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 libraries
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>,
446 <literal>hslibs</literal> and <literal>libraries</literal>
447 modules (for a full list of the projects available, see
448 <xref linkend="projects">).</para>
453 <sect2 id="cvs-committing">
454 <title>Committing Changes</title>
456 <para>This is only if you have read-write access to the
457 repository. For anoncvs users, CVS will issue a "read-only
458 repository" error if you try to commit changes.</para>
462 <para>Build the software, if necessary. Unless you're just
463 working on documentation, you'll probably want to build the
464 software in order to test any changes you make.</para>
468 <para>Make changes. Preferably small ones first.</para>
472 <para>Test them. You can see exactly what changes you've
473 made by using the <literal>cvs diff</literal> command:</para>
477 <para>lists all the changes (using the
478 <literal>diff</literal> command) in and below the current
479 directory. In emacs, <literal>C-c C-v =</literal> runs
480 <literal>cvs diff</literal> on the current buffer and shows
481 you the results.</para>
485 <para>Before checking in a change, you need to update your
492 <para>This pulls in any changes that other people have made,
493 and merges them with yours. If there are any conflicts, CVS
494 will tell you, and you'll have to resolve them before you
495 can check your changes in. The documentation describes what
496 to do in the event of a conflict.</para>
498 <para>It's not always necessary to do a full cvs update
499 before checking in a change, since CVS will always tell you
500 if you try to check in a file that someone else has changed.
501 However, you should still update at regular intervals to
502 avoid making changes that don't work in conjuction with
503 changes that someone else made. Keeping an eye on what goes
504 by on the mailing list can help here.</para>
508 <para>When you're happy that your change isn't going to
509 break anything, check it in. For a one-file change:</para>
512 $ cvs commit <replaceable>filename</replaceable>
515 <para>CVS will then pop up an editor for you to enter a
516 "commit message", this is just a short description
517 of what your change does, and will be kept in the history of
520 <para>If you're using emacs, simply load up the file into a
521 buffer and type <literal>C-x C-q</literal>, and emacs will
522 prompt for a commit message and then check in the file for
525 <para>For a multiple-file change, things are a bit
526 trickier. There are several ways to do this, but this is the
527 way I find easiest. First type the commit message into a
528 temporary file. Then either</para>
531 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>file_1</replaceable> .... <replaceable>file_n</replaceable>
534 <para>or, if nothing else has changed in this part of the
538 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>directory</replaceable>
541 <para>where <replaceable>directory</replaceable> is a common
542 parent directory for all your changes, and
543 <replaceable>commit-message</replaceable> is the name of the
544 file containing the commit message.</para>
546 <para>Shortly afterwards, you'll get some mail from the
547 relevant mailing list saying which files changed, and giving
548 the commit message. For a multiple-file change, you should
549 still get only <emphasis>one</emphasis> message.</para>
554 <sect2 id="cvs-update">
555 <title>Updating Your Source Tree</title>
557 <para>It can be tempting to cvs update just part of a source
558 tree to bring in some changes that someone else has made, or
559 before committing your own changes. This is NOT RECOMMENDED!
560 Quite often changes in one part of the tree are dependent on
561 changes in another part of the tree (the
562 <literal>mk/*.mk</literal> files are a good example where
563 problems crop up quite often). Having an inconsistent tree is a
564 major cause of headaches. </para>
566 <para>So, to avoid a lot of hassle, follow this recipe for
567 updating your tree: </para>
571 $ cvs update -Pd 2>&1 | tee log</screen>
573 <para>Look at the log file, and fix any conflicts (denoted by a
574 <quote>C</quote> in the first column). If you're using multiple
575 build trees, then for every build tree you have pointing at this
576 source tree, you need to update the links in case any new files
577 have appeared: </para>
580 $ cd <replaceable>build-tree</replaceable>
581 $ lndir <replaceable>source-tree</replaceable>
584 <para>Some files might have been removed, so you need to remove
585 the links pointing to these non-existent files:</para>
588 $ find . -xtype l -exec rm '{}' \;
591 <para>To be <emphasis>really</emphasis> safe, you should do
594 <screen>$ gmake all</screen>
596 <para>from the top-level, to update the dependencies and build
597 any changed files. </para>
600 <sect2 id="cvs-tags">
601 <title>GHC Tag Policy</title>
603 <para>If you want to check out a particular version of GHC,
604 you'll need to know how we tag versions in the repository. The
605 policy (as of 4.04) is:</para>
609 <para>The tree is branched before every major release. The
610 branch tag is <literal>ghc-x-xx-branch</literal>, where
611 <literal>x-xx</literal> is the version number of the release
612 with the <literal>'.'</literal> replaced by a
613 <literal>'-'</literal>. For example, the 4.04 release lives
614 on <literal>ghc-4-04-branch</literal>.</para>
618 <para>The release itself is tagged with
619 <literal>ghc-x-xx</literal> (on the branch). eg. 4.06 is
620 called <literal>ghc-4-06</literal>.</para>
624 <para>We didn't always follow these guidelines, so to see
625 what tags there are for previous versions, do <literal>cvs
626 log</literal> on a file that's been around for a while (like
627 <literal>fptools/ghc/README</literal>).</para>
631 <para>So, to check out a fresh GHC 4.06 tree you would
635 $ cvs co -r ghc-4-06 fpconfig
637 $ cvs co -r ghc-4-06 ghc hslibs
641 <sect2 id="cvs-hints">
642 <title>General Hints</title>
646 <para>As a general rule: commit changes in small units,
647 preferably addressing one issue or implementing a single
648 feature. Provide a descriptive log message so that the
649 repository records exactly which changes were required to
650 implement a given feature/fix a bug. I've found this
651 <emphasis>very</emphasis> useful in the past for finding out
652 when a particular bug was introduced: you can just wind back
653 the CVS tree until the bug disappears.</para>
657 <para>Keep the sources at least *buildable* at any given
658 time. No doubt bugs will creep in, but it's quite easy to
