentation is generally supported everywhere, and DocBook
3538 documentation might support the other formats (depending on what
3539 other tools you have installed).</para>
3541 <para>All of these targets are recursive; that is, saying
3542 <literal>make html</literal> will make HTML docs for all the
3543 documents recursively below the current directory.</para>
3545 <para>Because there are many different formats that the DocBook
3546 documentation can be generated in, you have to select which ones
3547 you want by setting the <literal>SGMLDocWays</literal> variable
3548 to a list of them. For example, in
3549 <filename>build.mk</filename> you might have a line:</para>
3552 SGMLDocWays = html ps
3555 <para>This will cause the documentation to be built in the requested
3556 formats as part of the main build (the default is not to build
3557 any documentation at all).</para>
3561 <title>Installing the documentation</title>
3563 <para>To install the documentation, use:</para>
3569 <para>This will install the documentation into
3570 <literal>$(datadir)</literal> (which defaults to
3571 <literal>$(prefix)/share</literal>). The exception is HTML
3572 documentation, which goes into
3573 <literal>$(datadir)/html</literal>, to keep things tidy.</para>
3575 <para>Note that unless you set <literal>$(SGMLDocWays)</literal>
3576 to a list of formats, the <literal>install-docs</literal> target
3577 won't do anything for SGML documentation.</para>
3583 <sect1 id="sec-porting-ghc">
3584 <title>Porting GHC</title>
3586 <para>This section describes how to port GHC to a currenly
3587 unsupported platform. There are two distinct
3588 possibilities:</para>
3592 <para>The hardware architecture for your system is already
3593 supported by GHC, but you're running an OS that isn't
3594 supported (or perhaps has been supported in the past, but
3595 currently isn't). This is the easiest type of porting job,
3596 but it still requires some careful bootstrapping. Proceed to
3597 <xref linkend="sec-booting-from-hc">.</para>
3601 <para>Your system's hardware architecture isn't supported by
3602 GHC. This will be a more difficult port (though by comparison
3603 perhaps not as difficult as porting gcc). Proceed to <xref
3604 linkend="unregisterised-porting">.</para>
3608 <sect2 id="sec-booting-from-hc">
3609 <title>Booting/porting from C (<filename>.hc</filename>) files</title>
3611 <indexterm><primary>building GHC from .hc files</primary></indexterm>
3612 <indexterm><primary>booting GHC from .hc files</primary></indexterm>
3613 <indexterm><primary>porting GHC</primary></indexterm>
3615 <para>Bootstrapping GHC on a system without GHC already
3616 installed is achieved by taking the intermediate C files (known
3617 as HC files) from a GHC compilation on a supported system to the
3618 target machine, and compiling them using gcc to get a working
3621 <para><emphasis>NOTE: GHC versions 5.xx and later are
3622 significantly harder to bootstrap from C than earlier versions.
3623 We recommend starting from version 4.08.2 if you need to
3624 bootstrap in this way.</emphasis></para>
3626 <para>HC files are architecture-dependent (but not
3627 OS-dependent), so you have to get a set that were generated on
3628 similar hardware. There may be some supplied on the GHC
3629 download page, otherwise you'll have to compile some up
3630 yourself, or start from <emphasis>unregisterised</emphasis> HC
3631 files - see <xref linkend="unregisterised-porting">.</para>
3633 <para>The following steps should result in a working GHC build
3634 with full libraries:</para>
3638 <para>Unpack the HC files on top of a fresh source tree
3639 (make sure the source tree version matches the version of
3640 the HC files <emphasis>exactly</emphasis>!). This will
3641 place matching <filename>.hc</filename> files next to the
3642 corresponding Haskell source (<filename>.hs</filename> or
3643 <filename>.lhs</filename>) in the compiler subdirectory
3644 <filename>ghc/compiler</filename> and in the libraries
3645 (subdirectories of <filename>hslibs</filename> and
3646 <literal>libraries</literal>).</para>
3650 <para>The actual build process is fully automated by the
3651 <filename>hc-build</filename> script located in the
3652 <filename>distrib</filename> directory. If you eventually
3653 want to install GHC into the directory
3654 <replaceable>dir</replaceable>, the following
3655 command will execute the whole build process (it won't
3656 install yet):</para>
3659 foo% distrib/hc-build --prefix=<replaceable>dir</replaceable>
3661 <indexterm><primary>--hc-build</primary></indexterm>
3663 <para>By default, the installation directory is
3664 <filename>/usr/local</filename>. If that is what you want,
3665 you may omit the argument to <filename>hc-build</filename>.
