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
264 <para>Send a message to to the CVS repository
265 administrator (currently Jeff Lewis
266 <email>jeff@galconn.com</email>), containing:</para>
269 <para>Your desired user-name.</para>
272 <para>Your <literal>.ssh/id_dsa.pub</literal> (or
273 <literal>.ssh/identity.pub</literal>).</para>
276 <para>He will set up your account.</para>
280 <para>Set the following environment variables:</para>
284 <constant>$HOME</constant>: points to your home directory. This is where CVS
285 will look for its <filename>.cvsrc</filename> file.
291 <constant>$CVS_RSH</constant> to <filename>ssh</filename>
293 <para>[Windows users.] Setting your <literal>CVS_RSH</literal> to
294 <literal>ssh</literal> assumes that your CVS client
295 understands how to execute shell script
296 ("#!"s,really), which is what
297 <literal>ssh</literal> is. This may not be the case on
298 Win32 platforms, so in that case set <literal>CVS_RSH</literal> to
299 <literal>ssh1</literal>.</para>
303 <para><literal>$CVSROOT</literal> to
304 <literal>:ext:</literal><replaceable>your-username</replaceable>
305 <literal>@cvs.haskell.org:/home/cvs/root</literal>
306 where <replaceable>your-username</replaceable> is your user name on
307 <literal>cvs.haskell.org</literal>.
309 <para>The <literal>CVSROOT</literal> environment variable will
310 be recorded in the checked-out tree, so you don't need to set
311 this every time. </para>
317 <constant>$CVSEDITOR</constant>: <filename>bin/gnuclient.exe</filename>
318 if you want to use an Emacs buffer for typing in those long commit messages.
324 <constant>$SHELL</constant>: To use bash as the shell in Emacs, you need to
325 set this to point to <filename>bash.exe</filename>.
336 Put the following in <filename>$HOME/.cvsrc</filename>:
347 These are the default options for the specified CVS commands,
348 and represent better defaults than the usual ones. (Feel
349 free to change them.)
353 [Windows users.] Filenames starting with <filename>.</filename> were illegal in
354 the 8.3 DOS filesystem, but that restriction should have
355 been lifted by now (i.e., you're using VFAT or later filesystems.) If
356 you're still having problems creating it, don't worry; <filename>.cvsrc</filename> is entirely
364 <para>[Experts.] Once your account is set up, you can get
365 access from other machines without bothering Jeff, thus:</para>
368 <para>Generate a public/private key pair on the new
372 <para>Use ssh to log in to
373 <literal>cvs.haskell.org</literal>, from your old
377 <para>Add the public key for the new machine to the file
378 <literal>$HOME/ssh/authorized_keys</literal> on
379 <literal>cvs.haskell.org</literal>.
380 (<literal>authorized_keys2</literal>, I think, for Version
384 <para>Make sure that the new version of
385 <literal>authorized_keys</literal> still has 600 file
394 <sect2 id="cvs-first">
395 <title>Checking Out a Source Tree</title>
399 <para>Make sure you set your <literal>CVSROOT</literal>
400 environment variable according to either of the remote
401 methods above. The Approved Way to check out a source tree
402 is as follows:</para>
405 $ cvs checkout fpconfig
408 <para>At this point you have a new directory called
409 <literal>fptools</literal> which contains the basic stuff
410 for the fptools suite, including the configuration files and
411 some other junk. </para>
413 <para>[Windows users.] The following messages appear to be harmless:
415 setsockopt IPTOS_LOWDELAY: Invalid argument
416 setsockopt IPTOS_THROUGHPUT: Invalid argument
421 <para>You can call the fptools directory whatever you like,
422 CVS won't mind: </para>
425 $ mv fptools <replaceable>directory</replaceable>
428 <para> NB: after you've read the CVS manual you might be
429 tempted to try</para>
431 $ cvs checkout -d <replaceable>directory</replaceable> fpconfig
434 <para>instead of checking out <literal>fpconfig</literal>
435 and then renaming it. But this doesn't work, and will
436 result in checking out the entire repository instead of just
437 the <literal>fpconfig</literal> bit.</para>
439 $ cd <replaceable>directory</replaceable>
440 $ cvs checkout ghc hslibs
443 <para>The second command here checks out the relevant
444 modules you want to work on. For a GHC build, for instance,
445 you need at least the <literal>ghc</literal> and
446 <literal>hslibs</literal> modules (for a full list of the
447 projects available, see <xref linkend="projects">).</para>
452 <sect2 id="cvs-committing">
453 <title>Committing Changes</title>
455 <para>This is only if you have read-write access to the
456 repository. For anoncvs users, CVS will issue a "read-only
457 repository" error if you try to commit changes.</para>
461 <para>Build the software, if necessary. Unless you're just
462 working on documentation, you'll probably want to build the
463 software in order to test any changes you make.</para>
467 <para>Make changes. Preferably small ones first.</para>
471 <para>Test them. You can see exactly what changes you've
472 made by using the <literal>cvs diff</literal> command:</para>
476 <para>lists all the changes (using the
477 <literal>diff</literal> command) in and below the current
478 directory. In emacs, <literal>C-c C-v =</literal> runs
479 <literal>cvs diff</literal> on the current buffer and shows
480 you the results.</para>
484 <para>Before checking in a change, you need to update your
491 <para>This pulls in any changes that other people have made,
492 and merges them with yours. If there are any conflicts, CVS
493 will tell you, and you'll have to resolve them before you
494 can check your changes in. The documentation describes what
495 to do in the event of a conflict.</para>
497 <para>It's not always necessary to do a full cvs update
498 before checking in a change, since CVS will always tell you
499 if you try to check in a file that someone else has changed.
500 However, you should still update at regular intervals to
501 avoid making changes that don't work in conjuction with
502 changes that someone else made. Keeping an eye on what goes
503 by on the mailing list can help here.</para>
507 <para>When you're happy that your change isn't going to
508 break anything, check it in. For a one-file change:</para>
511 $ cvs commit <replaceable>filename</replaceable>
514 <para>CVS will then pop up an editor for you to enter a
515 "commit message", this is just a short description
516 of what your change does, and will be kept in the history of
519 <para>If you're using emacs, simply load up the file into a
520 buffer and type <literal>C-x C-q</literal>, and emacs will
521 prompt for a commit message and then check in the file for
524 <para>For a multiple-file change, things are a bit
525 trickier. There are several ways to do this, but this is the
526 way I find easiest. First type the commit message into a
527 temporary file. Then either</para>
530 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>file_1</replaceable> .... <replaceable>file_n</replaceable>
533 <para>or, if nothing else has changed in this part of the
537 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>directory</replaceable>
540 <para>where <replaceable>directory</replaceable> is a common
541 parent directory for all your changes, and
542 <replaceable>commit-message</replaceable> is the name of the
543 file containing the commit message.</para>
545 <para>Shortly afterwards, you'll get some mail from the
546 relevant mailing list saying which files changed, and giving
547 the commit message. For a multiple-file change, you should
548 still get only <emphasis>one</emphasis> message.</para>
553 <sect2 id="cvs-update">
554 <title>Updating Your Source Tree</title>
556 <para>It can be tempting to cvs update just part of a source
557 tree to bring in some changes that someone else has made, or
558 before committing your own changes. This is NOT RECOMMENDED!
559 Quite often changes in one part of the tree are dependent on
560 changes in another part of the tree (the
561 <literal>mk/*.mk</literal> files are a good example where
562 problems crop up quite often). Having an inconsistent tree is a
563 major cause of headaches. </para>
565 <para>So, to avoid a lot of hassle, follow this recipe for
566 updating your tree: </para>
570 $ cvs update -Pd 2>&1 | tee log</screen>
572 <para>Look at the log file, and fix any conflicts (denoted by a
573 <quote>C</quote> in the first column). If you're using multiple
574 build trees, then for every build tree you have pointing at this
575 source tree, you need to update the links in case any new files
576 have appeared: </para>
579 $ cd <replaceable>build-tree</replaceable>
580 $ lndir <replaceable>source-tree</replaceable>
583 <para>Some files might have been removed, so you need to remove
584 the links pointing to these non-existent files:</para>
587 $ find . -xtype l -exec rm '{}' \;
590 <para>To be <emphasis>really</emphasis> safe, you should do
593 <screen>$ gmake all</screen>
595 <para>from the top-level, to update the dependencies and build
596 any changed files. </para>
599 <sect2 id="cvs-tags">
600 <title>GHC Tag Policy</title>
602 <para>If you want to check out a particular version of GHC,
603 you'll need to know how we tag versions in the repository. The
604 policy (as of 4.04) is:</para>
608 <para>The tree is branched before every major release. The
609 branch tag is <literal>ghc-x-xx-branch</literal>, where
610 <literal>x-xx</literal> is the version number of the release
611 with the <literal>'.'</literal> replaced by a
612 <literal>'-'</literal>. For example, the 4.04 release lives
613 on <literal>ghc-4-04-branch</literal>.</para>
617 <para>The release itself is tagged with
618 <literal>ghc-x-xx</literal> (on the branch). eg. 4.06 is
619 called <literal>ghc-4-06</literal>.</para>
623 <para>We didn't always follow these guidelines, so to see
624 what tags there are for previous versions, do <literal>cvs
625 log</literal> on a file that's been around for a while (like
626 <literal>fptools/ghc/README</literal>).</para>
630 <para>So, to check out a fresh GHC 4.06 tree you would
634 $ cvs co -r ghc-4-06 fpconfig
636 $ cvs co -r ghc-4-06 ghc hslibs
640 <sect2 id="cvs-hints">
641 <title>General Hints</title>
645 <para>As a general rule: commit changes in small units,
646 preferably addressing one issue or implementing a single
647 feature. Provide a descriptive log message so that the
648 repository records exactly which changes were required to
649 implement a given feature/fix a bug. I've found this
650 <emphasis>very</emphasis> useful in the past for finding out
651 when a particular bug was introduced: you can just wind back
652 the CVS tree until the bug disappears.</para>
656 <para>Keep the sources at least *buildable* at any given
657 time. No doubt bugs will creep in, but it's quite easy to
658 ensure that any change made at least leaves the tree in a
659 buildable state. We do nightly builds of GHC to keep an eye
660 on what things work/don't work each day and how we're doing
661 in relation to previous verions. This idea is truely wrecked
662 if the compiler won't build in the first place!</para>
666 <para>To check out extra bits into an already-checked-out
667 tree, use the following procedure. Suppose you have a
668 checked-out fptools tree containing just ghc, and you want
669 to add nofib to it:</para>
680 $ cvs update -d nofib
683 <para>(the -d flag tells update to create a new
684 directory). If you just want part of the nofib suite, you
689 $ cvs checkout nofib/spectral
692 <para>This works because <literal>nofib</literal> is a
693 module in its own right, and spectral is a subdirectory of
694 the nofib module. The path argument to checkout must always
695 start with a module name. There's no equivalent form of this
696 command using <literal>update</literal>.</para>
702 <sect1 id="projects">
703 <title>What projects are there?</title>
705 <para>The <literal>fptools</literal> suite consists of several
706 <firstterm>projects</firstterm>, most of which can be downloaded,
707 built and installed individually. Each project corresponds to a
708 subdirectory in the source tree, and if checking out from CVS then
709 each project can be checked out individually by sitting in the top
710 level of your source tree and typing <command>cvs checkout
711 <replaceable>project</replaceable></command>.</para>
713 <para>Here is a list of the projects currently available:</para>
717 <term><literal>ghc</literal></term>
718 <indexterm><primary><literal>ghc</literal></primary>
719 <secondary>project</secondary></indexterm>
721 <para>The <ulink url="http://www.haskell.org/ghc/">Glasgow
722 Haskell Compiler</ulink> (minus libraries). Absolutely
723 required for building GHC.</para>
728 <term><literal>glafp-utils</literal></term>
729 <indexterm><primary><literal>glafp-utils</literal></primary><secondary>project</secondary></indexterm>
731 <para>Utility programs, some of which are used by the
732 build/installation system. Required for pretty much
738 <term><literal>green-card</literal></term>
739 <indexterm><primary><literal>green-card</literal></primary><secondary>project</secondary></indexterm>
742 url="http://www.haskell.org/greencard/">Green Card</ulink>
743 system for generating Haskell foreign function
749 <term><literal>haggis</literal></term>
750 <indexterm><primary><literal>haggis</literal></primary><secondary>project</secondary></indexterm>
753 url="http://www.dcs.gla.ac.uk/fp/software/haggis/">Haggis</ulink>
754 Haskell GUI framework.</para>
759 <term><literal>haddock</literal></term>
760 <indexterm><primary><literal>haddock</literal></primary><secondary>project</secondary></indexterm>
763 url="http://www.haskell.org/haddock/">Haddock</ulink>
764 documentation tool.</para>
769 <term><literal>happy</literal></term>
770 <indexterm><primary><literal>happy</literal></primary><secondary>project</secondary></indexterm>
773 url="http://www.haskell.org/happy/">Happy</ulink> Parser
779 <term><literal>hdirect</literal></term>
780 <indexterm><primary><literal>hdirect</literal></primary><secondary>project</secondary></indexterm>
783 url="http://www.haskell.org/hdirect/">H/Direct</ulink>
784 Haskell interoperability tool.</para>
789 <term><literal>hood</literal></term>
790 <indexterm><primary><literal>hood</literal></primary><secondary>project</secondary></indexterm>
792 <para>The <ulink url="http://www.haskell.org/hood/">Haskell
793 Object Observation Debugger</ulink>.</para>
798 <term><literal>hslibs</literal></term>
799 <indexterm><primary><literal>hslibs</literal></primary><secondary>project</secondary></indexterm>
801 <para>Supplemental libraries for GHC
802 (<emphasis>required</emphasis> for building GHC).</para>
807 <term><literal>libraries</literal></term>
808 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
810 <para>Hierarchical Haskell library suite
811 (<emphasis>required</emphasis> for building GHC).</para>
816 <term><literal>mhms</literal></term>
817 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
819 <para>The Modular Haskell Metric System.</para>
824 <term><literal>nofib</literal></term>
825 <indexterm><primary><literal>nofib</literal></primary><secondary>project</secondary></indexterm>
827 <para>The NoFib suite: A collection of Haskell programs used
828 primarily for benchmarking.</para>
833 <term><literal>testsuite</literal></term>
834 <indexterm><primary><literal>testsuite</literal></primary><secondary>project</secondary></indexterm>
836 <para>A testing framework, including GHC's regression test
842 <para>So, to build GHC you need at least the
843 <literal>ghc</literal>, <literal>libraries</literal> and
844 <literal>hslibs</literal> projects (a GHC source distribution will
845 already include the bits you need).</para>
848 <sect1 id="sec-build-checks">
849 <title>Things to check before you start</title>
851 <para>Here's a list of things to check before you get
857 <indexterm><primary>Disk space needed</primary></indexterm>
858 <para>Disk space needed: from about 100Mb for a basic GHC
859 build, up to probably 500Mb for a GHC build with everything
860 included (libraries built several different ways,
865 <para>Use an appropriate machine / operating system. <xref
866 linkend="sec-port-info"> lists the supported platforms; if
867 yours isn't amongst these then you can try porting GHC (see
868 <xref linkend="sec-porting-ghc">).</para>
872 <para>Be sure that the “pre-supposed” utilities are
873 installed. <Xref LinkEnd="sec-pre-supposed">
878 <para>If you have any problem when building or installing the
879 Glasgow tools, please check the “known pitfalls” (<Xref
880 LinkEnd="sec-build-pitfalls">). Also check the FAQ for the
881 version you're building, which is part of the User's Guide and
882 available on the <ulink URL="http://www.haskell.org/ghc/" >GHC web
885 <indexterm><primary>bugs</primary><secondary>known</secondary></indexterm>
887 <para>If you feel there is still some shortcoming in our
888 procedure or instructions, please report it.</para>
890 <para>For GHC, please see the <ulink
891 url="http://www.haskell.org/ghc/docs/latest/set/bug-reporting.html">bug-reporting
892 section of the GHC Users' Guide</ulink>, to maximise the
893 usefulness of your report.</para>
895 <indexterm><primary>bugs</primary><secondary>seporting</secondary></indexterm>
896 <para>If in doubt, please send a message to
897 <email>glasgow-haskell-bugs@haskell.org</email>.