659 ensure that any change made at least leaves the tree in a
660 buildable state. We do nightly builds of GHC to keep an eye
661 on what things work/don't work each day and how we're doing
662 in relation to previous verions. This idea is truely wrecked
663 if the compiler won't build in the first place!</para>
667 <para>To check out extra bits into an already-checked-out
668 tree, use the following procedure. Suppose you have a
669 checked-out fptools tree containing just ghc, and you want
670 to add nofib to it:</para>
681 $ cvs update -d nofib
684 <para>(the -d flag tells update to create a new
685 directory). If you just want part of the nofib suite, you
690 $ cvs checkout nofib/spectral
693 <para>This works because <literal>nofib</literal> is a
694 module in its own right, and spectral is a subdirectory of
695 the nofib module. The path argument to checkout must always
696 start with a module name. There's no equivalent form of this
697 command using <literal>update</literal>.</para>
703 <sect1 id="projects">
704 <title>What projects are there?</title>
706 <para>The <literal>fptools</literal> suite consists of several
707 <firstterm>projects</firstterm>, most of which can be downloaded,
708 built and installed individually. Each project corresponds to a
709 subdirectory in the source tree, and if checking out from CVS then
710 each project can be checked out individually by sitting in the top
711 level of your source tree and typing <command>cvs checkout
712 <replaceable>project</replaceable></command>.</para>
714 <para>Here is a list of the projects currently available:</para>
718 <term><literal>ghc</literal></term>
719 <indexterm><primary><literal>ghc</literal></primary>
720 <secondary>project</secondary></indexterm>
722 <para>The <ulink url="http://www.haskell.org/ghc/">Glasgow
723 Haskell Compiler</ulink> (minus libraries). Absolutely
724 required for building GHC.</para>
729 <term><literal>glafp-utils</literal></term>
730 <indexterm><primary><literal>glafp-utils</literal></primary><secondary>project</secondary></indexterm>
732 <para>Utility programs, some of which are used by the
733 build/installation system. Required for pretty much
739 <term><literal>green-card</literal></term>
740 <indexterm><primary><literal>green-card</literal></primary><secondary>project</secondary></indexterm>
743 url="http://www.haskell.org/greencard/">Green Card</ulink>
744 system for generating Haskell foreign function
750 <term><literal>haggis</literal></term>
751 <indexterm><primary><literal>haggis</literal></primary><secondary>project</secondary></indexterm>
754 url="http://www.dcs.gla.ac.uk/fp/software/haggis/">Haggis</ulink>
755 Haskell GUI framework.</para>
760 <term><literal>haddock</literal></term>
761 <indexterm><primary><literal>haddock</literal></primary><secondary>project</secondary></indexterm>
764 url="http://www.haskell.org/haddock/">Haddock</ulink>
765 documentation tool.</para>
770 <term><literal>happy</literal></term>
771 <indexterm><primary><literal>happy</literal></primary><secondary>project</secondary></indexterm>
774 url="http://www.haskell.org/happy/">Happy</ulink> Parser
780 <term><literal>hdirect</literal></term>
781 <indexterm><primary><literal>hdirect</literal></primary><secondary>project</secondary></indexterm>
784 url="http://www.haskell.org/hdirect/">H/Direct</ulink>
785 Haskell interoperability tool.</para>
790 <term><literal>hood</literal></term>
791 <indexterm><primary><literal>hood</literal></primary><secondary>project</secondary></indexterm>
793 <para>The <ulink url="http://www.haskell.org/hood/">Haskell
794 Object Observation Debugger</ulink>.</para>
799 <term><literal>hslibs</literal></term>
800 <indexterm><primary><literal>hslibs</literal></primary><secondary>project</secondary></indexterm>
802 <para>Supplemental libraries for GHC
803 (<emphasis>required</emphasis> for building GHC).</para>
808 <term><literal>libraries</literal></term>
809 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
811 <para>Hierarchical Haskell library suite
812 (<emphasis>required</emphasis> for building GHC).</para>
817 <term><literal>mhms</literal></term>
818 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
820 <para>The Modular Haskell Metric System.</para>
825 <term><literal>nofib</literal></term>
826 <indexterm><primary><literal>nofib</literal></primary><secondary>project</secondary></indexterm>
828 <para>The NoFib suite: A collection of Haskell programs used
829 primarily for benchmarking.</para>
834 <term><literal>testsuite</literal></term>
835 <indexterm><primary><literal>testsuite</literal></primary><secondary>project</secondary></indexterm>
837 <para>A testing framework, including GHC's regression test
843 <para>So, to build GHC you need at least the
844 <literal>ghc</literal>, <literal>libraries</literal> and
845 <literal>hslibs</literal> projects (a GHC source distribution will
846 already include the bits you need).</para>
849 <sect1 id="sec-build-checks">
850 <title>Things to check before you start</title>
852 <para>Here's a list of things to check before you get
858 <indexterm><primary>Disk space needed</primary></indexterm>
859 <para>Disk space needed: from about 100Mb for a basic GHC
860 build, up to probably 500Mb for a GHC build with everything
861 included (libraries built several different ways,
866 <para>Use an appropriate machine / operating system. <xref
867 linkend="sec-port-info"> lists the supported platforms; if
868 yours isn't amongst these then you can try porting GHC (see
869 <xref linkend="sec-porting-ghc">).</para>
873 <para>Be sure that the “pre-supposed” utilities are
874 installed. <Xref LinkEnd="sec-pre-supposed">
879 <para>If you have any problem when building or installing the
880 Glasgow tools, please check the “known pitfalls” (<Xref
881 LinkEnd="sec-build-pitfalls">). Also check the FAQ for the
882 version you're building, which is part of the User's Guide and
883 available on the <ulink URL="http://www.haskell.org/ghc/" >GHC web
886 <indexterm><primary>bugs</primary><secondary>known</secondary></indexterm>
888 <para>If you feel there is still some shortcoming in our
889 procedure or instructions, please report it.</para>
891 <para>For GHC, please see the <ulink
892 url="http://www.haskell.org/ghc/docs/latest/set/bug-reporting.html">bug-reporting
893 section of the GHC Users' Guide</ulink>, to maximise the
894 usefulness of your report.</para>
896 <indexterm><primary>bugs</primary><secondary>seporting</secondary></indexterm>
897 <para>If in doubt, please send a message to
898 <email>glasgow-haskell-bugs@haskell.org</email>.