3666 Generally, any option given to <filename>hc-build</filename>
3667 is passed through to the configuration script
3668 <filename>configure</filename>. If
3669 <filename>hc-build</filename> successfully completes the
3670 build process, you can install the resulting system, as
3680 <sect2 id="unregisterised-porting">
3681 <title>Porting GHC to a new architecture</title>
3683 <para>The first step in porting to a new architecture is to get
3684 an <firstterm>unregisterised</firstterm> build working. An
3685 unregisterised build is one that compiles via vanilla C only.
3686 By contrast, a registerised build uses the following
3687 architecture-specific hacks for speed:</para>
3691 <para>Global register variables: certain abstract machine
3692 <quote>registers</quote> are mapped to real machine
3693 registers, depending on how many machine registers are
3695 <filename>ghc/includes/MachRegs.h</filename>).</para>
3699 <para>Assembly-mangling: when compiling via C, we feed the
3700 assembly generated by gcc though a Perl script known as the
3701 <firstterm>mangler</firstterm> (see
3702 <filename>ghc/driver/mangler/ghc-asm.lprl</filename>). The
3703 mangler rearranges the assembly to support tail-calls and
3704 various other optimisations.</para>
3708 <para>In an unregisterised build, neither of these hacks are
3709 used — the idea is that the C code generated by the
3710 compiler should compile using gcc only. The lack of these
3711 optimisations costs about a factor of two in performance, but
3712 since unregisterised compilation is usually just a step on the
3713 way to a full registerised port, we don't mind too much.</para>
3716 <title>Building an unregisterised port</title>
3718 <para>The first step is to get some unregisterised HC files.
3719 Either (a) download them from the GHC site (if there are
3720 some available for the right version of GHC), or
3721 (b) build them yourself on any machine with a working
3722 GHC. If at all possible this should be a machine with the
3723 same word size as the target.</para>
3725 <para>There is a script available which should automate the
3726 process of doing the 2-stage bootstrap necessary to get the
3727 unregisterised HC files - it's available in <ulink
3728 url="http://cvs.haskell.org/cgi-bin/cvsweb.cgi/fptools/distrib/cross-port"><filename>fptools/distrib/cross-port</filename></ulink>
3731 <para>Now take these unregisterised HC files to the target
3732 platform and bootstrap a compiler from them as per the
3733 instructions in <xref linkend="sec-booting-from-hc">. In
3734 <filename>build.mk</filename>, you need to tell the build
3735 system that the compiler you're building is
3736 (a) unregisterised itself, and (b) builds
3737 unregisterised binaries. This varies depending on the GHC
3738 version you're bootstraping:</para>
3741 # build.mk for GHC 4.08.x
3742 GhcWithRegisterised=NO
3746 # build.mk for GHC 5.xx and 6.x
3747 GhcUnregisterised=YES
3750 <para>Versions 5.xx and 6.x only: use the option
3751 <option>--enable-hc-boot-unregisterised</option> instead of
3752 <option>--enable-hc-boot</option> when running
3753 <filename>./configure</filename>.</para>
3755 <para>The build may not go through cleanly. We've tried to
3756 stick to writing portable code in most parts of the compiler,
3757 so it should compile on any POSIXish system with gcc, but in
3758 our experience most systems differ from the standards in one
3759 way or another. Deal with any problems as they arise - if you
3760 get stuck, ask the experts on
3761 <email>glasgow-haskell-users@haskell.org</email>.</para>
3763 <para>Once you have the unregisterised compiler up and
3764 running, you can use it to start a registerised port. The
3765 following sections describe the various parts of the system
3766 that will need architecture-specific tweaks in order to get a
3767 registerised build going.</para>
3769 <para>Lots of useful information about the innards of GHC is
3770 available in the <ulink
3771 url="http://www.cse.unsw.edu.au/~chak/haskell/ghc/comm/">GHC
3772 Commentary</ulink>, which might be helpful if you run into
3773 some code which needs tweaking for your system.</para>
3777 <title>Porting the RTS</title>
3779 <para>The following files need architecture-specific code for a
3780 registerised build:</para>
3784 <term><filename>ghc/includes/MachRegs.h</filename></term>
3785 <indexterm><primary><filename>MachRegs.h</filename></primary>
3788 <para>Defines the STG-register to machine-register
3789 mapping. You need to know your platform's C calling
3790 convention, and which registers are generally available
3791 for mapping to global register variables. There are
3792 plenty of useful comments in this file.</para>
3796 <term><filename>ghc/includes/TailCalls.h</filename></term>
3797 <indexterm><primary><filename>TailCalls.h</filename></primary>
3800 <para>Macros that cooperate with the mangler (see <xref
3801 linkend="sec-mangler">) to make proper tail-calls
3806 <term><filename>ghc/rts/Adjustor.c</filename></term>
3807 <indexterm><primary><filename>Adjustor.c</filename></primary>
3811 <literal>foreign import "wrapper"</literal>
3813 <literal>foreign export dynamic</literal>).