898 <indexterm><primary>bugs</primary><secondary>mailing
899 list</secondary></indexterm></para>
904 <sect1 id="sec-port-info">
905 <title>What machines the Glasgow tools run on</title>
907 <indexterm><primary>ports</primary><secondary>GHC</secondary></indexterm>
908 <indexterm><primary>GHC</primary><secondary>ports</secondary></indexterm>
909 <indexterm><primary>platforms</primary><secondary>supported</secondary></indexterm>
911 <para>The main question is whether or not the Haskell compiler
912 (GHC) runs on your platform.</para>
914 <para>A “platform” is a
915 architecture/manufacturer/operating-system combination, such as
916 <literal>sparc-sun-solaris2</literal>. Other common ones are
917 <literal>alpha-dec-osf2</literal>,
918 <literal>hppa1.1-hp-hpux9</literal>,
919 <literal>i386-unknown-linux</literal>,
920 <literal>i386-unknown-solaris2</literal>,
921 <literal>i386-unknown-freebsd</literal>,
922 <literal>i386-unknown-cygwin32</literal>,
923 <literal>m68k-sun-sunos4</literal>,
924 <literal>mips-sgi-irix5</literal>,
925 <literal>sparc-sun-sunos4</literal>,
926 <literal>sparc-sun-solaris2</literal>,
927 <literal>powerpc-ibm-aix</literal>.</para>
929 <para>Some libraries may only work on a limited number of
930 platforms; for example, a sockets library is of no use unless the
931 operating system supports the underlying BSDisms.</para>
934 <title>What platforms the Haskell compiler (GHC) runs on</title>
936 <indexterm><primary>fully-supported platforms</primary></indexterm>
937 <indexterm><primary>native-code generator</primary></indexterm>
938 <indexterm><primary>registerised ports</primary></indexterm>
939 <indexterm><primary>unregisterised ports</primary></indexterm>
941 <para>The GHC hierarchy of Porting Goodness: (a) Best is a
942 native-code generator; (b) next best is a
943 “registerised” port; (c) the bare minimum is an
944 “unregisterised” port.
945 (“Unregisterised” is so terrible that we won't say
946 more about it).</para>
948 <para>We use Sparcs running Solaris 2.7 and x86 boxes running
949 FreeBSD and Linux, so those are the best supported platforms,
950 unsurprisingly.</para>
952 <para>Here's everything that's known about GHC ports. We
953 identify platforms by their “canonical”
954 CPU/Manufacturer/OS triple.</para>
958 <term>alpha-dec-{osf,linux,freebsd,openbsd,netbsd}:</term>
959 <indexterm><primary>alpha-dec-osf</primary></indexterm>
960 <indexterm><primary>alpha-dec-linux</primary></indexterm>
961 <indexterm><primary>alpha-dec-freebsd</primary></indexterm>
962 <indexterm><primary>alpha-dec-openbsd</primary></indexterm>
963 <indexterm><primary>alpha-dec-netbsd</primary></indexterm>
966 <para>The OSF port is currently working (as of GHC version
967 5.02.1) and well supported. The native code generator is
968 currently non-working. Other operating systems will
969 require some minor porting.</para>
974 <term>sparc-sun-sunos4</term>
975 <indexterm><primary>sparc-sun-sunos4</primary></indexterm>
977 <para>Probably works with minor tweaks, hasn't been tested
983 <term>sparc-sun-solaris2</term>
984 <indexterm><primary>sparc-sun-solaris2</primary></indexterm>
986 <para>Fully supported (at least for Solaris 2.7),
987 including native-code generator.</para>
992 <term>hppa1.1-hp-hpux (HP-PA boxes running HPUX 9.x)</term>
993 <indexterm><primary>hppa1.1-hp-hpux</primary></indexterm>
995 <para>A registerised port is available for version 4.08,
996 but GHC hasn't been built on that platform since (as far
997 as we know). No native-code generator.</para>
1002 <term>i386-unknown-linux (PCs running Linux, ELF binary format)</term>
1003 <indexterm><primary>i386-*-linux</primary></indexterm>
1005 <para>GHC works registerised and has a native code
1006 generator. You <Emphasis>must</Emphasis> have GCC 2.7.x
1007 or later. NOTE about <literal>glibc</literal> versions:
1008 GHC binaries built on a system running <literal>glibc
1009 2.0</literal> won't work on a system running
1010 <literal>glibc 2.1</literal>, and vice versa. In general,
1011 don't expect compatibility between
1012 <literal>glibc</literal> versions, even if the shared
1013 library version hasn't changed.</para>
1018 <term>i386-unknown-freebsd (PCs running FreeBSD 2.2 or
1020 <indexterm><primary>i386-unknown-freebsd</primary></indexterm>
1022 <para>GHC works registerised. Pre-built packages are
1023 available in the native package format, so if you just
1024 need binaries you're better off just installing the
1025 package (it might even be on your installation
1031 <term>i386-unknown-openbsd (PCs running OpenBSD)</term>
1032 <indexterm><primary>i386-unknown-openbsd</primary></indexterm>
1034 <para>Supported, with native code generator. Packages are
1035 available through the ports system in the native package
1041 <term>i386-unknown-netbsd (PCs running NetBSD and
1043 <indexterm><primary>i386-unknown-netbsd</primary></indexterm>
1045 <para>Will require some minor porting effort, but should
1046 work registerised.</para>
1051 <term>i386-unknown-mingw32 (PCs running Windows)</term>
1052 <indexterm><primary>i386-unknown-mingw32</primary></indexterm>
1054 <para>Fully supported under Win9x, WinNT, Win2k, and
1055 WinXP. Includes a native code generator. Building from
1056 source requires a recent <ulink
1057 url="http://www.cygwin.com/">Cygwin</ulink> distribution
1058 to be installed.</para>
1063 <term>mips-sgi-irix5</term>
1064 <indexterm><primary>mips-sgi-irix[5-6]</primary></indexterm>
1066 <para>Port has worked in the past, but hasn't been tested
1067 for some time (and will certainly have rotted in various
1068 ways). As usual, we don't have access to machines and
1069 there hasn't been an overwhelming demand for this port,
1070 but feel free to get in touch.</para>
1075 <term>powerpc-ibm-aix</term>
1076 <indexterm><primary>powerpc-ibm-aix</primary></indexterm>
1078 <para>Port currently doesn't work, needs some minimal
1079 porting effort. As usual, we don't have access to
1080 machines and there hasn't been an overwhelming demand for
1081 this port, but feel free to get in touch.</para>
1086 <term>powerpc-apple-darwin</term>
1087 <indexterm><primary>powerpc-apple-darwin</primary></indexterm>
1089 <para>Supported registerised. No native code
1095 <term>powerpc-apple-linux</term>
1096 <indexterm><primary>powerpc-apple-linux</primary></indexterm>
1098 <para>Not supported (yet).</para>
1103 <para>Various other systems have had GHC ported to them in the
1104 distant past, including various Motorola 68k boxes. The 68k
1105 support still remains, but porting to one of these systems will
1106 certainly be a non-trivial task.</para>
1110 <title>What machines the other tools run on</title>
1112 <para>Unless you hear otherwise, the other tools work if GHC
1118 <sect1 id="sec-pre-supposed">
1119 <title>Installing pre-supposed utilities</title>
1121 <indexterm><primary>pre-supposed utilities</primary></indexterm>
1122 <indexterm><primary>utilities, pre-supposed</primary></indexterm>
1124 <para>Here are the gory details about some utility programs you
1125 may need; <command>perl</command>, <command>gcc</command> and
1126 <command>happy</command> are the only important
1127 ones. (PVM<indexterm><primary>PVM</primary></indexterm> is
1128 important if you're going for Parallel Haskell.) The
1129 <command>configure</command><indexterm><primary>configure</primary></indexterm>
1130 script will tell you if you are missing something.</para>
1136 <indexterm><primary>pre-supposed: Perl</primary></indexterm>
1137 <indexterm><primary>Perl, pre-supposed</primary></indexterm>
1139 <para><emphasis>You have to have Perl to proceed!</emphasis>
1140 Perl version 5 at least is required. GHC has been known to
1141 tickle bugs in Perl, so if you find that Perl crashes when
1142 running GHC try updating (or downgrading) your Perl
1143 installation. Versions of Perl that we use and are known to
1144 be fairly stable are 5.005 and 5.6.1.</para>
1146 <para>For Win32 platforms, you should use the binary
1147 supplied in the InstallShield (copy it to
1148 <filename>/bin</filename>). The Cygwin-supplied Perl seems
1151 <para>Perl should be put somewhere so that it can be invoked
1152 by the <literal>#!</literal> script-invoking
1153 mechanism. The full pathname may need to be less than 32
1154 characters long on some systems.</para>
1159 <term>GNU C (<command>gcc</command>)</term>
1160 <indexterm><primary>pre-supposed: GCC (GNU C
1161 compiler)</primary></indexterm> <indexterm><primary>GCC (GNU C
1162 compiler), pre-supposed</primary></indexterm>
1164 <para>We recommend using GCC version 2.95.2 on all
1165 platforms. Failing that, version 2.7.2 is stable on most
1166 platforms. Earlier versions of GCC can be assumed not to
1167 work, and versions in between 2.7.2 and 2.95.2 (including
1168 <command>egcs</command>) have varying degrees of stability
1169 depending on the platform.</para>
1171 <para>If your GCC dies with “internal error” on
1172 some GHC source file, please let us know, so we can report
1173 it and get things improved. (Exception: on iX86
1174 boxes—you may need to fiddle with GHC's
1175 <option>-monly-N-regs</option> option; see the User's
1181 <term>GNU Make</term>
1182 <indexterm><primary>make</primary><secondary>GNU</secondary>
1185 <para>The fptools build system makes heavy use of features
1186 specific to GNU <command>make</command>, so you must have
1187 this installed in order to build any of the fptools
1194 <indexterm><primary>Happy</primary></indexterm>
1196 <para>Happy is a parser generator tool for Haskell, and is
1197 used to generate GHC's parsers. Happy is written in
1198 Haskell, and is a project in the CVS repository
1199 (<literal>fptools/happy</literal>). It can be built from
1200 source, but bear in mind that you'll need GHC installed in
1201 order to build it. To avoid the chicken/egg problem,
1202 install a binary distribtion of either Happy or GHC to get
1203 started. Happy distributions are available from <ulink
1204 url="http://www.haskell.org/happy/">Happy's Web
1205 Page</ulink>.</para>
1210 <term>Autoconf</term>
1211 <indexterm><primary>pre-supposed: Autoconf</primary></indexterm>
1212 <indexterm><primary>Autoconf, pre-supposed</primary></indexterm>
1214 <para>GNU Autoconf is needed if you intend to build from the
1215 CVS sources, it is <emphasis>not</emphasis> needed if you
1216 just intend to build a standard source distribution.</para>
1218 <para>Autoconf builds the <command>configure</command>
1219 script from <filename>configure.in</filename> and
1220 <filename>aclocal.m4</filename>. If you modify either of
1221 these files, you'll need <command>autoconf</command> to
1222 rebuild <filename>configure</filename>.</para>
1227 <term><command>sed</command></term>
1228 <indexterm><primary>pre-supposed: sed</primary></indexterm>
1229 <indexterm><primary>sed, pre-supposed</primary></indexterm>
1231 <para>You need a working <command>sed</command> if you are
1232 going to build from sources. The build-configuration stuff
1233 needs it. GNU sed version 2.0.4 is no good! It has a bug
1234 in it that is tickled by the build-configuration. 2.0.5 is
1235 OK. Others are probably OK too (assuming we don't create too
1236 elaborate configure scripts.)</para>
1241 <para>One <literal>fptools</literal> project is worth a quick note
1242 at this point, because it is useful for all the others:
1243 <literal>glafp-utils</literal> contains several utilities which
1244 aren't particularly Glasgow-ish, but Occasionally Indispensable.