899 <indexterm><primary>bugs</primary><secondary>mailing
900 list</secondary></indexterm></para>
905 <sect1 id="sec-port-info">
906 <title>What machines the Glasgow tools run on</title>
908 <indexterm><primary>ports</primary><secondary>GHC</secondary></indexterm>
909 <indexterm><primary>GHC</primary><secondary>ports</secondary></indexterm>
910 <indexterm><primary>platforms</primary><secondary>supported</secondary></indexterm>
912 <para>The main question is whether or not the Haskell compiler
913 (GHC) runs on your platform.</para>
915 <para>A “platform” is a
916 architecture/manufacturer/operating-system combination, such as
917 <literal>sparc-sun-solaris2</literal>. Other common ones are
918 <literal>alpha-dec-osf2</literal>,
919 <literal>hppa1.1-hp-hpux9</literal>,
920 <literal>i386-unknown-linux</literal>,
921 <literal>i386-unknown-solaris2</literal>,
922 <literal>i386-unknown-freebsd</literal>,
923 <literal>i386-unknown-cygwin32</literal>,
924 <literal>m68k-sun-sunos4</literal>,
925 <literal>mips-sgi-irix5</literal>,
926 <literal>sparc-sun-sunos4</literal>,
927 <literal>sparc-sun-solaris2</literal>,
928 <literal>powerpc-ibm-aix</literal>.</para>
930 <para>Some libraries may only work on a limited number of
931 platforms; for example, a sockets library is of no use unless the
932 operating system supports the underlying BSDisms.</para>
935 <title>What platforms the Haskell compiler (GHC) runs on</title>
937 <indexterm><primary>fully-supported platforms</primary></indexterm>
938 <indexterm><primary>native-code generator</primary></indexterm>
939 <indexterm><primary>registerised ports</primary></indexterm>
940 <indexterm><primary>unregisterised ports</primary></indexterm>
942 <para>The GHC hierarchy of Porting Goodness: (a) Best is a
943 native-code generator; (b) next best is a
944 “registerised” port; (c) the bare minimum is an
945 “unregisterised” port.
946 (“Unregisterised” is so terrible that we won't say
947 more about it).</para>
949 <para>We use Sparcs running Solaris 2.7 and x86 boxes running
950 FreeBSD and Linux, so those are the best supported platforms,
951 unsurprisingly.</para>
953 <para>Here's everything that's known about GHC ports. We
954 identify platforms by their “canonical”
955 CPU/Manufacturer/OS triple.</para>
959 <term>alpha-dec-{osf,linux,freebsd,openbsd,netbsd}:</term>
960 <indexterm><primary>alpha-dec-osf</primary></indexterm>
961 <indexterm><primary>alpha-dec-linux</primary></indexterm>
962 <indexterm><primary>alpha-dec-freebsd</primary></indexterm>
963 <indexterm><primary>alpha-dec-openbsd</primary></indexterm>
964 <indexterm><primary>alpha-dec-netbsd</primary></indexterm>
967 <para>The OSF port is currently working (as of GHC version
968 5.02.1) and well supported. The native code generator is
969 currently non-working. Other operating systems will
970 require some minor porting.</para>
975 <term>sparc-sun-sunos4</term>
976 <indexterm><primary>sparc-sun-sunos4</primary></indexterm>
978 <para>Probably works with minor tweaks, hasn't been tested
984 <term>sparc-sun-solaris2</term>
985 <indexterm><primary>sparc-sun-solaris2</primary></indexterm>
987 <para>Fully supported (at least for Solaris 2.7),
988 including native-code generator.</para>
993 <term>hppa1.1-hp-hpux (HP-PA boxes running HPUX 9.x)</term>
994 <indexterm><primary>hppa1.1-hp-hpux</primary></indexterm>
996 <para>A registerised port is available for version 4.08,
997 but GHC hasn't been built on that platform since (as far
998 as we know). No native-code generator.</para>
1003 <term>i386-unknown-linux (PCs running Linux, ELF binary format)</term>
1004 <indexterm><primary>i386-*-linux</primary></indexterm>
1006 <para>GHC works registerised and has a native code
1007 generator. You <Emphasis>must</Emphasis> have GCC 2.7.x
1008 or later. NOTE about <literal>glibc</literal> versions:
1009 GHC binaries built on a system running <literal>glibc
1010 2.0</literal> won't work on a system running
1011 <literal>glibc 2.1</literal>, and vice versa. In general,
1012 don't expect compatibility between
1013 <literal>glibc</literal> versions, even if the shared
1014 library version hasn't changed.</para>
1019 <term>i386-unknown-freebsd (PCs running FreeBSD 2.2 or
1021 <indexterm><primary>i386-unknown-freebsd</primary></indexterm>
1023 <para>GHC works registerised. Pre-built packages are
1024 available in the native package format, so if you just
1025 need binaries you're better off just installing the
1026 package (it might even be on your installation
1032 <term>i386-unknown-openbsd (PCs running OpenBSD)</term>
1033 <indexterm><primary>i386-unknown-openbsd</primary></indexterm>
1035 <para>Supported, with native code generator. Packages are
1036 available through the ports system in the native package
1042 <term>i386-unknown-netbsd (PCs running NetBSD and
1044 <indexterm><primary>i386-unknown-netbsd</primary></indexterm>
1046 <para>Will require some minor porting effort, but should
1047 work registerised.</para>
1052 <term>i386-unknown-mingw32 (PCs running Windows)</term>
1053 <indexterm><primary>i386-unknown-mingw32</primary></indexterm>
1055 <para>Fully supported under Win9x, WinNT, Win2k, and
1056 WinXP. Includes a native code generator. Building from
1057 source requires a recent <ulink
1058 url="http://www.cygwin.com/">Cygwin</ulink> distribution
1059 to be installed.</para>
1064 <term>ia64-unknown-linux</term>
1065 <indexterm><primary>ia64-unknown-linux</primary></indexterm>
1067 <para>GHC currently works unregisterised. A registerised
1068 port is in progress.</para>
1073 <term>mips-sgi-irix5</term>
1074 <indexterm><primary>mips-sgi-irix[5-6]</primary></indexterm>