3814 Not essential for getting GHC bootstrapped, so this file
3815 can be deferred until later if necessary.</para>
3819 <term><filename>ghc/rts/StgCRun.c</filename></term>
3820 <indexterm><primary><filename>StgCRun.c</filename></primary>
3823 <para>The little assembly layer between the C world and
3824 the Haskell world. See the comments and code for the
3825 other architectures in this file for pointers.</para>
3829 <term><filename>ghc/rts/MBlock.h</filename></term>
3830 <term><filename>ghc/rts/MBlock.c</filename></term>
3831 <indexterm><primary><filename>MBlock.h</filename></primary>
3833 <indexterm><primary><filename>MBlock.c</filename></primary>
3836 <para>These files are really OS-specific rather than
3837 architecture-specific. In <filename>MBlock.h</filename>
3838 is specified the absolute location at which the RTS
3839 should try to allocate memory on your platform (try to
3840 find an area which doesn't conflict with code or dynamic
3841 libraries). In <filename>Mblock.c</filename> you might
3842 need to tweak the call to <literal>mmap()</literal> for
3849 <sect3 id="sec-mangler">
3850 <title>The mangler</title>
3852 <para>The mangler is an evil Perl-script that rearranges the
3853 assembly code output from gcc to do two main things:</para>
3857 <para>Remove function prologues and epilogues, and all
3858 movement of the C stack pointer. This is to support
3859 tail-calls: every code block in Haskell code ends in an
3860 explicit jump, so we don't want the C-stack overflowing
3861 while we're jumping around between code blocks.</para>
3864 <para>Move the <firstterm>info table</firstterm> for a
3865 closure next to the entry code for that closure. In
3866 unregisterised code, info tables contain a pointer to the
3867 entry code, but in registerised compilation we arrange
3868 that the info table is shoved right up against the entry
3869 code, and addressed backwards from the entry code pointer
3870 (this saves a word in the info table and an extra
3871 indirection when jumping to the closure entry
3876 <para>The mangler is abstracted to a certain extent over some
3877 architecture-specific things such as the particular assembler
3878 directives used to herald symbols. Take a look at the
3879 definitions for other architectures and use these as a
3880 starting point.</para>
3884 <title>The native code generator</title>
3886 <para>The native code generator isn't essential to getting a
3887 registerised build going, but it's a desirable thing to have
3888 because it can cut compilation times in half. The native code
3889 generator is described in some detail in the <ulink
3890 url="http://www.cse.unsw.edu.au/~chak/haskell/ghc/comm/">GHC
3891 commentary</ulink>.</para>
3897 <para>To support GHCi, you need to port the dynamic linker
3898 (<filename>fptools/ghc/rts/Linker.c</filename>). The linker
3899 currently supports the ELF and PEi386 object file formats - if
3900 your platform uses one of these then you probably don't have
3901 to do anything except fiddle with the
3902 <literal>#ifdef</literal>s at the top of
3903 <filename>Linker.c</filename> to tell it about your OS.</para>
3905 <para>If your system uses a different object file format, then
3906 you have to write a linker — good luck!</para>
3912 <sect1 id="sec-build-pitfalls">
3913 <title>Known pitfalls in building Glasgow Haskell
3915 <indexterm><primary>problems, building</primary></indexterm>
3916 <indexterm><primary>pitfalls, in building</primary></indexterm>
3917 <indexterm><primary>building pitfalls</primary></indexterm></title>
3920 WARNINGS about pitfalls and known “problems”:
3929 One difficulty that comes up from time to time is running out of space
3930 in <literal>TMPDIR</literal>. (It is impossible for the configuration stuff to
3931 compensate for the vagaries of different sysadmin approaches to temp
3933 <indexterm><primary>tmp, running out of space in</primary></indexterm>
3935 The quickest way around it is <command>setenv TMPDIR /usr/tmp</command><indexterm><primary>TMPDIR</primary></indexterm> or
3936 even <command>setenv TMPDIR .</command> (or the equivalent incantation with your shell
3939 The best way around it is to say
3942 export TMPDIR=<dir>
3945 in your <filename>build.mk</filename> file.