1245 Like <command>lndir</command> for creating symbolic link
1248 <sect2 id="pre-supposed-gph-tools">
1249 <title>Tools for building parallel GHC (GPH)</title>
1253 <term>PVM version 3:</term>
1254 <indexterm><primary>pre-supposed: PVM3 (Parallel Virtual Machine)</primary></indexterm>
1255 <indexterm><primary>PVM3 (Parallel Virtual Machine), pre-supposed</primary></indexterm>
1257 <para>PVM is the Parallel Virtual Machine on which
1258 Parallel Haskell programs run. (You only need this if you
1259 plan to run Parallel Haskell. Concurent Haskell, which
1260 runs concurrent threads on a uniprocessor doesn't need
1261 it.) Underneath PVM, you can have (for example) a network
1262 of workstations (slow) or a multiprocessor box
1265 <para>The current version of PVM is 3.3.11; we use 3.3.7.
1266 It is readily available on the net; I think I got it from
1267 <literal>research.att.com</literal>, in
1268 <filename>netlib</filename>.</para>
1270 <para>A PVM installation is slightly quirky, but easy to
1271 do. Just follow the <filename>Readme</filename>
1272 instructions.</para>
1277 <term><command>bash</command>:</term>
1278 <indexterm><primary>bash, presupposed (Parallel Haskell only)</primary></indexterm>
1280 <para>Sadly, the <command>gr2ps</command> script, used to
1281 convert “parallelism profiles” to PostScript,
1282 is written in Bash (GNU's Bourne Again shell). This bug
1283 will be fixed (someday).</para>
1289 <sect2 id="pre-supposed-doc-tools">
1290 <title>Tools for building the Documentation</title>
1292 <para>The following additional tools are required if you want to
1293 format the documentation that comes with the
1294 <literal>fptools</literal> projects:</para>
1298 <term>DocBook</term>
1299 <indexterm><primary>pre-supposed: DocBook</primary></indexterm>
1300 <indexterm><primary>DocBook, pre-supposed</primary></indexterm>
1302 <para>All our documentation is written in SGML, using the
1303 DocBook DTD. Instructions on installing and configuring
1304 the DocBook tools are in the installation guide (in the
1305 GHC user guide).</para>
1311 <indexterm><primary>pre-supposed: TeX</primary></indexterm>
1312 <indexterm><primary>TeX, pre-supposed</primary></indexterm>
1314 <para>A decent TeX distribution is required if you want to
1315 produce printable documentation. We recomment teTeX,
1316 which includes just about everything you need.</para>
1321 <para> In order to actually build any documentation, you need to
1322 set <constant>SGMLDocWays</constant> in your
1323 <filename>build.mk</filename>. Valid values to add to this list
1324 are: <literal>dvi</literal>, <literal>ps</literal>,
1325 <literal>pdf</literal>, <literal>html</literal>, and
1326 <literal>rtf</literal>.</para>
1329 <sect2 id="pre-supposed-other-tools">
1330 <title>Other useful tools</title>
1335 <indexterm><primary>pre-supposed: flex</primary></indexterm>
1336 <indexterm><primary>flex, pre-supposed</primary></indexterm>
1338 <para>This is a quite-a-bit-better-than-Lex lexer. Used
1339 to build a couple of utilities in
1340 <literal>glafp-utils</literal>. Depending on your
1341 operating system, the supplied <command>lex</command> may
1342 or may not work; you should get the GNU version.</para>
1349 <sect1 id="sec-building-from-source">
1350 <title>Building from source</title>
1352 <indexterm><primary>Building from source</primary></indexterm>
1353 <indexterm><primary>Source, building from</primary></indexterm>
1355 <para>You've been rash enough to want to build some of the Glasgow
1356 Functional Programming tools (GHC, Happy, nofib, etc.) from
1357 source. You've slurped the source, from the CVS repository or
1358 from a source distribution, and now you're sitting looking at a
1359 huge mound of bits, wondering what to do next.</para>
1361 <para>Gingerly, you type <command>make</command>. Wrong
1364 <para>This rest of this guide is intended for duffers like me, who
1365 aren't really interested in Makefiles and systems configurations,
1366 but who need a mental model of the interlocking pieces so that
1367 they can make them work, extend them consistently when adding new
1368 software, and lay hands on them gently when they don't
1371 <sect2 id="sec-source-tree">
1372 <title>Your source tree</title>
1374 <para>The source code is held in your <emphasis>source
1375 tree</emphasis>. The root directory of your source tree
1376 <emphasis>must</emphasis> contain the following directories and
1381 <para><filename>Makefile</filename>: the root
1386 <para><filename>mk/</filename>: the directory that contains
1387 the main Makefile code, shared by all the
1388 <literal>fptools</literal> software.</para>
1392 <para><filename>configure.in</filename>,
1393 <filename>config.sub</filename>,
1394 <filename>config.guess</filename>: these files support the
1395 configuration process.</para>
1399 <para><filename>install-sh</filename>.</para>
1403 <para>All the other directories are individual
1404 <emphasis>projects</emphasis> of the <literal>fptools</literal>
1405 system—for example, the Glasgow Haskell Compiler
1406 (<literal>ghc</literal>), the Happy parser generator
1407 (<literal>happy</literal>), the <literal>nofib</literal>
1408 benchmark suite, and so on. You can have zero or more of these.
1409 Needless to say, some of them are needed to build others.</para>
1411 <para>The important thing to remember is that even if you want
1412 only one project (<literal>happy</literal>, say), you must have
1413 a source tree whose root directory contains
1414 <filename>Makefile</filename>, <filename>mk/</filename>,
1415 <filename>configure.in</filename>, and the project(s) you want
1416 (<filename>happy/</filename> in this case). You cannot get by
1417 with just the <filename>happy/</filename> directory.</para>
1421 <title>Build trees</title>
1422 <indexterm><primary>build trees</primary></indexterm>
1423 <indexterm><primary>link trees, for building</primary></indexterm>
1425 <para>If you just want to build the software once on a single
1426 platform, then your source tree can also be your build tree, and
1427 you can skip the rest of this section.</para>
1429 <para>We often want to build multiple versions of our software
1430 for different architectures, or with different options
1431 (e.g. profiling). It's very desirable to share a single copy of
1432 the source code among all these builds.</para>
1434 <para>So for every source tree we have zero or more
1435 <emphasis>build trees</emphasis>. Each build tree is initially
1436 an exact copy of the source tree, except that each file is a
1437 symbolic link to the source file, rather than being a copy of
1438 the source file. There are “standard” Unix
1439 utilities that make such copies, so standard that they go by
1441 <command>lndir</command><indexterm><primary>lndir</primary></indexterm>,
1442 <command>mkshadowdir</command><indexterm><primary>mkshadowdir</primary></indexterm>
1443 are two (If you don't have either, the source distribution
1444 includes sources for the X11
1445 <command>lndir</command>—check out
1446 <filename>fptools/glafp-utils/lndir</filename>). See <Xref
1447 LinkEnd="sec-storysofar"> for a typical invocation.</para>
1449 <para>The build tree does not need to be anywhere near the
1450 source tree in the file system. Indeed, one advantage of
1451 separating the build tree from the source is that the build tree
1452 can be placed in a non-backed-up partition, saving your systems
1453 support people from backing up untold megabytes of
1454 easily-regenerated, and rapidly-changing, gubbins. The golden
1455 rule is that (with a single exception—<XRef
1456 LinkEnd="sec-build-config">) <emphasis>absolutely everything in
1457 the build tree is either a symbolic link to the source tree, or
1458 else is mechanically generated</emphasis>. It should be
1459 perfectly OK for your build tree to vanish overnight; an hour or
1460 two compiling and you're on the road again.</para>
1462 <para>You need to be a bit careful, though, that any new files
1463 you create (if you do any development work) are in the source
1464 tree, not a build tree!</para>
1466 <para>Remember, that the source files in the build tree are
1467 <emphasis>symbolic links</emphasis> to the files in the source
1468 tree. (The build tree soon accumulates lots of built files like
1469 <filename>Foo.o</filename>, as well.) You can
1470 <emphasis>delete</emphasis> a source file from the build tree
1471 without affecting the source tree (though it's an odd thing to
1472 do). On the other hand, if you <emphasis>edit</emphasis> a
1473 source file from the build tree, you'll edit the source-tree
1474 file directly. (You can set up Emacs so that if you edit a
1475 source file from the build tree, Emacs will silently create an
1476 edited copy of the source file in the build tree, leaving the
1477 source file unchanged; but the danger is that you think you've
1478 edited the source file whereas actually all you've done is edit
1479 the build-tree copy. More commonly you do want to edit the
1480 source file.)</para>
1482 <para>Like the source tree, the top level of your build tree
1483 must be (a linked copy of) the root directory of the
1484 <literal>fptools</literal> suite. Inside Makefiles, the root of
1485 your build tree is called
1486 <constant>$(FPTOOLS_TOP)</constant><indexterm><primary>FPTOOLS_TOP</primary></indexterm>.
1487 In the rest of this document path names are relative to
1488 <constant>$(FPTOOLS_TOP)</constant> unless
1489 otherwise stated. For example, the file
1490 <filename>ghc/mk/target.mk</filename> is actually
1491 <filename><constant>$(FPTOOLS_TOP)</constant>/ghc/mk/target.mk</filename>.</para>
1494 <sect2 id="sec-build-config">
1495 <title>Getting the build you want</title>
1497 <para>When you build <literal>fptools</literal> you will be
1498 compiling code on a particular <emphasis>host
1499 platform</emphasis>, to run on a particular <emphasis>target
1500 platform</emphasis> (usually the same as the host
1501 platform)<indexterm><primary>platform</primary></indexterm>.
1502 The difficulty is that there are minor differences between
1503 different platforms; minor, but enough that the code needs to be
1504 a bit different for each. There are some big differences too:
1505 for a different architecture we need to build GHC with a
1506 different native-code generator.</para>
1508 <para>There are also knobs you can turn to control how the
1509 <literal>fptools</literal> software is built. For example, you
1510 might want to build GHC optimised (so that it runs fast) or
1511 unoptimised (so that you can compile it fast after you've
1512 modified it. Or, you might want to compile it with debugging on
1513 (so that extra consistency-checking code gets included) or off.
1516 <para>All of this stuff is called the
1517 <emphasis>configuration</emphasis> of your build. You set the
1518 configuration using a three-step process.</para>
1522 <term>Step 1: get ready for configuration.</term>
1524 <para>Change directory to
1525 <constant>$(FPTOOLS_TOP)</constant> and
1527 <command>autoconf</command><indexterm><primary>autoconf</primary></indexterm>
1528 (with no arguments). This GNU program converts
1529 <filename><constant>$(FPTOOLS_TOP)</constant>/configure.in</filename>
1530 to a shell script called
1531 <filename><constant>$(FPTOOLS_TOP)</constant>/configure</filename>.
1534 <para>Some projects, including GHC, have their own
1535 configure script. If there's an
1536 <constant>$(FPTOOLS_TOP)/<project>/configure.in</constant>,
1537 then you need to run <command>autoconf</command> in that
1538 directory too.</para>
1540 <para>Both these steps are completely
1541 platform-independent; they just mean that the
1542 human-written file (<filename>configure.in</filename>) can
1543 be short, although the resulting shell script,
1544 <command>configure</command>, and
1545 <filename>mk/config.h.in</filename>, are long.</para>
1547 <para>In case you don't have <command>autoconf</command>
1548 we distribute the results, <command>configure</command>,
1549 and <filename>mk/config.h.in</filename>, with the source
1550 distribution. They aren't kept in the repository,
1556 <term>Step 2: system configuration.</term>
1558 <para>Runs the newly-created <command>configure</command>
1559 script, thus:</para>
1562 ./configure <optional><parameter>args</parameter></optional>
1565 <para><command>configure</command>'s mission is to scurry
1566 round your computer working out what architecture it has,
1567 what operating system, whether it has the
1568 <Function>vfork</Function> system call, where
1569 <command>yacc</command> is kept, whether
1570 <command>gcc</command> is available, where various obscure
1571 <literal>#include</literal> files are, whether it's a
1572 leap year, and what the systems manager had for lunch. It
1573 communicates these snippets of information in two
1580 <filename>mk/config.mk.in</filename><indexterm><primary>config.mk.in</primary></indexterm>
1582 <filename>mk/config.mk</filename><indexterm><primary>config.mk</primary></indexterm>,
1583 substituting for things between
1584 “<literal>@</literal>” brackets. So,
1585 “<literal>@HaveGcc@</literal>” will be
1586 replaced by “<literal>YES</literal>” or
1587 “<literal>NO</literal>” depending on what
1588 <command>configure</command> finds.