1076 <para>Port has worked in the past, but hasn't been tested
1077 for some time (and will certainly have rotted in various
1078 ways). As usual, we don't have access to machines and
1079 there hasn't been an overwhelming demand for this port,
1080 but feel free to get in touch.</para>
1085 <term>powerpc-ibm-aix</term>
1086 <indexterm><primary>powerpc-ibm-aix</primary></indexterm>
1088 <para>Port currently doesn't work, needs some minimal
1089 porting effort. As usual, we don't have access to
1090 machines and there hasn't been an overwhelming demand for
1091 this port, but feel free to get in touch.</para>
1096 <term>powerpc-apple-darwin</term>
1097 <indexterm><primary>powerpc-apple-darwin</primary></indexterm>
1099 <para>Supported registerised. No native code
1105 <term>powerpc-apple-linux</term>
1106 <indexterm><primary>powerpc-apple-linux</primary></indexterm>
1108 <para>Not supported (yet).</para>
1113 <para>Various other systems have had GHC ported to them in the
1114 distant past, including various Motorola 68k boxes. The 68k
1115 support still remains, but porting to one of these systems will
1116 certainly be a non-trivial task.</para>
1120 <title>What machines the other tools run on</title>
1122 <para>Unless you hear otherwise, the other tools work if GHC
1128 <sect1 id="sec-pre-supposed">
1129 <title>Installing pre-supposed utilities</title>
1131 <indexterm><primary>pre-supposed utilities</primary></indexterm>
1132 <indexterm><primary>utilities, pre-supposed</primary></indexterm>
1134 <para>Here are the gory details about some utility programs you
1135 may need; <command>perl</command>, <command>gcc</command> and
1136 <command>happy</command> are the only important
1137 ones. (PVM<indexterm><primary>PVM</primary></indexterm> is
1138 important if you're going for Parallel Haskell.) The
1139 <command>configure</command><indexterm><primary>configure</primary></indexterm>
1140 script will tell you if you are missing something.</para>
1146 <indexterm><primary>pre-supposed: GHC</primary></indexterm>
1147 <indexterm><primary>GHC, pre-supposed</primary></indexterm>
1149 <para>GHC is required to build many of the tools, including
1150 GHC itself. If you need to port GHC to your platform
1151 because there isn't a binary distribution of GHC available,
1152 then see <xref linkend="sec-porting-ghc">.</para>
1154 <para>Which version of GHC you need will depend on the
1155 packages you intend to build. GHC itself will normally
1156 build using one of several older versions of itself - check
1157 the announcement or release notes for details.</para>
1163 <indexterm><primary>pre-supposed: Perl</primary></indexterm>
1164 <indexterm><primary>Perl, pre-supposed</primary></indexterm>
1166 <para><emphasis>You have to have Perl to proceed!</emphasis>
1167 Perl version 5 at least is required. GHC has been known to
1168 tickle bugs in Perl, so if you find that Perl crashes when
1169 running GHC try updating (or downgrading) your Perl
1170 installation. Versions of Perl that we use and are known to
1171 be fairly stable are 5.005 and 5.6.1.</para>
1173 <para>For Win32 platforms, you should use the binary
1174 supplied in the InstallShield (copy it to
1175 <filename>/bin</filename>). The Cygwin-supplied Perl seems
1178 <para>Perl should be put somewhere so that it can be invoked
1179 by the <literal>#!</literal> script-invoking
1180 mechanism. The full pathname may need to be less than 32
1181 characters long on some systems.</para>
1186 <term>GNU C (<command>gcc</command>)</term>
1187 <indexterm><primary>pre-supposed: GCC (GNU C
1188 compiler)</primary></indexterm> <indexterm><primary>GCC (GNU C
1189 compiler), pre-supposed</primary></indexterm>
1191 <para>We recommend using GCC version 2.95.2 on all
1192 platforms. Failing that, version 2.7.2 is stable on most
1193 platforms. Earlier versions of GCC can be assumed not to
1194 work, and versions in between 2.7.2 and 2.95.2 (including
1195 <command>egcs</command>) have varying degrees of stability
1196 depending on the platform.</para>
1198 <para>If your GCC dies with “internal error” on
1199 some GHC source file, please let us know, so we can report
1200 it and get things improved. (Exception: on iX86
1201 boxes—you may need to fiddle with GHC's
1202 <option>-monly-N-regs</option> option; see the User's
1208 <term>GNU Make</term>
1209 <indexterm><primary>make</primary><secondary>GNU</secondary>
1212 <para>The fptools build system makes heavy use of features
1213 specific to GNU <command>make</command>, so you must have
1214 this installed in order to build any of the fptools
1221 <indexterm><primary>Happy</primary></indexterm>
1223 <para>Happy is a parser generator tool for Haskell, and is
1224 used to generate GHC's parsers. Happy is written in
1225 Haskell, and is a project in the CVS repository
1226 (<literal>fptools/happy</literal>). It can be built from
1227 source, but bear in mind that you'll need GHC installed in
1228 order to build it. To avoid the chicken/egg problem,
1229 install a binary distribtion of either Happy or GHC to get
1230 started. Happy distributions are available from <ulink
1231 url="http://www.haskell.org/happy/">Happy's Web
1232 Page</ulink>.</para>
1237 <term>Autoconf</term>
1238 <indexterm><primary>pre-supposed: Autoconf</primary></indexterm>
1239 <indexterm><primary>Autoconf, pre-supposed</primary></indexterm>
1241 <para>GNU Autoconf is needed if you intend to build from the
1242 CVS sources, it is <emphasis>not</emphasis> needed if you
1243 just intend to build a standard source distribution.</para>
1245 <para>Autoconf builds the <command>configure</command>