3946 Then GHC and the other <literal>fptools</literal> programs will use the appropriate directory
3955 In compiling some support-code bits, e.g., in <filename>ghc/rts/gmp</filename> and even
3956 in <filename>ghc/lib</filename>, you may get a few C-compiler warnings. We think these
3964 When compiling via C, you'll sometimes get “warning: assignment from
3965 incompatible pointer type” out of GCC. Harmless.
3972 Similarly, <command>ar</command>chiving warning messages like the following are not
3976 ar: filename GlaIOMonad__1_2s.o truncated to GlaIOMonad_
3977 ar: filename GlaIOMonad__2_2s.o truncated to GlaIOMonad_
3987 In compiling the compiler proper (in <filename>compiler/</filename>), you <emphasis>may</emphasis>
3988 get an “Out of heap space” error message. These can vary with the
3989 vagaries of different systems, it seems. The solution is simple:
3996 If you're compiling with GHC 4.00 or later, then the
3997 <emphasis>maximum</emphasis> heap size must have been reached. This
3998 is somewhat unlikely, since the maximum is set to 64M by default.
3999 Anyway, you can raise it with the
4000 <option>-optCrts-M<size></option> flag (add this flag to
4001 <constant><module>_HC_OPTS</constant>
4002 <command>make</command> variable in the appropriate
4003 <filename>Makefile</filename>).
4010 For GHC < 4.00, add a suitable <option>-H</option> flag to the <filename>Makefile</filename>, as
4019 and try again: <command>gmake</command>. (see <Xref LinkEnd="sec-suffix"> for information about
4020 <constant><module>_HC_OPTS</constant>.)
4022 Alternatively, just cut to the chase:
4026 % make EXTRA_HC_OPTS=-optCrts-M128M
4035 If you try to compile some Haskell, and you get errors from GCC about
4036 lots of things from <filename>/usr/include/math.h</filename>, then your GCC was
4037 mis-installed. <command>fixincludes</command> wasn't run when it should've been.
4039 As <command>fixincludes</command> is now automagically run as part of GCC installation,
4040 this bug also suggests that you have an old GCC.
4048 You <emphasis>may</emphasis> need to re-<command>ranlib</command><indexterm><primary>ranlib</primary></indexterm> your libraries (on Sun4s).
4052 % cd $(libdir)/ghc-x.xx/sparc-sun-sunos4
4053 % foreach i ( `find . -name '*.a' -print` ) # or other-shell equiv...
4055 ? # or, on some machines: ar s $i
4060 We'd be interested to know if this is still necessary.
4068 GHC's sources go through <command>cpp</command> before being compiled, and <command>cpp</command> varies
4069 a bit from one Unix to another. One particular gotcha is macro calls
4074 SLIT("Hello, world")
4078 Some <command>cpp</command>s treat the comma inside the string as separating two macro
4079 arguments, so you get
4083 :731: macro `SLIT' used with too many (2) args
4087 Alas, <command>cpp</command> doesn't tell you the offending file!
4089 Workaround: don't put weird things in string args to <command>cpp</command> macros.
4100 <Sect1 id="winbuild"><Title>Notes for building under Windows</Title>
4103 This section summarises how to get the utilities you need on your
4104 Win95/98/NT/2000 machine to use CVS and build GHC. Similar notes for
4105 installing and running GHC may be found in the user guide. In general,
4106 Win95/Win98 behave the same, and WinNT/Win2k behave the same.
4107 You should read the GHC installation guide sections on Windows (in the user
4108 guide) before continuing to read these notes.