1589 <filename>mk/config.mk</filename> is included by every
1590 Makefile (directly or indirectly), so the
1591 configuration information is thereby communicated to
1592 all Makefiles.</para>
1596 <para> It translates
1597 <filename>mk/config.h.in</filename><indexterm><primary>config.h.in</primary></indexterm>
1599 <filename>mk/config.h</filename><indexterm><primary>config.h</primary></indexterm>.
1600 The latter is <literal>#include</literal>d by
1601 various C programs, which can thereby make use of
1602 configuration information.</para>
1606 <para><command>configure</command> takes some optional
1607 arguments. Use <literal>./configure --help</literal> to
1608 get a list of the available arguments. Here are some of
1609 the ones you might need:</para>
1613 <term><literal>--with-ghc=<parameter>path</parameter></literal></term>
1614 <indexterm><primary><literal>--with-ghc</literal></primary>
1617 <para>Specifies the path to an installed GHC which
1618 you would like to use. This compiler will be used
1619 for compiling GHC-specific code (eg. GHC itself).
1620 This option <emphasis>cannot</emphasis> be specified
1621 using <filename>build.mk</filename> (see later),
1622 because <command>configure</command> needs to
1623 auto-detect the version of GHC you're using. The
1624 default is to look for a compiler named
1625 <literal>ghc</literal> in your path.</para>
1630 <term><literal>--with-hc=<parameter>path</parameter></literal></term>
1631 <indexterm><primary><literal>--with-hc</literal></primary>
1634 <para>Specifies the path to any installed Haskell
1635 compiler. This compiler will be used for compiling
1636 generic Haskell code. The default is to use
1637 <literal>ghc</literal>.</para>
1642 <term><literal>--with-gcc=<parameter>path</parameter></literal></term>
1643 <indexterm><primary><literal>--with-gcc</literal></primary>
1646 <para>Specifies the path to the installed GCC. This
1647 compiler will be used to compile all C files,
1648 <emphasis>except</emphasis> any generated by the
1649 installed Haskell compiler, which will have its own
1650 idea of which C compiler (if any) to use. The
1651 default is to use <literal>gcc</literal>.</para>
1656 <para><command>configure</command> caches the results of
1657 its run in <filename>config.cache</filename>. Quite often
1658 you don't want that; you're running
1659 <command>configure</command> a second time because
1660 something has changed. In that case, simply delete
1661 <filename>config.cache</filename>.</para>
1666 <term>Step 3: build configuration.</term>
1668 <para>Next, you say how this build of
1669 <literal>fptools</literal> is to differ from the standard
1670 defaults by creating a new file
1671 <filename>mk/build.mk</filename><indexterm><primary>build.mk</primary></indexterm>
1672 <emphasis>in the build tree</emphasis>. This file is the
1673 one and only file you edit in the build tree, precisely
1674 because it says how this build differs from the source.
1675 (Just in case your build tree does die, you might want to
1676 keep a private directory of <filename>build.mk</filename>
1677 files, and use a symbolic link in each build tree to point
1678 to the appropriate one.) So
1679 <filename>mk/build.mk</filename> never exists in the
1680 source tree—you create one in each build tree from
1681 the template. We'll discuss what to put in it
1687 <para>And that's it for configuration. Simple, eh?</para>
1689 <para>What do you put in your build-specific configuration file
1690 <filename>mk/build.mk</filename>? <emphasis>For almost all
1691 purposes all you will do is put make variable definitions that
1692 override those in</emphasis>
1693 <filename>mk/config.mk.in</filename>. The whole point of
1694 <filename>mk/config.mk.in</filename>—and its derived
1695 counterpart <filename>mk/config.mk</filename>—is to define
1696 the build configuration. It is heavily commented, as you will
1697 see if you look at it. So generally, what you do is look at
1698 <filename>mk/config.mk.in</filename>, and add definitions in
1699 <filename>mk/build.mk</filename> that override any of the
1700 <filename>config.mk</filename> definitions that you want to
1701 change. (The override occurs because the main boilerplate file,
1702 <filename>mk/boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>,
1703 includes <filename>build.mk</filename> after
1704 <filename>config.mk</filename>.)</para>
1706 <para>For example, <filename>config.mk.in</filename> contains
1707 the definition:</para>
1710 GhcHcOpts=-O -Rghc-timing
1713 <para>The accompanying comment explains that this is the list of
1714 flags passed to GHC when building GHC itself. For doing
1715 development, it is wise to add <literal>-DDEBUG</literal>, to
1716 enable debugging code. So you would add the following to
1717 <filename>build.mk</filename>:</para>
1719 <para>or, if you prefer,</para>
1722 GhcHcOpts += -DDEBUG
1725 <para>GNU <command>make</command> allows existing definitions to
1726 have new text appended using the “<literal>+=</literal>”
1727 operator, which is quite a convenient feature.)</para>
1729 <para>If you want to remove the <literal>-O</literal> as well (a
1730 good idea when developing, because the turn-around cycle gets a
1731 lot quicker), you can just override
1732 <literal>GhcLibHcOpts</literal> altogether:</para>
1735 GhcHcOpts=-DDEBUG -Rghc-timing
1738 <para>When reading <filename>config.mk.in</filename>, remember
1739 that anything between “@...@” signs is going to be substituted
1740 by <command>configure</command> later. You
1741 <emphasis>can</emphasis> override the resulting definition if
1742 you want, but you need to be a bit surer what you are doing.
1743 For example, there's a line that says:</para>
1749 <para>This defines the Make variables <constant>YACC</constant>
1750 to the pathname for a <command>yacc</command> that
1751 <command>configure</command> finds somewhere. If you have your
1752 own pet <command>yacc</command> you want to use instead, that's
1753 fine. Just add this line to <filename>mk/build.mk</filename>:</para>
1759 <para>You do not <emphasis>have</emphasis> to have a
1760 <filename>mk/build.mk</filename> file at all; if you don't,
1761 you'll get all the default settings from
1762 <filename>mk/config.mk.in</filename>.</para>
1764 <para>You can also use <filename>build.mk</filename> to override
1765 anything that <command>configure</command> got wrong. One place
1766 where this happens often is with the definition of
1767 <constant>FPTOOLS_TOP_ABS</constant>: this
1768 variable is supposed to be the canonical path to the top of your
1769 source tree, but if your system uses an automounter then the
1770 correct directory is hard to find automatically. If you find
1771 that <command>configure</command> has got it wrong, just put the
1772 correct definition in <filename>build.mk</filename>.</para>
1776 <sect2 id="sec-storysofar">
1777 <title>The story so far</title>
1779 <para>Let's summarise the steps you need to carry to get
1780 yourself a fully-configured build tree from scratch.</para>
1784 <para> Get your source tree from somewhere (CVS repository
1785 or source distribution). Say you call the root directory
1786 <filename>myfptools</filename> (it does not have to be
1787 called <filename>fptools</filename>). Make sure that you
1788 have the essential files (see <XRef
1789 LinkEnd="sec-source-tree">).</para>
1794 <para>(Optional) Use <command>lndir</command> or
1795 <command>mkshadowdir</command> to create a build tree.</para>
1799 $ mkshadowdir . /scratch/joe-bloggs/myfptools-sun4
1802 <para>(N.B. <command>mkshadowdir</command>'s first argument
1803 is taken relative to its second.) You probably want to give
1804 the build tree a name that suggests its main defining
1805 characteristic (in your mind at least), in case you later
1810 <para>Change directory to the build tree. Everything is
1811 going to happen there now.</para>
1814 $ cd /scratch/joe-bloggs/myfptools-sun4
1820 <para>Prepare for system configuration:</para>
1826 <para>(You can skip this step if you are starting from a
1827 source distribution, and you already have
1828 <filename>configure</filename> and
1829 <filename>mk/config.h.in</filename>.)</para>
1831 <para>Some projects, including GHC itself, have their own
1832 configure scripts, so it is necessary to run autoconf again
1833 in the appropriate subdirectories. eg:</para>
1836 $ (cd ghc; autoconf)
1841 <para>Do system configuration:</para>
1847 <para>Don't forget to check whether you need to add any
1848 arguments to <literal>configure</literal>; for example, a
1849 common requirement is to specify which GHC to use with
1850 <option>--with-ghc=<replaceable>ghc</replaceable></option>.</para>
1854 <para>Create the file <filename>mk/build.mk</filename>,
1855 adding definitions for your desired configuration
1864 <para>You can make subsequent changes to
1865 <filename>mk/build.mk</filename> as often as you like. You do
1866 not have to run any further configuration programs to make these
1867 changes take effect. In theory you should, however, say
1868 <command>gmake clean</command>, <command>gmake all</command>,
1869 because configuration option changes could affect
1870 anything—but in practice you are likely to know what's
1875 <title>Making things</title>
1877 <para>At this point you have made yourself a fully-configured
1878 build tree, so you are ready to start building real
1881 <para>The first thing you need to know is that <emphasis>you
1882 must use GNU <command>make</command>, usually called
1883 <command>gmake</command>, not standard Unix
1884 <command>make</command></emphasis>. If you use standard Unix
1885 <command>make</command> you will get all sorts of error messages
1886 (but no damage) because the <literal>fptools</literal>
1887 <command>Makefiles</command> use GNU <command>make</command>'s
1888 facilities extensively.</para>
1890 <para>To just build the whole thing, <command>cd</command> to
1891 the top of your <literal>fptools</literal> tree and type
1892 <command>gmake</command>. This will prepare the tree and build
1893 the various projects in the correct order.</para>
1897 <sect2 id="sec-standard-targets">
1898 <title>Standard Targets</title>
1899 <indexterm><primary>targets, standard makefile</primary></indexterm>
1900 <indexterm><primary>makefile targets</primary></indexterm>
1902 <para>In any directory you should be able to make the following:</para>
1906 <term><literal>boot</literal></term>
1908 <para>does the one-off preparation required to get ready
1909 for the real work. Notably, it does <command>gmake
1910 depend</command> in all directories that contain programs.
1911 It also builds the necessary tools for compilation to
1914 <para>Invoking the <literal>boot</literal> target
1915 explicitly is not normally necessary. From the top-level
1916 <literal>fptools</literal> directory, invoking
1917 <literal>gmake</literal> causes <literal>gmake boot
1918 all</literal> to be invoked in each of the project
1919 subdirectories, in the order specified by
1920 <literal>$(AllTargets)</literal> in
1921 <literal>config.mk</literal>.</para>
1923 <para>If you're working in a subdirectory somewhere and
1924 need to update the dependencies, <literal>gmake
1925 boot</literal> is a good way to do it.</para>
1930 <term><literal>all</literal></term>
1932 <para>makes all the final target(s) for this Makefile.