1246 script from <filename>configure.in</filename> and
1247 <filename>aclocal.m4</filename>. If you modify either of
1248 these files, you'll need <command>autoconf</command> to
1249 rebuild <filename>configure</filename>.</para>
1254 <term><command>sed</command></term>
1255 <indexterm><primary>pre-supposed: sed</primary></indexterm>
1256 <indexterm><primary>sed, pre-supposed</primary></indexterm>
1258 <para>You need a working <command>sed</command> if you are
1259 going to build from sources. The build-configuration stuff
1260 needs it. GNU sed version 2.0.4 is no good! It has a bug
1261 in it that is tickled by the build-configuration. 2.0.5 is
1262 OK. Others are probably OK too (assuming we don't create too
1263 elaborate configure scripts.)</para>
1268 <para>One <literal>fptools</literal> project is worth a quick note
1269 at this point, because it is useful for all the others:
1270 <literal>glafp-utils</literal> contains several utilities which
1271 aren't particularly Glasgow-ish, but Occasionally Indispensable.
1272 Like <command>lndir</command> for creating symbolic link
1275 <sect2 id="pre-supposed-gph-tools">
1276 <title>Tools for building parallel GHC (GPH)</title>
1280 <term>PVM version 3:</term>
1281 <indexterm><primary>pre-supposed: PVM3 (Parallel Virtual Machine)</primary></indexterm>
1282 <indexterm><primary>PVM3 (Parallel Virtual Machine), pre-supposed</primary></indexterm>
1284 <para>PVM is the Parallel Virtual Machine on which
1285 Parallel Haskell programs run. (You only need this if you
1286 plan to run Parallel Haskell. Concurent Haskell, which
1287 runs concurrent threads on a uniprocessor doesn't need
1288 it.) Underneath PVM, you can have (for example) a network
1289 of workstations (slow) or a multiprocessor box
1292 <para>The current version of PVM is 3.3.11; we use 3.3.7.
1293 It is readily available on the net; I think I got it from
1294 <literal>research.att.com</literal>, in
1295 <filename>netlib</filename>.</para>
1297 <para>A PVM installation is slightly quirky, but easy to
1298 do. Just follow the <filename>Readme</filename>
1299 instructions.</para>
1304 <term><command>bash</command>:</term>
1305 <indexterm><primary>bash, presupposed (Parallel Haskell only)</primary></indexterm>
1307 <para>Sadly, the <command>gr2ps</command> script, used to
1308 convert “parallelism profiles” to PostScript,
1309 is written in Bash (GNU's Bourne Again shell). This bug
1310 will be fixed (someday).</para>
1316 <sect2 id="pre-supposed-other-tools">
1317 <title>Other useful tools</title>
1322 <indexterm><primary>pre-supposed: flex</primary></indexterm>
1323 <indexterm><primary>flex, pre-supposed</primary></indexterm>
1325 <para>This is a quite-a-bit-better-than-Lex lexer. Used
1326 to build a couple of utilities in
1327 <literal>glafp-utils</literal>. Depending on your
1328 operating system, the supplied <command>lex</command> may
1329 or may not work; you should get the GNU version.</para>
1334 <para>More tools are required if you want to format the documentation
1335 that comes with GHC and other fptools projects. See <xref
1336 linkend="building-docs">.</para>
1340 <sect1 id="sec-building-from-source">
1341 <title>Building from source</title>
1343 <indexterm><primary>Building from source</primary></indexterm>
1344 <indexterm><primary>Source, building from</primary></indexterm>
1346 <para>You've been rash enough to want to build some of the Glasgow
1347 Functional Programming tools (GHC, Happy, nofib, etc.) from
1348 source. You've slurped the source, from the CVS repository or
1349 from a source distribution, and now you're sitting looking at a
1350 huge mound of bits, wondering what to do next.</para>
1352 <para>Gingerly, you type <command>make</command>. Wrong
1355 <para>This rest of this guide is intended for duffers like me, who
1356 aren't really interested in Makefiles and systems configurations,
1357 but who need a mental model of the interlocking pieces so that
1358 they can make them work, extend them consistently when adding new
1359 software, and lay hands on them gently when they don't
1362 <sect2 id="quick-start">
1363 <title>Quick Start</title>
1365 <para>If you are starting from a source distribution, and just
1366 want a completely standard build, then the following should
1369 <screen>$ ./configure
1374 <para>For GHC, this will do a 2-stage bootstrap build of the
1375 compiler, with profiling libraries, and install the
1378 <para>If you want to do anything at all non-standard, or you
1379 want to do some development, read on...</para>
1382 <sect2 id="sec-source-tree">
1383 <title>Your source tree</title>
1385 <para>The source code is held in your <emphasis>source
1386 tree</emphasis>. The root directory of your source tree
1387 <emphasis>must</emphasis> contain the following directories and
1392 <para><filename>Makefile</filename>: the root
1397 <para><filename>mk/</filename>: the directory that contains
1398 the main Makefile code, shared by all the
1399 <literal>fptools</literal> software.</para>
1403 <para><filename>configure.in</filename>,
1404 <filename>config.sub</filename>,
1405 <filename>config.guess</filename>: these files support the
1406 configuration process.</para>
1410 <para><filename>install-sh</filename>.</para>
1414 <para>All the other directories are individual
1415 <emphasis>projects</emphasis> of the <literal>fptools</literal>
1416 system—for example, the Glasgow Haskell Compiler
1417 (<literal>ghc</literal>), the Happy parser generator
1418 (<literal>happy</literal>), the <literal>nofib</literal>
1419 benchmark suite, and so on. You can have zero or more of these.