4112 <sect2 id="cygwin-and-mingw"><Title>Cygwin and MinGW</Title>
4114 <para> The Windows situation for building GHC is rather confusing. This section
4115 tries to clarify, and to establish terminology.</para>
4117 <sect3 id="ghc-mingw"><title>GHC-mingw</title>
4119 <para> <ulink url="http://www.mingw.org">MinGW (Minimalist GNU for Windows)</ulink>
4120 is a collection of header
4121 files and import libraries that allow one to use <command>gcc</command> and produce
4122 native Win32 programs that do not rely on any third-party DLLs. The
4123 current set of tools include GNU Compiler Collection (<command>gcc</command>), GNU Binary
4124 Utilities (Binutils), GNU debugger (Gdb), GNU make, and a assorted
4127 <para>The GHC that we distribute includes, inside the distribution itself, the MinGW <command>gcc</command>,
4128 <command>as</command>, <command>ld</command>, and a bunch of input/output libraries.
4129 GHC compiles Haskell to C (or to
4130 assembly code), and then invokes these MinGW tools to generate an executable binary.
4131 The resulting binaries can run on any Win32 system.
4133 <para> We will call a GHC that targets MinGW in this way <emphasis>GHC-mingw</emphasis>.</para>
4135 <para> The down-side of GHC-mingw is that the MinGW libraries do not support anything like the full
4136 Posix interface. So programs compiled with GHC-mingw cannot import the (Haskell) Posix
4137 library; they have to do
4138 their input output using standard Haskell I/O libraries, or native Win32 bindings.
4142 <sect3 id="ghc-cygwin"><title>GHC-cygwin</title>
4144 <para>There <emphasis>is</emphasis> a way to get the full Posix interface, which is to use Cygwin.
4145 <ulink url="http://www.cygwin.com">Cygwin</ulink> is a complete Unix simulation that runs on Win32.
4146 Cygwin comes with a shell, and all the usual Unix commands: <command>mv</command>, <command>rm</command>,
4147 <command>ls</command>, plus of course <command>gcc</command>, <command>ld</command> and so on.
4148 A C program compiled with the Cygwin <command>gcc</command> certainly can use all of Posix.
4150 <para>So why doesn't GHC use the Cygwin <command>gcc</command> and libraries? Because
4151 Cygwin comes with a DLL <emphasis>that must be linked with every runnable Cygwin-compiled program</emphasis>.
4152 A program compiled by the Cygwin tools cannot run at all unless Cygwin is installed.
4153 If GHC targeted Cygwin, users would have to install Cygwin just to run the Haskell programs
4154 that GHC compiled; and the Cygwin DLL would have to be in the DLL load path.
4155 Worse, Cygwin is a moving target. The name of the main DLL, <literal>cygwin1.dll</literal>
4156 does not change, but the implementation certainly does. Even the interfaces to functions
4157 it exports seem to change occasionally. So programs compiled by GHC might only run with
4158 particular versions of Cygwin. All of this seems very undesirable.
4161 Nevertheless, it is certainly possible to build a version of GHC that targets Cygwin;
4162 we will call that <emphasis>GHC-cygwin</emphasis>. The up-side of GHC-cygwin is
4163 that Haskell programs compiled by GHC-cygwin can import the (Haskell) Posix library.