1933 Depending on which directory you are in a “final
1934 target” may be an executable program, a library
1935 archive, a shell script, or a Postscript file. Typing
1936 <command>gmake</command> alone is generally the same as
1937 typing <command>gmake all</command>.</para>
1942 <term><literal>install</literal></term>
1944 <para>installs the things built by <literal>all</literal>
1945 (except for the documentation). Where does it install
1946 them? That is specified by
1947 <filename>mk/config.mk.in</filename>; you can override it
1948 in <filename>mk/build.mk</filename>, or by running
1949 <command>configure</command> with command-line arguments
1950 like <literal>--bindir=/home/simonpj/bin</literal>; see
1951 <literal>./configure --help</literal> for the full
1957 <term><literal>install-docs</literal></term>
1959 <para>installs the documentation. Otherwise behaves just
1960 like <literal>install</literal>.</para>
1965 <term><literal>uninstall</literal></term>
1967 <para>reverses the effect of
1968 <literal>install</literal>.</para>
1973 <term><literal>clean</literal></term>
1975 <para>Delete all files from the current directory that are
1976 normally created by building the program. Don't delete
1977 the files that record the configuration, or files
1978 generated by <command>gmake boot</command>. Also preserve
1979 files that could be made by building, but normally aren't
1980 because the distribution comes with them.</para>
1985 <term><literal>distclean</literal></term>
1987 <para>Delete all files from the current directory that are
1988 created by configuring or building the program. If you
1989 have unpacked the source and built the program without
1990 creating any other files, <literal>make
1991 distclean</literal> should leave only the files that were
1992 in the distribution.</para>
1997 <term><literal>mostlyclean</literal></term>
1999 <para>Like <literal>clean</literal>, but may refrain from
2000 deleting a few files that people normally don't want to
2006 <term><literal>maintainer-clean</literal></term>
2008 <para>Delete everything from the current directory that
2009 can be reconstructed with this Makefile. This typically
2010 includes everything deleted by
2011 <literal>distclean</literal>, plus more: C source files
2012 produced by Bison, tags tables, Info files, and so
2015 <para>One exception, however: <literal>make
2016 maintainer-clean</literal> should not delete
2017 <filename>configure</filename> even if
2018 <filename>configure</filename> can be remade using a rule
2019 in the <filename>Makefile</filename>. More generally,
2020 <literal>make maintainer-clean</literal> should not delete
2021 anything that needs to exist in order to run
2022 <filename>configure</filename> and then begin to build the
2028 <term><literal>check</literal></term>
2030 <para>run the test suite.</para>
2035 <para>All of these standard targets automatically recurse into
2036 sub-directories. Certain other standard targets do not:</para>
2040 <term><literal>configure</literal></term>
2042 <para>is only available in the root directory
2043 <constant>$(FPTOOLS_TOP)</constant>; it has
2044 been discussed in <XRef
2045 LinkEnd="sec-build-config">.</para>
2050 <term><literal>depend</literal></term>
2052 <para>make a <filename>.depend</filename> file in each
2053 directory that needs it. This <filename>.depend</filename>
2054 file contains mechanically-generated dependency
2055 information; for example, suppose a directory contains a
2056 Haskell source module <filename>Foo.lhs</filename> which
2057 imports another module <literal>Baz</literal>. Then the
2058 generated <filename>.depend</filename> file will contain
2059 the dependency:</para>
2065 <para>which says that the object file
2066 <filename>Foo.o</filename> depends on the interface file
2067 <filename>Baz.hi</filename> generated by compiling module
2068 <literal>Baz</literal>. The <filename>.depend</filename>
2069 file is automatically included by every Makefile.</para>
2074 <term><literal>binary-dist</literal></term>
2076 <para>make a binary distribution. This is the target we
2077 use to build the binary distributions of GHC and
2083 <term><literal>dist</literal></term>
2085 <para>make a source distribution. Note that this target
2086 does “make distclean” as part of its work;
2087 don't use it if you want to keep what you've built.</para>
2092 <para>Most <filename>Makefile</filename>s have targets other
2093 than these. You can discover them by looking in the
2094 <filename>Makefile</filename> itself.</para>
2098 <title>Using a project from the build tree</title>
2100 <para>If you want to build GHC (say) and just use it direct from
2101 the build tree without doing <literal>make install</literal>
2102 first, you can run the in-place driver script:
2103 <filename>ghc/compiler/ghc-inplace</filename>.</para>
2105 <para> Do <emphasis>NOT</emphasis> use
2106 <filename>ghc/compiler/ghc</filename>, or
2107 <filename>ghc/compiler/ghc-5.xx</filename>, as these are the
2108 scripts intended for installation, and contain hard-wired paths
2109 to the installed libraries, rather than the libraries in the
2112 <para>Happy can similarly be run from the build tree, using
2113 <filename>happy/src/happy-inplace</filename>.</para>
2117 <title>Fast Making</title>
2119 <indexterm><primary>fastmake</primary></indexterm>
2120 <indexterm><primary>dependencies, omitting</primary></indexterm>
2121 <indexterm><primary>FAST, makefile variable</primary></indexterm>
2123 <para>Sometimes the dependencies get in the way: if you've made
2124 a small change to one file, and you're absolutely sure that it
2125 won't affect anything else, but you know that
2126 <command>make</command> is going to rebuild everything anyway,
2127 the following hack may be useful:</para>
2133 <para>This tells the make system to ignore dependencies and just
2134 build what you tell it to. In other words, it's equivalent to
2135 temporarily removing the <filename>.depend</filename> file in
2136 the current directory (where <command>mkdependHS</command> and
2137 friends store their dependency information).</para>
2139 <para>A bit of history: GHC used to come with a
2140 <command>fastmake</command> script that did the above job, but
2141 GNU make provides the features we need to do it without
2142 resorting to a script. Also, we've found that fastmaking is
2143 less useful since the advent of GHC's recompilation checker (see
2144 the User's Guide section on "Separate Compilation").</para>
2148 <sect1 id="sec-makefile-arch">
2149 <title>The <filename>Makefile</filename> architecture</title>
2150 <indexterm><primary>makefile architecture</primary></indexterm>
2152 <para><command>make</command> is great if everything
2153 works—you type <command>gmake install</command> and lo! the
2154 right things get compiled and installed in the right places. Our
2155 goal is to make this happen often, but somehow it often doesn't;
2156 instead some weird error message eventually emerges from the
2157 bowels of a directory you didn't know existed.</para>
2159 <para>The purpose of this section is to give you a road-map to
2160 help you figure out what is going right and what is going
2164 <title>Debugging</title>
2166 <para>Debugging <filename>Makefile</filename>s is something of a
2167 black art, but here's a couple of tricks that we find
2168 particularly useful. The following command allows you to see
2169 the contents of any make variable in the context of the current
2170 <filename>Makefile</filename>:</para>
2172 <screen>$ make show VALUE=HS_SRCS</screen>
2174 <para>where you can replace <literal>HS_SRCS</literal> with the
2175 name of any variable you wish to see the value of.</para>
2177 <para>GNU make has a <option>-d</option> option which generates
2178 a dump of the decision procedure used to arrive at a conclusion
2179 about which files should be recompiled. Sometimes useful for
2180 tracking down problems with superfluous or missing
2181 recompilations.</para>
2185 <title>A small project</title>
2187 <para>To get started, let us look at the
2188 <filename>Makefile</filename> for an imaginary small
2189 <literal>fptools</literal> project, <literal>small</literal>.
2190 Each project in <literal>fptools</literal> has its own directory
2191 in <constant>FPTOOLS_TOP</constant>, so the
2192 <literal>small</literal> project will have its own directory
2193 <constant>FPOOLS_TOP/small/</constant>. Inside the
2194 <filename>small/</filename> directory there will be a
2195 <filename>Makefile</filename>, looking something like
2198 <indexterm><primary>Makefile, minimal</primary></indexterm>
2201 # Makefile for fptools project "small"
2204 include $(TOP)/mk/boilerplate.mk
2206 SRCS = $(wildcard *.lhs) $(wildcard *.c)
2209 include $(TOP)/target.mk
2212 <para>this <filename>Makefile</filename> has three
2217 <para>The first section includes
2220 One of the most important
2221 features of GNU <command>make</command> that we use is the ability for a <filename>Makefile</filename> to
2222 include another named file, very like <command>cpp</command>'s <literal>#include</literal>
2227 a file of “boilerplate” code from the level
2228 above (which in this case will be
2229 <filename><constant>FPTOOLS_TOP</constant>/mk/boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>).
2230 As its name suggests, <filename>boilerplate.mk</filename>
2231 consists of a large quantity of standard
2232 <filename>Makefile</filename> code. We discuss this
2233 boilerplate in more detail in <XRef LinkEnd="sec-boiler">.
2234 <indexterm><primary>include, directive in
2235 Makefiles</primary></indexterm> <indexterm><primary>Makefile
2236 inclusion</primary></indexterm></para>
2238 <para>Before the <literal>include</literal> statement, you
2239 must define the <command>make</command> variable
2240 <constant>TOP</constant><indexterm><primary>TOP</primary></indexterm>
2241 to be the directory containing the <filename>mk</filename>
2242 directory in which the <filename>boilerplate.mk</filename>
2243 file is. It is <emphasis>not</emphasis> OK to simply say</para>
2246 include ../mk/boilerplate.mk # NO NO NO
2250 <para>Why? Because the <filename>boilerplate.mk</filename>
2251 file needs to know where it is, so that it can, in turn,
2252 <literal>include</literal> other files. (Unfortunately,
2253 when an <literal>include</literal>d file does an
2254 <literal>include</literal>, the filename is treated relative
2255 to the directory in which <command>gmake</command> is being
2256 run, not the directory in which the
2257 <literal>include</literal>d sits.) In general,
2258 <emphasis>every file <filename>foo.mk</filename> assumes
2260 <filename><constant>$(TOP)</constant>/mk/foo.mk</filename>
2261 refers to itself.</emphasis> It is up to the
2262 <filename>Makefile</filename> doing the
2263 <literal>include</literal> to ensure this is the case.</para>
2265 <para>Files intended for inclusion in other
2266 <filename>Makefile</filename>s are written to have the
2267 following property: <emphasis>after
2268 <filename>foo.mk</filename> is <literal>include</literal>d,
2269 it leaves <constant>TOP</constant> containing the same value
2270 as it had just before the <literal>include</literal>
2271 statement</emphasis>. In our example, this invariant
2272 guarantees that the <literal>include</literal> for
2273 <filename>target.mk</filename> will look in the same
2274 directory as that for <filename>boilerplate.mk</filename>.</para>
2278 <para> The second section defines the following standard
2279 <command>make</command> variables:
2280 <constant>SRCS</constant><indexterm><primary>SRCS</primary></indexterm>
2281 (the source files from which is to be built), and
2282 <constant>HS_PROG</constant><indexterm><primary>HS_PROG</primary></indexterm>
2283 (the executable binary to be built). We will discuss in
2284 more detail what the “standard variables” are,
2285 and how they affect what happens, in <XRef
2286 LinkEnd="sec-targets">.</para>
2288 <para>The definition for <constant>SRCS</constant> uses the
2289 useful GNU <command>make</command> construct
2290 <literal>$(wildcard $pat$)</literal><indexterm><primary>wildcard</primary></indexterm>,
2291 which expands to a list of all the files matching the
2292 pattern <literal>pat</literal> in the current directory. In
2293 this example, <constant>SRCS</constant> is set to the list
2294 of all the <filename>.lhs</filename> and
2295 <filename>.c</filename> files in the directory. (Let's
2296 suppose there is one of each, <filename>Foo.lhs</filename>
2297 and <filename>Baz.c</filename>.)</para>
2301 <para>The last section includes a second file of standard
2303 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>.
2304 It contains the rules that tell <command>gmake</command> how
2305 to make the standard targets (<Xref
2306 LinkEnd="sec-standard-targets">). Why, you ask, can't this
2307 standard code be part of
2308 <filename>boilerplate.mk</filename>? Good question. We
2309 discuss the reason later, in <Xref
2310 LinkEnd="sec-boiler-arch">.</para>
2312 <para>You do not <emphasis>have</emphasis> to
2313 <literal>include</literal> the
2314 <filename>target.mk</filename> file. Instead, you can write
2315 rules of your own for all the standard targets. Usually,
2316 though, you will find quite a big payoff from using the
2317 canned rules in <filename>target.mk</filename>; the price
2318 tag is that you have to understand what canned rules get
2319 enabled, and what they do (<Xref
2320 LinkEnd="sec-targets">).</para>
2324 <para>In our example <filename>Makefile</filename>, most of the
2325 work is done by the two <literal>include</literal>d files. When
2326 you say <command>gmake all</command>, the following things
2331 <para><command>gmake</command> figures out that the object
2332 files are <filename>Foo.o</filename> and
2333 <filename>Baz.o</filename>.</para>
2337 <para>It uses a boilerplate pattern rule to compile
2338 <filename>Foo.lhs</filename> to <filename>Foo.o</filename>
2339 using a Haskell compiler. (Which one? That is set in the
2340 build configuration.)</para>
2344 <para>It uses another standard pattern rule to compile
2345 <filename>Baz.c</filename> to <filename>Baz.o</filename>,
2346 using a C compiler. (Ditto.)</para>
2350 <para>It links the resulting <filename>.o</filename> files
2351 together to make <literal>small</literal>, using the Haskell
2352 compiler to do the link step. (Why not use
2353 <command>ld</command>? Because the Haskell compiler knows
2354 what standard libraries to link in. How did
2355 <command>gmake</command> know to use the Haskell compiler to
2356 do the link, rather than the C compiler? Because we set the
2357 variable <constant>HS_PROG</constant> rather than
2358 <constant>C_PROG</constant>.)</para>
2362 <para>All <filename>Makefile</filename>s should follow the above
2363 three-section format.</para>
2367 <title>A larger project</title>
2369 <para>Larger projects are usually structured into a number of
2370 sub-directories, each of which has its own
2371 <filename>Makefile</filename>. (In very large projects, this
2372 sub-structure might be iterated recursively, though that is
2373 rare.) To give you the idea, here's part of the directory
2374 structure for the (rather large) GHC project:</para>
2384 ...source files for documentation...
2387 ...source files for driver...
2390 parser/...source files for parser...
2391 renamer/...source files for renamer...
2395 <para>The sub-directories <filename>docs</filename>,
2396 <filename>driver</filename>, <filename>compiler</filename>, and
2397 so on, each contains a sub-component of GHC, and each has its
2398 own <filename>Makefile</filename>. There must also be a
2399 <filename>Makefile</filename> in
2400 <filename><constant>$(FPTOOLS_TOP)</constant>/ghc</filename>.