1420 Needless to say, some of them are needed to build others.</para>
1422 <para>The important thing to remember is that even if you want
1423 only one project (<literal>happy</literal>, say), you must have
1424 a source tree whose root directory contains
1425 <filename>Makefile</filename>, <filename>mk/</filename>,
1426 <filename>configure.in</filename>, and the project(s) you want
1427 (<filename>happy/</filename> in this case). You cannot get by
1428 with just the <filename>happy/</filename> directory.</para>
1432 <title>Build trees</title>
1433 <indexterm><primary>build trees</primary></indexterm>
1434 <indexterm><primary>link trees, for building</primary></indexterm>
1436 <para>If you just want to build the software once on a single
1437 platform, then your source tree can also be your build tree, and
1438 you can skip the rest of this section.</para>
1440 <para>We often want to build multiple versions of our software
1441 for different architectures, or with different options
1442 (e.g. profiling). It's very desirable to share a single copy of
1443 the source code among all these builds.</para>
1445 <para>So for every source tree we have zero or more
1446 <emphasis>build trees</emphasis>. Each build tree is initially
1447 an exact copy of the source tree, except that each file is a
1448 symbolic link to the source file, rather than being a copy of
1449 the source file. There are “standard” Unix
1450 utilities that make such copies, so standard that they go by
1452 <command>lndir</command><indexterm><primary>lndir</primary></indexterm>,
1453 <command>mkshadowdir</command><indexterm><primary>mkshadowdir</primary></indexterm>
1454 are two (If you don't have either, the source distribution
1455 includes sources for the X11
1456 <command>lndir</command>—check out
1457 <filename>fptools/glafp-utils/lndir</filename>). See <Xref
1458 LinkEnd="sec-storysofar"> for a typical invocation.</para>
1460 <para>The build tree does not need to be anywhere near the
1461 source tree in the file system. Indeed, one advantage of
1462 separating the build tree from the source is that the build tree
1463 can be placed in a non-backed-up partition, saving your systems
1464 support people from backing up untold megabytes of
1465 easily-regenerated, and rapidly-changing, gubbins. The golden
1466 rule is that (with a single exception—<XRef
1467 LinkEnd="sec-build-config">) <emphasis>absolutely everything in
1468 the build tree is either a symbolic link to the source tree, or
1469 else is mechanically generated</emphasis>. It should be
1470 perfectly OK for your build tree to vanish overnight; an hour or
1471 two compiling and you're on the road again.</para>
1473 <para>You need to be a bit careful, though, that any new files
1474 you create (if you do any development work) are in the source
1475 tree, not a build tree!</para>
1477 <para>Remember, that the source files in the build tree are
1478 <emphasis>symbolic links</emphasis> to the files in the source
1479 tree. (The build tree soon accumulates lots of built files like
1480 <filename>Foo.o</filename>, as well.) You can
1481 <emphasis>delete</emphasis> a source file from the build tree
1482 without affecting the source tree (though it's an odd thing to
1483 do). On the other hand, if you <emphasis>edit</emphasis> a
1484 source file from the build tree, you'll edit the source-tree
1485 file directly. (You can set up Emacs so that if you edit a
1486 source file from the build tree, Emacs will silently create an
1487 edited copy of the source file in the build tree, leaving the
1488 source file unchanged; but the danger is that you think you've
1489 edited the source file whereas actually all you've done is edit
1490 the build-tree copy. More commonly you do want to edit the
1491 source file.)</para>
1493 <para>Like the source tree, the top level of your build tree
1494 must be (a linked copy of) the root directory of the
1495 <literal>fptools</literal> suite. Inside Makefiles, the root of
1496 your build tree is called
1497 <constant>$(FPTOOLS_TOP)</constant><indexterm><primary>FPTOOLS_TOP</primary></indexterm>.
1498 In the rest of this document path names are relative to
1499 <constant>$(FPTOOLS_TOP)</constant> unless
1500 otherwise stated. For example, the file
1501 <filename>ghc/mk/target.mk</filename> is actually
1502 <filename><constant>$(FPTOOLS_TOP)</constant>/ghc/mk/target.mk</filename>.</para>
1505 <sect2 id="sec-build-config">
1506 <title>Getting the build you want</title>
1508 <para>When you build <literal>fptools</literal> you will be
1509 compiling code on a particular <emphasis>host
1510 platform</emphasis>, to run on a particular <emphasis>target
1511 platform</emphasis> (usually the same as the host
1512 platform)<indexterm><primary>platform</primary></indexterm>.