4167 <sect3><title>HOST_OS vs TARGET_OS</title>
4170 In the source code you'll find various ifdefs looking like:
4172 #ifdef mingw32_HOST_OS
4178 #ifdef mingw32_TARGET_OS
4182 These macros are set by the configure script (via the file config.h).
4183 Which is which? The criterion is this. In the ifdefs in GHC's source code:
4186 The "host" system is the one on which GHC itself will be run.
4189 The "target" system is the one for which the program compiled by GHC will be run.
4192 For a stage-2 compiler, in which GHCi is available, the "host" and "target" systems must be the same.
4193 So then it doesn't really matter whether you use the HOST_OS or TARGET_OS cpp macros.
4198 <sect3><title>Summary</title>
4200 <para>Notice that "GHC-mingw" means "GHC that <emphasis>targets</emphasis> MinGW". It says nothing about
4201 how that GHC was <emphasis>built</emphasis>. It is entirely possible to have a GHC-mingw that was built
4202 by compiling GHC's Haskell sources with a GHC-cygwin, or vice versa.</para>
4204 <para>We distribute only a GHC-mingw built by a GHC-mingw; supporting
4205 GHC-cygwin too is beyond our resources. The GHC we distribute
4206 therefore does not require Cygwin to run, nor do the programs it
4207 compiles require Cygwin.</para>
4209 <para>The instructions that follow describe how to build GHC-mingw. It is
4210 possible to build GHC-cygwin, but it's not a supported route, and the build system might
4213 <para>In your build tree, you build a compiler called <Command>ghc-inplace</Command>. It
4214 uses the <Command>gcc</Command> that you specify using the
4215 <option>--with-gcc</option> flag when you run
4216 <Command>configure</Command> (see below).
4217 The makefiles are careful to use <Command>ghc-inplace</Command> (not <Command>gcc</Command>)
4218 to compile any C files, so that it will in turn invoke the right <Command>gcc</Command> rather that
4219 whatever one happens to be in your path. However, the makefiles do use whatever <Command>ld</Command>
4220 and <Command>ar</Command> happen to be in your path. This is a bit naughty, but (a) they are only
4221 used to glom together .o files into a bigger .o file, or a .a file,
4222 so they don't ever get libraries (which would be bogus; they might be the wrong libraries), and (b)
4223 Cygwin and Mingw use the same .o file format. So its ok.
4228 <Sect2><Title>Installing and configuring Cygwin</Title>
4230 <para>You don't need Cygwin to <emphasis>use</emphasis> GHC,
4231 but you do need it to <emphasis>build</emphasis> GHC.</para>
4233 <para> Install Cygwin from <ulink url="http://www.cygwin.com/">http://www.cygwin.com/</ulink>.
4234 The installation process is straightforward; we install it in <Filename>c:/cygwin</Filename>.
4235 During the installation dialogue, make sure that you select:
4236 <command>cvs</command>, <command>openssh</command>,
4237 <command>autoconf</command>,
4238 <command>binutils</command> (includes ld and (I think) ar),
4239 <command>gcc</command>,
4240 <command>flex</command>,
4241 <command>make</command>.
4244 <para> Now set the following user environment variables:
4247 <listitem><para> Add <filename>c:/cygwin/bin</filename> and <filename>c:/cygwin/usr/bin</filename> to your
4248 <constant>PATH</constant></para></listitem>
4252 Set <constant>MAKE_MODE</constant> to <Literal>UNIX</Literal>. If you
4253 don't do this you get very weird messages when you type
4254 <Command>make</Command>, such as:
4256 /c: /c: No such file or directory
4261 <listitem><para> Set <constant>SHELL</constant> to
4262 <Filename>c:/cygwin/bin/sh</Filename>. When you invoke a shell in Emacs, this
4263 <constant>SHELL</constant> is what you get.
4266 <listitem><para> Set <constant>HOME</constant> to point to your
4267 home directory. This is where, for example,
4268 <command>bash</command> will look for your <filename>.bashrc</filename>
4269 file. Ditto <command>emacs</command> looking for <filename>.emacsrc</filename>
4275 There are a few other things to do:
4279 By default, cygwin provides the command shell <filename>ash</filename>
4280 as <filename>sh.exe</filename>. We have often seen build-system problems that
4281 turn out to be due to bugs in <filename>ash</filename>
4283 and length of command lines). On the other hand <filename>bash</filename> seems
4285 So, in <filename>cygwin/bin</filename>
4286 remove the supplied <filename>sh.exe</filename> (or rename it as <filename>ash.exe</filename>),
4287 and copy <filename>bash.exe</filename> to <filename>sh.exe</filename>.
4288 You'll need to do this in Windows Explorer or the Windows <command>cmd</command> shell, because
4289 you can't rename a running program!
4295 Some script files used in the make system start with "<Command>#!/bin/perl</Command>",
4296 (and similarly for <Command>sh</Command>). Notice the hardwired path!