2401 It does most of its work by recursively invoking
2402 <command>gmake</command> on the <filename>Makefile</filename>s
2403 in the sub-directories. We say that
2404 <filename>ghc/Makefile</filename> is a <emphasis>non-leaf
2405 <filename>Makefile</filename></emphasis>, because it does little
2406 except organise its children, while the
2407 <filename>Makefile</filename>s in the sub-directories are all
2408 <emphasis>leaf <filename>Makefile</filename>s</emphasis>. (In
2409 principle the sub-directories might themselves contain a
2410 non-leaf <filename>Makefile</filename> and several
2411 sub-sub-directories, but that does not happen in GHC.)</para>
2413 <para>The <filename>Makefile</filename> in
2414 <filename>ghc/compiler</filename> is considered a leaf
2415 <filename>Makefile</filename> even though the
2416 <filename>ghc/compiler</filename> has sub-directories, because
2417 these sub-directories do not themselves have
2418 <filename>Makefile</filename>s in them. They are just used to
2419 structure the collection of modules that make up GHC, but all
2420 are managed by the single <filename>Makefile</filename> in
2421 <filename>ghc/compiler</filename>.</para>
2423 <para>You will notice that <filename>ghc/</filename> also
2424 contains a directory <filename>ghc/mk/</filename>. It contains
2425 GHC-specific <filename>Makefile</filename> boilerplate code.
2426 More precisely:</para>
2430 <para><filename>ghc/mk/boilerplate.mk</filename> is included
2431 at the top of <filename>ghc/Makefile</filename>, and of all
2432 the leaf <filename>Makefile</filename>s in the
2433 sub-directories. It in turn <literal>include</literal>s the
2434 main boilerplate file
2435 <filename>mk/boilerplate.mk</filename>.</para>
2439 <para><filename>ghc/mk/target.mk</filename> is
2440 <literal>include</literal>d at the bottom of
2441 <filename>ghc/Makefile</filename>, and of all the leaf
2442 <filename>Makefile</filename>s in the sub-directories. It
2443 in turn <literal>include</literal>s the file
2444 <filename>mk/target.mk</filename>.</para>
2448 <para>So these two files are the place to look for GHC-wide
2449 customisation of the standard boilerplate.</para>
2452 <sect2 id="sec-boiler-arch">
2453 <title>Boilerplate architecture</title>
2454 <indexterm><primary>boilerplate architecture</primary></indexterm>
2456 <para>Every <filename>Makefile</filename> includes a
2457 <filename>boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>
2458 file at the top, and
2459 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>
2460 file at the bottom. In this section we discuss what is in these
2461 files, and why there have to be two of them. In general:</para>
2465 <para><filename>boilerplate.mk</filename> consists of:</para>
2469 <para><emphasis>Definitions of millions of
2470 <command>make</command> variables</emphasis> that
2471 collectively specify the build configuration. Examples:
2472 <constant>HC_OPTS</constant><indexterm><primary>HC_OPTS</primary></indexterm>,
2473 the options to feed to the Haskell compiler;
2474 <constant>NoFibSubDirs</constant><indexterm><primary>NoFibSubDirs</primary></indexterm>,
2475 the sub-directories to enable within the
2476 <literal>nofib</literal> project;
2477 <constant>GhcWithHc</constant><indexterm><primary>GhcWithHc</primary></indexterm>,
2478 the name of the Haskell compiler to use when compiling
2479 GHC in the <literal>ghc</literal> project.</para>
2483 <para><emphasis>Standard pattern rules</emphasis> that
2484 tell <command>gmake</command> how to construct one file
2485 from another.</para>
2489 <para><filename>boilerplate.mk</filename> needs to be
2490 <literal>include</literal>d at the <emphasis>top</emphasis>
2491 of each <filename>Makefile</filename>, so that the user can
2492 replace the boilerplate definitions or pattern rules by
2493 simply giving a new definition or pattern rule in the
2494 <filename>Makefile</filename>. <command>gmake</command>
2495 simply takes the last definition as the definitive one.</para>
2497 <para>Instead of <emphasis>replacing</emphasis> boilerplate
2498 definitions, it is also quite common to
2499 <emphasis>augment</emphasis> them. For example, a
2500 <filename>Makefile</filename> might say:</para>
2506 <para>thereby adding “<option>-O</option>” to
2508 <constant>SRC_HC_OPTS</constant><indexterm><primary>SRC_HC_OPTS</primary></indexterm>.</para>
2512 <para><filename>target.mk</filename> contains
2513 <command>make</command> rules for the standard targets
2514 described in <Xref LinkEnd="sec-standard-targets">. These
2515 rules are selectively included, depending on the setting of
2516 certain <command>make</command> variables. These variables
2517 are usually set in the middle section of the
2518 <filename>Makefile</filename> between the two
2519 <literal>include</literal>s.</para>
2521 <para><filename>target.mk</filename> must be included at the
2522 end (rather than being part of
2523 <filename>boilerplate.mk</filename>) for several tiresome
2529 <para><command>gmake</command> commits target and
2530 dependency lists earlier than it should. For example,
2531 <FIlename>target.mk</FIlename> has a rule that looks
2535 $(HS_PROG) : $(OBJS)
2536 $(HC) $(LD_OPTS) $< -o $@
2539 <para>If this rule was in
2540 <filename>boilerplate.mk</filename> then
2541 <constant>$(HS_PROG)</constant><indexterm><primary>HS_PROG</primary></indexterm>
2543 <constant>$(OBJS)</constant><indexterm><primary>OBJS</primary></indexterm>
2544 would not have their final values at the moment
2545 <command>gmake</command> encountered the rule. Alas,
2546 <command>gmake</command> takes a snapshot of their
2547 current values, and wires that snapshot into the rule.
2548 (In contrast, the commands executed when the rule
2549 “fires” are only substituted at the moment
2550 of firing.) So, the rule must follow the definitions
2551 given in the <filename>Makefile</filename> itself.</para>
2555 <para>Unlike pattern rules, ordinary rules cannot be
2556 overriden or replaced by subsequent rules for the same
2557 target (at least, not without an error message).
2558 Including ordinary rules in
2559 <filename>boilerplate.mk</filename> would prevent the
2560 user from writing rules for specific targets in specific
2565 <para>There are a couple of other reasons I've
2566 forgotten, but it doesn't matter too much.</para>
2573 <sect2 id="sec-boiler">
2574 <title>The main <filename>mk/boilerplate.mk</filename> file</title>
2575 <indexterm><primary>boilerplate.mk</primary></indexterm>
2577 <para>If you look at
2578 <filename><constant>$(FPTOOLS_TOP)</constant>/mk/boilerplate.mk</filename>
2579 you will find that it consists of the following sections, each
2580 held in a separate file:</para>
2584 <term><filename>config.mk</filename></term>
2585 <indexterm><primary>config.mk</primary></indexterm>
2587 <para>is the build configuration file we discussed at
2588 length in <Xref LinkEnd="sec-build-config">.</para>
2593 <term><filename>paths.mk</filename></term>
2594 <indexterm><primary>paths.mk</primary></indexterm>
2596 <para>defines <command>make</command> variables for
2597 pathnames and file lists. This file contains code for
2598 automatically compiling lists of source files and deriving
2599 lists of object files from those. The results can be
2600 overriden in the <filename>Makefile</filename>, but in
2601 most cases the automatic setup should do the right
2604 <para>The following variables may be set in the
2605 <filename>Makefile</filename> to affect how the automatic
2606 source file search is done:</para>
2610 <term><literal>ALL_DIRS</literal></term>
2611 <indexterm><primary><literal>ALL_DIRS</literal></primary>
2614 <para>Set to a list of directories to search in
2615 addition to the current directory for source
2621 <term><literal>EXCLUDE_SRCS</literal></term>
2622 <indexterm><primary><literal>EXCLUDE_SRCS</literal></primary>
2625 <para>Set to a list of source files (relative to the
2626 current directory) to omit from the automatic
2627 search. The source searching machinery is clever
2628 enough to know that if you exclude a source file
2629 from which other sources are derived, then the
2630 derived sources should also be excluded. For
2631 example, if you set <literal>EXCLUDED_SRCS</literal>
2632 to include <filename>Foo.y</filename>, then
2633 <filename>Foo.hs</filename> will also be
2639 <term><literal>EXTRA_SRCS</literal></term>
2640 <indexterm><primary><literal>EXCLUDE_SRCS</literal></primary>
2643 <para>Set to a list of extra source files (perhaps
2644 in directories not listed in
2645 <literal>ALL_DIRS</literal>) that should be
2651 <para>The results of the automatic source file search are
2652 placed in the following make variables:</para>
2656 <term><literal>SRCS</literal></term>
2657 <indexterm><primary><literal>SRCS</literal></primary></indexterm>
2659 <para>All source files found, sorted and without
2660 duplicates, including those which might not exist
2661 yet but will be derived from other existing sources.
2662 <literal>SRCS</literal> <emphasis>can</emphasis> be
2663 overriden if necessary, in which case the variables
2664 below will follow suit.</para>
2669 <term><literal>HS_SRCS</literal></term>
2670 <indexterm><primary><literal>HS_SRCS</literal></primary></indexterm>
2672 <para>all Haskell source files in the current
2673 directory, including those derived from other source
2674 files (eg. Happy sources also give rise to Haskell
2680 <term><literal>HS_OBJS</literal></term>
2681 <indexterm><primary><literal>HS_OBJS</literal></primary></indexterm>
2683 <para>Object files derived from
2684 <literal>HS_SRCS</literal>.</para>
2689 <term><literal>HS_IFACES</literal></term>
2690 <indexterm><primary><literal>HS_IFACES</literal></primary></indexterm>
2692 <para>Interface files (<literal>.hi</literal> files)
2693 derived from <literal>HS_SRCS</literal>.</para>
2698 <term><literal>C_SRCS</literal></term>
2699 <indexterm><primary><literal>C_SRCS</literal></primary></indexterm>
2701 <para>All C source files found.</para>
2706 <term><literal>C_OBJS</literal></term>
2707 <indexterm><primary><literal>C_OBJS</literal></primary></indexterm>
2709 <para>Object files derived from
2710 <literal>C_SRCS</literal>.</para>
2715 <term><literal>SCRIPT_SRCS</literal></term>
2716 <indexterm><primary><literal>SCRIPT_SRCS</literal></primary></indexterm>
2718 <para>All script source files found
2719 (<literal>.lprl</literal> files).</para>
2724 <term><literal>SCRIPT_OBJS</literal></term>
2725 <indexterm><primary><literal>SCRIPT_OBJS</literal></primary></indexterm>
2727 <para><quote>object</quote> files derived from
2728 <literal>SCRIPT_SRCS</literal>
2729 (<literal>.prl</literal> files).</para>
2734 <term><literal>HSC_SRCS</literal></term>
2735 <indexterm><primary><literal>HSC_SRCS</literal></primary></indexterm>
2737 <para>All <literal>hsc2hs</literal> source files
2738 (<literal>.hsc</literal> files).</para>
2743 <term><literal>HAPPY_SRCS</literal></term>
2744 <indexterm><primary><literal>HAPPY_SRCS</literal></primary></indexterm>
2746 <para>All <literal>happy</literal> source files
2747 (<literal>.y</literal> or <literal>.hy</literal> files).</para>
2752 <term><literal>OBJS</literal></term>
2753 <indexterm><primary>OBJS</primary></indexterm>
2755 <para>the concatenation of
2756 <literal>$(HS_OBJS)</literal>,
2757 <literal>$(C_OBJS)</literal>, and
2758 <literal>$(SCRIPT_OBJS)</literal>.</para>
2763 <para>Any or all of these definitions can easily be
2764 overriden by giving new definitions in your
2765 <filename>Makefile</filename>.</para>
2767 <para>What, exactly, does <filename>paths.mk</filename>
2768 consider a <quote>source file</quote> to be? It's based
2769 on the file's suffix (e.g. <filename>.hs</filename>,
2770 <filename>.lhs</filename>, <filename>.c</filename>,
2771 <filename>.hy</filename>, etc), but this is the kind of
2772 detail that changes, so rather than enumerate the source
2773 suffices here the best thing to do is to look in
2774 <filename>paths.mk</filename>.</para>
2779 <term><filename>opts.mk</filename></term>
2780 <indexterm><primary>opts.mk</primary></indexterm>
2782 <para>defines <command>make</command> variables for option
2783 strings to pass to each program. For example, it defines
2784 <constant>HC_OPTS</constant><indexterm><primary>HC_OPTS</primary></indexterm>,
2785 the option strings to pass to the Haskell compiler. See
2786 <Xref LinkEnd="sec-suffix">.</para>
2791 <term><filename>suffix.mk</filename></term>
2792 <indexterm><primary>suffix.mk</primary></indexterm>
2794 <para>defines standard pattern rules—see <Xref
2795 LinkEnd="sec-suffix">.</para>
2800 <para>Any of the variables and pattern rules defined by the
2801 boilerplate file can easily be overridden in any particular
2802 <filename>Makefile</filename>, because the boilerplate
2803 <literal>include</literal> comes first. Definitions after this
2804 <literal>include</literal> directive simply override the default
2805 ones in <filename>boilerplate.mk</filename>.</para>
2808 <sect2 id="sec-suffix">
2809 <title>Pattern rules and options</title>
2810 <indexterm><primary>Pattern rules</primary></indexterm>
2813 <filename>suffix.mk</filename><indexterm><primary>suffix.mk</primary></indexterm>
2814 defines standard <emphasis>pattern rules</emphasis> that say how
2815 to build one kind of file from another, for example, how to
2816 build a <filename>.o</filename> file from a
2817 <filename>.c</filename> file. (GNU <command>make</command>'s
2818 <emphasis>pattern rules</emphasis> are more powerful and easier
2819 to use than Unix <command>make</command>'s <emphasis>suffix
2820 rules</emphasis>.)</para>
2822 <para>Almost all the rules look something like this:</para>
2827 $(CC) $(CC_OPTS) -c $< -o $@
2830 <para>Here's how to understand the rule. It says that
2831 <emphasis>something</emphasis><filename>.o</filename> (say
2832 <filename>Foo.o</filename>) can be built from
2833 <emphasis>something</emphasis><filename>.c</filename>
2834 (<filename>Foo.c</filename>), by invoking the C compiler (path
2835 name held in <constant>$(CC)</constant>), passing to it
2836 the options <constant>$(CC_OPTS)</constant> and
2837 the rule's dependent file of the rule
2838 <literal>$<</literal> (<filename>Foo.c</filename> in
2839 this case), and putting the result in the rule's target
2840 <literal>$@</literal> (<filename>Foo.o</filename> in this
2843 <para>Every program is held in a <command>make</command>
2844 variable defined in <filename>mk/config.mk</filename>—look
2845 in <filename>mk/config.mk</filename> for the complete list. One
2846 important one is the Haskell compiler, which is called
2847 <constant>$(HC)</constant>.</para>
2849 <para>Every program's options are are held in a
2850 <command>make</command> variables called
2851 <constant><prog>_OPTS</constant>. the
2852 <constant><prog>_OPTS</constant> variables are
2853 defined in <filename>mk/opts.mk</filename>. Almost all of them
2854 are defined like this:</para>
2857 CC_OPTS = $(SRC_CC_OPTS) $(WAY$(_way)_CC_OPTS) $($*_CC_OPTS) $(EXTRA_CC_OPTS)
2860 <para>The four variables from which
2861 <constant>CC_OPTS</constant> is built have the following