1513 The difficulty is that there are minor differences between
1514 different platforms; minor, but enough that the code needs to be
1515 a bit different for each. There are some big differences too:
1516 for a different architecture we need to build GHC with a
1517 different native-code generator.</para>
1519 <para>There are also knobs you can turn to control how the
1520 <literal>fptools</literal> software is built. For example, you
1521 might want to build GHC optimised (so that it runs fast) or
1522 unoptimised (so that you can compile it fast after you've
1523 modified it. Or, you might want to compile it with debugging on
1524 (so that extra consistency-checking code gets included) or off.
1527 <para>All of this stuff is called the
1528 <emphasis>configuration</emphasis> of your build. You set the
1529 configuration using a three-step process.</para>
1533 <term>Step 1: get ready for configuration.</term>
1535 <para>NOTE: if you're starting from a source distribution,
1536 rather than CVS sources, you can skip this step.</para>
1538 <para>Change directory to
1539 <constant>$(FPTOOLS_TOP)</constant> and
1541 <command>autoconf</command><indexterm><primary>autoconf</primary></indexterm>
1542 (with no arguments). This GNU program converts
1543 <filename><constant>$(FPTOOLS_TOP)</constant>/configure.in</filename>
1544 to a shell script called
1545 <filename><constant>$(FPTOOLS_TOP)</constant>/configure</filename>.
1548 <para>Some projects, including GHC, have their own
1549 configure script. If there's an
1550 <constant>$(FPTOOLS_TOP)/<project>/configure.in</constant>,
1551 then you need to run <command>autoconf</command> in that
1552 directory too.</para>
1554 <para>Both these steps are completely
1555 platform-independent; they just mean that the
1556 human-written file (<filename>configure.in</filename>) can
1557 be short, although the resulting shell script,
1558 <command>configure</command>, and
1559 <filename>mk/config.h.in</filename>, are long.</para>
1564 <term>Step 2: system configuration.</term>
1566 <para>Runs the newly-created <command>configure</command>
1567 script, thus:</para>
1570 ./configure <optional><parameter>args</parameter></optional>
1573 <para><command>configure</command>'s mission is to scurry
1574 round your computer working out what architecture it has,
1575 what operating system, whether it has the
1576 <Function>vfork</Function> system call, where
1577 <command>yacc</command> is kept, whether
1578 <command>gcc</command> is available, where various obscure
1579 <literal>#include</literal> files are, whether it's a
1580 leap year, and what the systems manager had for lunch. It
1581 communicates these snippets of information in two
1588 <filename>mk/config.mk.in</filename><indexterm><primary>config.mk.in</primary></indexterm>
1590 <filename>mk/config.mk</filename><indexterm><primary>config.mk</primary></indexterm>,
1591 substituting for things between
1592 “<literal>@</literal>” brackets. So,
1593 “<literal>@HaveGcc@</literal>” will be
1594 replaced by “<literal>YES</literal>” or
1595 “<literal>NO</literal>” depending on what
1596 <command>configure</command> finds.
1597 <filename>mk/config.mk</filename> is included by every
1598 Makefile (directly or indirectly), so the
1599 configuration information is thereby communicated to
1600 all Makefiles.</para>
1604 <para> It translates
1605 <filename>mk/config.h.in</filename><indexterm><primary>config.h.in</primary></indexterm>
1607 <filename>mk/config.h</filename><indexterm><primary>config.h</primary></indexterm>.
1608 The latter is <literal>#include</literal>d by
1609 various C programs, which can thereby make use of
1610 configuration information.</para>
1614 <para><command>configure</command> takes some optional
1615 arguments. Use <literal>./configure --help</literal> to
1616 get a list of the available arguments. Here are some of
1617 the ones you might need:</para>
1621 <term><literal>--with-ghc=<parameter>path</parameter></literal></term>
1622 <indexterm><primary><literal>--with-ghc</literal></primary>
1625 <para>Specifies the path to an installed GHC which
1626 you would like to use. This compiler will be used
1627 for compiling GHC-specific code (eg. GHC itself).
1628 This option <emphasis>cannot</emphasis> be specified
1629 using <filename>build.mk</filename> (see later),
1630 because <command>configure</command> needs to
1631 auto-detect the version of GHC you're using. The
1632 default is to look for a compiler named
1633 <literal>ghc</literal> in your path.</para>
1638 <term><literal>--with-hc=<parameter>path</parameter></literal></term>
1639 <indexterm><primary><literal>--with-hc</literal></primary>
1642 <para>Specifies the path to any installed Haskell
1643 compiler. This compiler will be used for compiling
1644 generic Haskell code. The default is to use
1645 <literal>ghc</literal>.</para>
1650 <term><literal>--with-gcc=<parameter>path</parameter></literal></term>
1651 <indexterm><primary><literal>--with-gcc</literal></primary>
1654 <para>Specifies the path to the installed GCC. This
1655 compiler will be used to compile all C files,
1656 <emphasis>except</emphasis> any generated by the
1657 installed Haskell compiler, which will have its own
1658 idea of which C compiler (if any) to use. The
1659 default is to use <literal>gcc</literal>.</para>
1664 <para><command>configure</command> caches the results of
1665 its run in <filename>config.cache</filename>. Quite often
1666 you don't want that; you're running
1667 <command>configure</command> a second time because
1668 something has changed. In that case, simply delete
1669 <filename>config.cache</filename>.</para>
1674 <term>Step 3: build configuration.</term>
1676 <para>Next, you say how this build of
1677 <literal>fptools</literal> is to differ from the standard
1678 defaults by creating a new file
1679 <filename>mk/build.mk</filename><indexterm><primary>build.mk</primary></indexterm>
1680 <emphasis>in the build tree</emphasis>. This file is the
1681 one and only file you edit in the build tree, precisely
1682 because it says how this build differs from the source.