4297 So you need to ensure that your <Filename>/bin</Filename> directory has the following
4300 <listitem> <para><Command>sh</Command></para></listitem>
4301 <listitem> <para><Command>perl</Command></para></listitem>
4302 <listitem> <para><Command>cat</Command></para></listitem>
4304 All these come in Cygwin's <Filename>bin</Filename> directory, which you probably have
4305 installed as <Filename>c:/cygwin/bin</Filename>. By default Cygwin mounts "<Filename>/</Filename>" as
4306 <Filename>c:/cygwin</Filename>, so if you just take the defaults it'll all work ok.
4307 (You can discover where your Cygwin
4308 root directory <Filename>/</Filename> is by typing <Command>mount</Command>.)
4309 Provided <Filename>/bin</Filename> points to the Cygwin <Filename>bin</Filename>
4310 directory, there's no need to copy anything. If not, copy these binaries from the <filename>cygwin/bin</filename>
4311 directory (after fixing the <filename>sh.exe</filename> stuff mentioned in the previous bullet).
4317 <para>Finally, here are some things to be aware of when using Cygwin:
4319 <listitem> <para>Cygwin doesn't deal well with filenames that include
4320 spaces. "<filename>Program Files</filename>" and "<filename>Local files</filename>" are
4324 <listitem> <para> Cygwin implements a symbolic link as a text file with some
4325 magical text in it. So other programs that don't use Cygwin's
4326 I/O libraries won't recognise such files as symlinks.
4327 In particular, programs compiled by GHC are meant to be runnable
4328 without having Cygwin, so they don't use the Cygwin library, so
4329 they don't recognise symlinks.
4333 Win32 has a <command>find</command> command which is not the same as Cygwin's find.
4334 You will probably discover that the Win32 <command>find</command> appears in your <constant>PATH</constant>
4335 before the Cygwin one, because it's in the <emphasis>system</emphasis> <constant>PATH</constant>
4336 environment variable, whereas you have probably modified the <emphasis>user</emphasis> <constant>PATH</constant>
4337 variable. You can always invoke <command>find</command> with an absolute path, or rename it.
4344 <Sect2><Title>Other things you need to install</Title>
4346 <para>You have to install the following other things to build GHC:
4350 Install an executable GHC, from <ulink url="http://www.haskell.org/ghc">http://www.haskell.org/ghc</ulink>.
4351 This is what you will use to compile GHC. Add it in your
4352 <constant>PATH</constant>: the installer tells you the path element
4353 you need to add upon completion.
4359 Install an executable Happy, from <ulink url="http://www.haskell.org/happy">http://www.haskell.org/happy</ulink>.
4360 Happy is a parser generator used to compile the Haskell grammar. Add it in your
4361 <constant>PATH</constant>.
4367 <para>GHC uses the <emphasis>mingw</emphasis> C compiler to
4368 generate code, so you have to install that (see <xref linkend="cygwin-and-mingw">).
4369 Just pick up a mingw bundle at
4370 <ulink url="http://www.mingw.org/">http://www.mingw.org/</ulink>.
4371 We install it in <filename>c:/mingw</filename>.
4373 <para>Do <emphasis>not</emphasis> add any of the <emphasis>mingw</emphasis> binaries to your path.
4374 They are only going to get used by explicit access (via the --with-gcc flag you
4375 give to <Command>configure</Command> later). If you do add them to your path
4376 you are likely to get into a mess because their names overlap with Cygwin binaries.
4382 <para>We use <command>emacs</command> a lot, so we install that too.
4383 When you are in <filename>fptools/ghc/compiler</filename>, you can use
4384 "<literal>make tags</literal>" to make a TAGS file for emacs. That uses the utility
4385 <filename>fptools/ghc/utils/hasktags/hasktags</filename>, so you need to make that first.
4386 The most convenient way to do this is by going <literal>make boot</literal> in <filename>fptools/ghc</filename>.
4387 The <literal>make tags</literal> command also uses <command>etags</command>, which comes with <command>emacs</command>,
4388 so you will need to add <filename>emacs/bin</filename> to your <literal>PATH</literal>.
4394 <para> Finally, check out a copy of GHC sources from
4395 the CVS repository, following the instructions above (<xref linkend="cvs-access">).