2866 <term><constant>SRC_CC_OPTS</constant><indexterm><primary>SRC_CC_OPTS</primary></indexterm>:</term>
2868 <para>options passed to all C compilations.</para>
2873 <term><constant>WAY_<way>_CC_OPTS</constant>:</term>
2875 <para>options passed to C compilations for way
2876 <literal><way></literal>. For example,
2877 <constant>WAY_mp_CC_OPTS</constant>
2878 gives options to pass to the C compiler when compiling way
2879 <literal>mp</literal>. The variable
2880 <constant>WAY_CC_OPTS</constant> holds
2881 options to pass to the C compiler when compiling the
2882 standard way. (<Xref LinkEnd="sec-ways"> dicusses
2883 multi-way compilation.)</para>
2888 <term><constant><module>_CC_OPTS</constant>:</term>
2890 <para>options to pass to the C compiler that are specific
2891 to module <literal><module></literal>. For example,
2892 <constant>SMap_CC_OPTS</constant> gives the
2893 specific options to pass to the C compiler when compiling
2894 <filename>SMap.c</filename>.</para>
2899 <term><constant>EXTRA_CC_OPTS</constant><indexterm><primary>EXTRA_CC_OPTS</primary></indexterm>:</term>
2901 <para>extra options to pass to all C compilations. This
2902 is intended for command line use, thus:</para>
2905 gmake libHS.a EXTRA_CC_OPTS="-v"
2912 <sect2 id="sec-targets">
2913 <title>The main <filename>mk/target.mk</filename> file</title>
2914 <indexterm><primary>target.mk</primary></indexterm>
2916 <para><filename>target.mk</filename> contains canned rules for
2917 all the standard targets described in <Xref
2918 LinkEnd="sec-standard-targets">. It is complicated by the fact
2919 that you don't want all of these rules to be active in every
2920 <filename>Makefile</filename>. Rather than have a plethora of
2921 tiny files which you can include selectively, there is a single
2922 file, <filename>target.mk</filename>, which selectively includes
2923 rules based on whether you have defined certain variables in
2924 your <filename>Makefile</filename>. This section explains what
2925 rules you get, what variables control them, and what the rules
2926 do. Hopefully, you will also get enough of an idea of what is
2927 supposed to happen that you can read and understand any weird
2928 special cases yourself.</para>
2932 <term><constant>HS_PROG</constant><indexterm><primary>HS_PROG</primary></indexterm>.</term>
2934 <para>If <constant>HS_PROG</constant> is defined,
2935 you get rules with the following targets:</para>
2939 <term><filename>HS_PROG</filename><indexterm><primary>HS_PROG</primary></indexterm></term>
2941 <para>itself. This rule links
2942 <constant>$(OBJS)</constant> with the Haskell
2943 runtime system to get an executable called
2944 <constant>$(HS_PROG)</constant>.</para>
2949 <term><literal>install</literal><indexterm><primary>install</primary></indexterm></term>
2952 <constant>$(HS_PROG)</constant> in
2953 <constant>$(bindir)</constant>.</para>
2962 <term><constant>C_PROG</constant><indexterm><primary>C_PROG</primary></indexterm></term>
2964 <para>is similar to <constant>HS_PROG</constant>,
2965 except that the link step links
2966 <constant>$(C_OBJS)</constant> with the C
2967 runtime system.</para>
2972 <term><constant>LIBRARY</constant><indexterm><primary>LIBRARY</primary></indexterm></term>
2974 <para>is similar to <constant>HS_PROG</constant>,
2975 except that it links
2976 <constant>$(LIB_OBJS)</constant> to make the
2977 library archive <constant>$(LIBRARY)</constant>,
2978 and <literal>install</literal> installs it in
2979 <constant>$(libdir)</constant>.</para>
2984 <term><constant>LIB_DATA</constant><indexterm><primary>LIB_DATA</primary></indexterm></term>
2986 <para>…</para>
2991 <term><constant>LIB_EXEC</constant><indexterm><primary>LIB_EXEC</primary></indexterm></term>
2993 <para>…</para>
2998 <term><constant>HS_SRCS</constant><indexterm><primary>HS_SRCS</primary></indexterm>, <constant>C_SRCS</constant><indexterm><primary>C_SRCS</primary></indexterm>.</term>
3000 <para>If <constant>HS_SRCS</constant> is defined
3001 and non-empty, a rule for the target
3002 <literal>depend</literal> is included, which generates
3003 dependency information for Haskell programs. Similarly
3004 for <constant>C_SRCS</constant>.</para>
3009 <para>All of these rules are “double-colon” rules,
3013 install :: $(HS_PROG)
3014 ...how to install it...
3017 <para>GNU <command>make</command> treats double-colon rules as
3018 separate entities. If there are several double-colon rules for
3019 the same target it takes each in turn and fires it if its
3020 dependencies say to do so. This means that you can, for
3021 example, define both <constant>HS_PROG</constant> and
3022 <constant>LIBRARY</constant>, which will generate two rules for
3023 <literal>install</literal>. When you type <command>gmake
3024 install</command> both rules will be fired, and both the program
3025 and the library will be installed, just as you wanted.</para>
3028 <sect2 id="sec-subdirs">
3029 <title>Recursion</title>
3030 <indexterm><primary>recursion, in makefiles</primary></indexterm>
3031 <indexterm><primary>Makefile, recursing into subdirectories</primary></indexterm>
3033 <para>In leaf <filename>Makefile</filename>s the variable
3034 <constant>SUBDIRS</constant><indexterm><primary>SUBDIRS</primary></indexterm>
3035 is undefined. In non-leaf <filename>Makefile</filename>s,
3036 <constant>SUBDIRS</constant> is set to the list of
3037 sub-directories that contain subordinate
3038 <filename>Makefile</filename>s. <emphasis>It is up to you to
3039 set <constant>SUBDIRS</constant> in the
3040 <filename>Makefile</filename>.</emphasis> There is no automation
3041 here—<constant>SUBDIRS</constant> is too important to
3044 <para>When <constant>SUBDIRS</constant> is defined,
3045 <filename>target.mk</filename> includes a rather neat rule for
3046 the standard targets (<Xref LinkEnd="sec-standard-targets"> that
3047 simply invokes <command>make</command> recursively in each of
3048 the sub-directories.</para>
3050 <para><emphasis>These recursive invocations are guaranteed to
3051 occur in the order in which the list of directories is specified
3052 in <constant>SUBDIRS</constant>. </emphasis>This guarantee can
3053 be important. For example, when you say <command>gmake
3054 boot</command> it can be important that the recursive invocation
3055 of <command>make boot</command> is done in one sub-directory
3056 (the include files, say) before another (the source files).
3057 Generally, put the most independent sub-directory first, and the
3058 most dependent last.</para>
3061 <sect2 id="sec-ways">
3062 <title>Way management</title>
3063 <indexterm><primary>way management</primary></indexterm>
3065 <para>We sometimes want to build essentially the same system in
3066 several different “ways”. For example, we want to build GHC's
3067 <literal>Prelude</literal> libraries with and without profiling,
3068 so that there is an appropriately-built library archive to link
3069 with when the user compiles his program. It would be possible
3070 to have a completely separate build tree for each such “way”,
3071 but it would be horribly bureaucratic, especially since often
3072 only parts of the build tree need to be constructed in multiple
3076 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>
3077 contains some clever magic to allow you to build several
3078 versions of a system; and to control locally how many versions
3079 are built and how they differ. This section explains the
3082 <para>The files for a particular way are distinguished by
3083 munging the suffix. The <quote>normal way</quote> is always
3084 built, and its files have the standard suffices
3085 <filename>.o</filename>, <filename>.hi</filename>, and so on.
3086 In addition, you can build one or more extra ways, each
3087 distinguished by a <emphasis>way tag</emphasis>. The object
3088 files and interface files for one of these extra ways are
3089 distinguished by their suffix. For example, way
3090 <literal>mp</literal> has files
3091 <filename>.mp_o</filename> and
3092 <filename>.mp_hi</filename>. Library archives have their
3093 way tag the other side of the dot, for boring reasons; thus,
3094 <filename>libHS_mp.a</filename>.</para>
3096 <para>A <command>make</command> variable called
3097 <constant>way</constant> holds the current way tag.
3098 <emphasis><constant>way</constant> is only ever set on the
3099 command line of <command>gmake</command></emphasis> (usually in
3100 a recursive invocation of <command>gmake</command> by the
3101 system). It is never set inside a
3102 <filename>Makefile</filename>. So it is a global constant for
3103 any one invocation of <command>gmake</command>. Two other
3104 <command>make</command> variables,
3105 <constant>way_</constant> and
3106 <constant>_way</constant> are immediately derived from
3107 <constant>$(way)</constant> and never altered. If
3108 <constant>way</constant> is not set, then neither are
3109 <constant>way_</constant> and
3110 <constant>_way</constant>, and the invocation of
3111 <command>make</command> will build the <quote>normal
3112 way</quote>. If <constant>way</constant> is set, then the other
3113 two variables are set in sympathy. For example, if
3114 <constant>$(way)</constant> is “<literal>mp</literal>”,
3115 then <constant>way_</constant> is set to
3116 “<literal>mp_</literal>” and
3117 <constant>_way</constant> is set to
3118 “<literal>_mp</literal>”. These three variables are
3119 then used when constructing file names.</para>
3121 <para>So how does <command>make</command> ever get recursively
3122 invoked with <constant>way</constant> set? There are two ways
3123 in which this happens:</para>
3127 <para>For some (but not all) of the standard targets, when
3128 in a leaf sub-directory, <command>make</command> is
3129 recursively invoked for each way tag in
3130 <constant>$(WAYS)</constant>. You set
3131 <constant>WAYS</constant> in the
3132 <filename>Makefile</filename> to the list of way tags you
3133 want these targets built for. The mechanism here is very
3134 much like the recursive invocation of
3135 <command>make</command> in sub-directories (<Xref
3136 LinkEnd="sec-subdirs">). It is up to you to set
3137 <constant>WAYS</constant> in your
3138 <filename>Makefile</filename>; this is how you control what
3139 ways will get built.</para>
3143 <para>For a useful collection of targets (such as
3144 <filename>libHS_mp.a</filename>,
3145 <filename>Foo.mp_o</filename>) there is a rule which
3146 recursively invokes <command>make</command> to make the
3147 specified target, setting the <constant>way</constant>
3148 variable. So if you say <command>gmake
3149 Foo.mp_o</command> you should see a recursive
3150 invocation <command>gmake Foo.mp_o way=mp</command>,
3151 and <emphasis>in this recursive invocation the pattern rule
3152 for compiling a Haskell file into a <filename>.o</filename>
3153 file will match</emphasis>. The key pattern rules (in
3154 <filename>suffix.mk</filename>) look like this:
3158 $(HC) $(HC_OPTS) $< -o $@
3165 <para>You can invoke <command>make</command> with a
3166 particular <literal>way</literal> setting yourself, in order
3167 to build files related to a particular
3168 <literal>way</literal> in the current directory. eg.
3174 will build files for the profiling way only in the current
3181 <title>When the canned rule isn't right</title>
3183 <para>Sometimes the canned rule just doesn't do the right thing.
3184 For example, in the <literal>nofib</literal> suite we want the
3185 link step to print out timing information. The thing to do here
3186 is <emphasis>not</emphasis> to define
3187 <constant>HS_PROG</constant> or
3188 <constant>C_PROG</constant>, and instead define a special
3189 purpose rule in your own <filename>Makefile</filename>. By
3190 using different variable names you will avoid the canned rules
3191 being included, and conflicting with yours.</para>
3195 <sect1 id="sec-porting-ghc">
3196 <title>Porting GHC</title>
3198 <para>This section describes how to port GHC to a currenly
3199 unsupported platform. There are two distinct
3200 possibilities:</para>
3204 <para>The hardware architecture for your system is already
3205 supported by GHC, but you're running an OS that isn't
3206 supported (or perhaps has been supported in the past, but
3207 currently isn't). This is the easiest type of porting job,
3208 but it still requires some careful bootstrapping. Proceed to
3209 <xref linkend="sec-booting-from-hc">.</para>
3213 <para>Your system's hardware architecture isn't supported by
3214 GHC. This will be a more difficult port (though by comparison
3215 perhaps not as difficult as porting gcc). Proceed to <xref
3216 linkend="unregisterised-porting">.</para>
3220 <sect2 id="sec-booting-from-hc">
3221 <title>Booting/porting from C (<filename>.hc</filename>) files</title>
3223 <indexterm><primary>building GHC from .hc files</primary></indexterm>
3224 <indexterm><primary>booting GHC from .hc files</primary></indexterm>
3225 <indexterm><primary>porting GHC</primary></indexterm>
3227 <para>Bootstrapping GHC on a system without GHC already
3228 installed is achieved by taking the intermediate C files (known
3229 as HC files) from a GHC compilation on a supported system to the
3230 target machine, and compiling them using gcc to get a working
3233 <para><emphasis>NOTE: GHC version 5.xx is significantly harder
3234 to bootstrap from C than previous versions. We recommend
3235 starting from version 4.08.2 if you need to bootstrap in this
3236 way.</emphasis></para>
3238 <para>HC files are architecture-dependent (but not
3239 OS-dependent), so you have to get a set that were generated on
3240 similar hardware. There may be some supplied on the GHC
3241 download page, otherwise you'll have to compile some up
3242 yourself, or start from <emphasis>unregisterised</emphasis> HC
3243 files - see <xref linkend="unregisterised-porting">.</para>
3245 <para>The following steps should result in a working GHC build
3246 with full libraries:</para>
3250 <para>Unpack the HC files on top of a fresh source tree
3251 (make sure the source tree version matches the version of
3252 the HC files <emphasis>exactly</emphasis>!). This will
3253 place matching <filename>.hc</filename> files next to the
3254 corresponding Haskell source (<filename>.hs</filename> or
3255 <filename>.lhs</filename>) in the compiler subdirectory
3256 <filename>ghc/compiler</filename> and in the libraries
3257 (<filename>ghc/lib</filename>, and subdirectories of
3258 <filename>hslibs</filename>).</para>
3262 <para>The actual build process is fully automated by the
3263 <filename>hc-build</filename> script located in the
3264 <filename>distrib</filename> directory. If you eventually
3265 want to install GHC into the directory
3266 <replaceable>dir</replaceable>, the following
3267 command will execute the whole build process (it won't
3268 install yet):</para>
3271 foo% distrib/hc-build --prefix=<replaceable>dir</replaceable>
3273 <indexterm><primary>--hc-build</primary></indexterm>
3275 <para>By default, the installation directory is
3276 <filename>/usr/local</filename>. If that is what you want,
3277 you may omit the argument to <filename>hc-build</filename>.