4402 <Sect2><Title>Building GHC</Title>
4405 Now go read the documentation above on building from source (<xref linkend="sec-building-from-source">);
4406 the bullets below only tell
4407 you about Windows-specific wrinkles.</para>
4411 Run <Command>autoconf</Command> both in <filename>fptools</filename>
4412 and in <filename>fptools/ghc</filename>. If you omit the latter step you'll
4413 get an error when you run <filename>./configure</filename>:
4416 creating mk/config.h
4417 mk/config.h is unchanged
4419 running /bin/sh ./configure --cache-file=.././config.cache --srcdir=.
4420 ./configure: ./configure: No such file or directory
4421 configure: error: ./configure failed for ghc
4426 <listitem> <para><command>autoconf</command> seems to create the file <filename>configure</filename>
4427 read-only. So if you need to run autoconf again (which I sometimes do for safety's sake),
4430 /usr/bin/autoconf: cannot create configure: permission denied
4432 Solution: delete <filename>configure</filename> first.
4437 You either need to add <filename>ghc</filename> to your
4438 <constant>PATH</constant> before you invoke
4439 <Command>configure</Command>, or use the <Command>configure</Command>
4440 option <option>--with-ghc=c:/ghc/ghc-some-version/bin/ghc</option>.
4445 If you are paranoid, delete <filename>config.cache</filename> if it exists.
4446 This file occasionally remembers out-of-date configuration information, which
4447 can be really confusing.
4453 After <command>autoconf</command> run <command>./configure</command> in
4454 <filename>fptools/</filename> thus:
4457 ./configure --host=i386-unknown-mingw32 --with-gcc=c:/mingw/bin/gcc
4459 This is the point at which you specify that you are building GHC-mingw
4460 (see <xref linkend="ghc-mingw">). </para>
4462 <para> Both these options are important! It's possible to get into
4463 trouble using the wrong C compiler!</para>
4465 Furthermore, it's <emphasis>very important</emphasis> that you specify a
4466 full MinGW path for <command>gcc</command>, not a Cygwin path, because GHC (which
4467 uses this path to invoke <command>gcc</command>) is a MinGW program and won't
4468 understand a Cygwin path. For example, if you
4469 say <literal>--with-gcc=/mingw/bin/gcc</literal>, it'll be interpreted as
4470 <filename>/cygdrive/c/mingw/bin/gcc</filename>, and GHC will fail the first
4471 time it tries to invoke it. Worse, the failure comes with
4472 no error message whatsoever. GHC simply fails silently when first invoked,
4473 typically leaving you with this:
4475 make[4]: Leaving directory `/cygdrive/e/fptools-stage1/ghc/rts/gmp'
4476 ../../ghc/compiler/ghc-inplace -optc-mno-cygwin -optc-O
4477 -optc-Wall -optc-W -optc-Wstrict-prototypes -optc-Wmissing-prototypes
4478 -optc-Wmissing-declarations -optc-Winline -optc-Waggregate-return
4479 -optc-Wbad-function-cast -optc-Wcast-align -optc-I../includes
4480 -optc-I. -optc-Iparallel -optc-DCOMPILING_RTS
4481 -optc-fomit-frame-pointer -O2 -static
4482 -package-name rts -O -dcore-lint -c Adjustor.c -o Adjustor.o
4483 make[2]: *** [Adjustor.o] Error 1
4484 make[1]: *** [all] Error 1
4485 make[1]: Leaving directory `/cygdrive/e/fptools-stage1/ghc'
4486 make: *** [all] Error 1
4492 If you want to build GHC-cygwin (<xref linkend="ghc-cygwin">)
4493 you'll have to do something more like:
4495 ./configure --with-gcc=...the Cygwin gcc...
4500 <listitem><para> You almost certainly want to set
4504 in your <filename>build.mk</filename> configuration file (see <xref linkend="sec-build-config">).
4505 This tells the build system not to split each library into a myriad of little object files, one
4506 for each function. Doing so reduces binary sizes for statically-linked binaries, but on Windows
4507 it dramatically increases the time taken to build the libraries in the first place.
4511 <listitem><para> Do not attempt to build the documentation.
4512 It needs all kinds of wierd Jade stuff that we haven't worked out for
4513 Win32.</para></listitem>