3278 Generally, any option given to <filename>hc-build</filename>
3279 is passed through to the configuration script
3280 <filename>configure</filename>. If
3281 <filename>hc-build</filename> successfully completes the
3282 build process, you can install the resulting system, as
3292 <sect2 id="unregisterised-porting">
3293 <title>Porting GHC to a new architecture</title>
3295 <para>The first step in porting to a new architecture is to get
3296 an <firstterm>unregisterised</firstterm> build working. An
3297 unregisterised build is one that compiles via vanilla C only.
3298 By contrast, a registerised build uses the following
3299 architecture-specific hacks for speed:</para>
3303 <para>Global register variables: certain abstract machine
3304 <quote>registers</quote> are mapped to real machine
3305 registers, depending on how many machine registers are
3307 <filename>ghc/includes/MachRegs.h</filename>).</para>
3311 <para>Assembly-mangling: when compiling via C, we feed the
3312 assembly generated by gcc though a Perl script known as the
3313 <firstterm>mangler</firstterm> (see
3314 <filename>ghc/driver/mangler/ghc-asm.lprl</filename>). The
3315 mangler rearranges the assembly to support tail-calls and
3316 various other optimisations.</para>
3320 <para>In an unregisterised build, neither of these hacks are
3321 used — the idea is that the C code generated by the
3322 compiler should compile using gcc only. The lack of these
3323 optimisations costs about a factor of two in performance, but
3324 since unregisterised compilation is usually just a step on the
3325 way to a full registerised port, we don't mind too much.</para>
3328 <title>Building an unregisterised port</title>
3330 <para>The first step is to get some unregisterised HC files.
3331 Either (a) download them from the GHC site (if there are
3332 some available for the right version of GHC), or
3333 (b) build them yourself on any machine with a working
3336 <para>There is a script available which should automate the
3337 process of doing the 2-stage bootstrap necessary to get the
3338 unregisterised HC files - it's available in <ulink
3339 url="http://cvs.haskell.org/cgi-bin/cvsweb.cgi/fptools/distrib/cross-port"><filename>fptools/distrib/cross-port</filename></ulink>
3342 <para>Now take these unregisterised HC files to the target
3343 platform and bootstrap a compiler from them as per the
3344 instructions in <xref linkend="sec-booting-from-hc">. In
3345 <filename>build.mk</filename>, you need to tell the build
3346 system that the compiler you're building is
3347 (a) unregisterised itself, and (b) builds
3348 unregisterised binaries. This varies depending on the GHC
3349 version you're bootstraping:</para>
3352 # build.mk for GHC 4.08.x
3353 GhcWithRegisterised=NO
3357 # build.mk for GHC 5.xx
3358 GhcUnregisterised=YES
3361 <para>Version 5.xx only: use the option
3362 <option>--enable-hc-boot-unregisterised</option> instead of
3363 <option>--enable-hc-boot</option> when running
3364 <filename>./configure</filename>.</para>
3366 <para>The build may not go through cleanly. We've tried to
3367 stick to writing portable code in most parts of the compiler,
3368 so it should compile on any POSIXish system with gcc, but in
3369 our experience most systems differ from the standards in one
3370 way or another. Deal with any problems as they arise - if you
3371 get stuck, ask the experts on
3372 <email>glasgow-haskell-users@haskell.org</email>.</para>
3374 <para>Once you have the unregisterised compiler up and
3375 running, you can use it to start a registerised port. The
3376 following sections describe the various parts of the system
3377 that will need architecture-specific tweaks in order to get a
3378 registerised build going.</para>
3380 <para>Lots of useful information about the innards of GHC is
3381 available in the <ulink
3382 url="http://www.cse.unsw.edu.au/~chak/haskell/ghc/comm/">GHC
3383 Commentary</ulink>, which might be helpful if you run into
3384 some code which needs tweaking for your system.</para>
3388 <title>Porting the RTS</title>
3390 <para>The following files need architecture-specific code for a
3391 registerised build:</para>
3395 <term><filename>ghc/includes/MachRegs.h</filename></term>
3396 <indexterm><primary><filename>MachRegs.h</filename></primary>
3399 <para>Defines the STG-register to machine-register
3400 mapping. You need to know your platform's C calling
3401 convention, and which registers are generally available
3402 for mapping to global register variables. There are
3403 plenty of useful comments in this file.</para>
3407 <term><filename>ghc/includes/TailCalls.h</filename></term>
3408 <indexterm><primary><filename>TailCalls.h</filename></primary>
3411 <para>Macros that cooperate with the mangler (see <xref
3412 linkend="sec-mangler">) to make proper tail-calls
3417 <term><filename>ghc/rts/Adjustor.c</filename></term>
3418 <indexterm><primary><filename>Adjustor.c</filename></primary>
3422 <literal>foreign import "wrapper"</literal>
3424 <literal>foreign export dynamic</literal>).
3425 Not essential for getting GHC bootstrapped, so this file
3426 can be deferred until later if necessary.</para>
3430 <term><filename>ghc/rts/StgCRun.c</filename></term>
3431 <indexterm><primary><filename>StgCRun.c</filename></primary>
3434 <para>The little assembly layer between the C world and
3435 the Haskell world. See the comments and code for the
3436 other architectures in this file for pointers.</para>
3440 <term><filename>ghc/rts/MBlock.h</filename></term>
3441 <term><filename>ghc/rts/MBlock.c</filename></term>
3442 <indexterm><primary><filename>MBlock.h</filename></primary>
3444 <indexterm><primary><filename>MBlock.c</filename></primary>
3447 <para>These files are really OS-specific rather than
3448 architecture-specific. In <filename>MBlock.h</filename>
3449 is specified the absolute location at which the RTS
3450 should try to allocate memory on your platform (try to
3451 find an area which doesn't conflict with code or dynamic
3452 libraries). In <filename>Mblock.c</filename> you might
3453 need to tweak the call to <literal>mmap()</literal> for
3460 <sect3 id="sec-mangler">
3461 <title>The mangler</title>
3463 <para>The mangler is an evil Perl-script that rearranges the
3464 assembly code output from gcc to do two main things:</para>
3468 <para>Remove function prologues and epilogues, and all
3469 movement of the C stack pointer. This is to support
3470 tail-calls: every code block in Haskell code ends in an
3471 explicit jump, so we don't want the C-stack overflowing
3472 while we're jumping around between code blocks.</para>
3475 <para>Move the <firstterm>info table</firstterm> for a
3476 closure next to the entry code for that closure. In
3477 unregisterised code, info tables contain a pointer to the
3478 entry code, but in registerised compilation we arrange
3479 that the info table is shoved right up against the entry
3480 code, and addressed backwards from the entry code pointer
3481 (this saves a word in the info table and an extra
3482 indirection when jumping to the closure entry
3487 <para>The mangler is abstracted to a certain extent over some
3488 architecture-specific things such as the particular assembler
3489 directives used to herald symbols. Take a look at the
3490 definitions for other architectures and use these as a
3491 starting point.</para>
3495 <title>The native code generator</title>
3497 <para>The native code generator isn't essential to getting a
3498 registerised build going, but it's a desirable thing to have
3499 because it can cut compilation times in half. The native code
3500 generator is described in some detail in the <ulink
3501 url="http://www.cse.unsw.edu.au/~chak/haskell/ghc/comm/">GHC
3502 commentary</ulink>.</para>
3508 <para>To support GHCi, you need to port the dynamic linker
3509 (<filename>fptools/ghc/rts/Linker.c</filename>). The linker
3510 currently supports the ELF and PEi386 object file formats - if
3511 your platform uses one of these then you probably don't have
3512 to do anything except fiddle with the
3513 <literal>#ifdef</literal>s at the top of
3514 <filename>Linker.c</filename> to tell it about your OS.</para>
3516 <para>If your system uses a different object file format, then
3517 you have to write a linker — good luck!</para>
3523 <sect1 id="sec-build-pitfalls">
3524 <title>Known pitfalls in building Glasgow Haskell
3526 <indexterm><primary>problems, building</primary></indexterm>
3527 <indexterm><primary>pitfalls, in building</primary></indexterm>
3528 <indexterm><primary>building pitfalls</primary></indexterm></title>
3531 WARNINGS about pitfalls and known “problems”:
3540 One difficulty that comes up from time to time is running out of space
3541 in <literal>TMPDIR</literal>. (It is impossible for the configuration stuff to
3542 compensate for the vagaries of different sysadmin approaches to temp
3544 <indexterm><primary>tmp, running out of space in</primary></indexterm>
3546 The quickest way around it is <command>setenv TMPDIR /usr/tmp</command><indexterm><primary>TMPDIR</primary></indexterm> or
3547 even <command>setenv TMPDIR .</command> (or the equivalent incantation with your shell
3550 The best way around it is to say
3553 export TMPDIR=<dir>
3556 in your <filename>build.mk</filename> file.
3557 Then GHC and the other <literal>fptools</literal> programs will use the appropriate directory
3566 In compiling some support-code bits, e.g., in <filename>ghc/rts/gmp</filename> and even
3567 in <filename>ghc/lib</filename>, you may get a few C-compiler warnings. We think these
3575 When compiling via C, you'll sometimes get “warning: assignment from
3576 incompatible pointer type” out of GCC. Harmless.
3583 Similarly, <command>ar</command>chiving warning messages like the following are not
3587 ar: filename GlaIOMonad__1_2s.o truncated to GlaIOMonad_
3588 ar: filename GlaIOMonad__2_2s.o truncated to GlaIOMonad_
3598 In compiling the compiler proper (in <filename>compiler/</filename>), you <emphasis>may</emphasis>
3599 get an “Out of heap space” error message. These can vary with the
3600 vagaries of different systems, it seems. The solution is simple:
3607 If you're compiling with GHC 4.00 or later, then the
3608 <emphasis>maximum</emphasis> heap size must have been reached. This
3609 is somewhat unlikely, since the maximum is set to 64M by default.
3610 Anyway, you can raise it with the
3611 <option>-optCrts-M<size></option> flag (add this flag to
3612 <constant><module>_HC_OPTS</constant>
3613 <command>make</command> variable in the appropriate
3614 <filename>Makefile</filename>).
3621 For GHC < 4.00, add a suitable <option>-H</option> flag to the <filename>Makefile</filename>, as
3630 and try again: <command>gmake</command>. (see <Xref LinkEnd="sec-suffix"> for information about
3631 <constant><module>_HC_OPTS</constant>.)
3633 Alternatively, just cut to the chase:
3637 % make EXTRA_HC_OPTS=-optCrts-M128M
3646 If you try to compile some Haskell, and you get errors from GCC about
3647 lots of things from <filename>/usr/include/math.h</filename>, then your GCC was
3648 mis-installed. <command>fixincludes</command> wasn't run when it should've been.
3650 As <command>fixincludes</command> is now automagically run as part of GCC installation,
3651 this bug also suggests that you have an old GCC.
3659 You <emphasis>may</emphasis> need to re-<command>ranlib</command><indexterm><primary>ranlib</primary></indexterm> your libraries (on Sun4s).
3663 % cd $(libdir)/ghc-x.xx/sparc-sun-sunos4
3664 % foreach i ( `find . -name '*.a' -print` ) # or other-shell equiv...
3666 ? # or, on some machines: ar s $i
3671 We'd be interested to know if this is still necessary.
3679 GHC's sources go through <command>cpp</command> before being compiled, and <command>cpp</command> varies
3680 a bit from one Unix to another. One particular gotcha is macro calls
3685 SLIT("Hello, world")
3689 Some <command>cpp</command>s treat the comma inside the string as separating two macro
3690 arguments, so you get
3694 :731: macro `SLIT' used with too many (2) args
3698 Alas, <command>cpp</command> doesn't tell you the offending file!
3700 Workaround: don't put weird things in string args to <command>cpp</command> macros.
3711 <sect1 id="winbuild"><title>Notes for building under Windows</title>
3714 This section summarises how to get the utilities you need on your
3715 Win95/98/NT/2000 machine to use CVS and build GHC. Similar notes for
3716 installing and running GHC may be found in the user guide. In general,
3717 Win95/Win98 behave the same, and WinNT/Win2k behave the same.
3718 You should read the GHC installation guide sections on Windows (in the user
3719 guide) before continuing to read these notes.
3723 <sect2><title>Before you start</title>
3728 Make sure that the user environment variable
3729 <constant>MAKE_MODE</constant> is set to <literal>UNIX</literal>. If you
3730 don't do this you get very weird messages when you type
3731 <command>make</command>, such as:
3733 /c: /c: No such file or directory
3739 <para>GHC uses the <emphasis>mingw</emphasis> C compiler to
3740 generate code, so you have to install that. Just pick up a mingw bundle at
3741 <ulink url="http://www.mingw.org/">http://www.mingw.org/</ulink>.
3742 We install it in <filename>c:/mingw</filename>.
3748 Install a version of GHC, and put it in your
3749 <constant>PATH</constant> (the installer tells you the path element
3750 you need to add upon completion.)