1683 (Just in case your build tree does die, you might want to
1684 keep a private directory of <filename>build.mk</filename>
1685 files, and use a symbolic link in each build tree to point
1686 to the appropriate one.) So
1687 <filename>mk/build.mk</filename> never exists in the
1688 source tree—you create one in each build tree from
1689 the template. We'll discuss what to put in it
1695 <para>And that's it for configuration. Simple, eh?</para>
1697 <para>What do you put in your build-specific configuration file
1698 <filename>mk/build.mk</filename>? <emphasis>For almost all
1699 purposes all you will do is put make variable definitions that
1700 override those in</emphasis>
1701 <filename>mk/config.mk.in</filename>. The whole point of
1702 <filename>mk/config.mk.in</filename>—and its derived
1703 counterpart <filename>mk/config.mk</filename>—is to define
1704 the build configuration. It is heavily commented, as you will
1705 see if you look at it. So generally, what you do is look at
1706 <filename>mk/config.mk.in</filename>, and add definitions in
1707 <filename>mk/build.mk</filename> that override any of the
1708 <filename>config.mk</filename> definitions that you want to
1709 change. (The override occurs because the main boilerplate file,
1710 <filename>mk/boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>,
1711 includes <filename>build.mk</filename> after
1712 <filename>config.mk</filename>.)</para>
1714 <para>For example, <filename>config.mk.in</filename> contains
1715 the definition:</para>
1718 GhcHcOpts=-O -Rghc-timing
1721 <para>The accompanying comment explains that this is the list of
1722 flags passed to GHC when building GHC itself. For doing
1723 development, it is wise to add <literal>-DDEBUG</literal>, to
1724 enable debugging code. So you would add the following to
1725 <filename>build.mk</filename>:</para>
1727 <para>or, if you prefer,</para>
1730 GhcHcOpts += -DDEBUG
1733 <para>GNU <command>make</command> allows existing definitions to
1734 have new text appended using the “<literal>+=</literal>”
1735 operator, which is quite a convenient feature.)</para>
1737 <para>If you want to remove the <literal>-O</literal> as well (a
1738 good idea when developing, because the turn-around cycle gets a
1739 lot quicker), you can just override
1740 <literal>GhcLibHcOpts</literal> altogether:</para>
1743 GhcHcOpts=-DDEBUG -Rghc-timing
1746 <para>When reading <filename>config.mk.in</filename>, remember
1747 that anything between “@...@” signs is going to be substituted
1748 by <command>configure</command> later. You
1749 <emphasis>can</emphasis> override the resulting definition if
1750 you want, but you need to be a bit surer what you are doing.
1751 For example, there's a line that says:</para>
1757 <para>This defines the Make variables <constant>YACC</constant>
1758 to the pathname for a <command>yacc</command> that
1759 <command>configure</command> finds somewhere. If you have your
1760 own pet <command>yacc</command> you want to use instead, that's
1761 fine. Just add this line to <filename>mk/build.mk</filename>:</para>
1767 <para>You do not <emphasis>have</emphasis> to have a
1768 <filename>mk/build.mk</filename> file at all; if you don't,
1769 you'll get all the default settings from
1770 <filename>mk/config.mk.in</filename>.</para>
1772 <para>You can also use <filename>build.mk</filename> to override
1773 anything that <command>configure</command> got wrong. One place
1774 where this happens often is with the definition of
1775 <constant>FPTOOLS_TOP_ABS</constant>: this
1776 variable is supposed to be the canonical path to the top of your
1777 source tree, but if your system uses an automounter then the
1778 correct directory is hard to find automatically. If you find
1779 that <command>configure</command> has got it wrong, just put the
1780 correct definition in <filename>build.mk</filename>.</para>
1784 <sect2 id="sec-storysofar">
1785 <title>The story so far</title>
1787 <para>Let's summarise the steps you need to carry to get
1788 yourself a fully-configured build tree from scratch.</para>
1792 <para> Get your source tree from somewhere (CVS repository
1793 or source distribution). Say you call the root directory
1794 <filename>myfptools</filename> (it does not have to be
1795 called <filename>fptools</filename>). Make sure that you
1796 have the essential files (see <XRef
1797 LinkEnd="sec-source-tree">).</para>
1802 <para>(Optional) Use <command>lndir</command> or
1803 <command>mkshadowdir</command> to create a build tree.</para>
1807 $ mkshadowdir . /scratch/joe-bloggs/myfptools-sun4
1810 <para>(N.B. <command>mkshadowdir</command>'s first argument
1811 is taken relative to its second.) You probably want to give
1812 the build tree a name that suggests its main defining
1813 characteristic (in your mind at least), in case you later
1818 <para>Change directory to the build tree. Everything is
1819 going to happen there now.</para>
1822 $ cd /scratch/joe-bloggs/myfptools-sun4
1828 <para>Prepare for system configuration:</para>
1834 <para>(You can skip this step if you are starting from a
1835 source distribution, and you already have
1836 <filename>configure</filename> and
1837 <filename>mk/config.h.in</filename>.)</para>
1839 <para>Some projects, including GHC itself, have their own
1840 configure scripts, so it is necessary to run autoconf again
1841 in the appropriate subdirectories. eg:</para>
1844 $ (cd ghc; autoconf)
1849 <para>Do system configuration:</para>
1855 <para>Don't forget to check whether you need to add any
1856 arguments to <literal>configure</literal>; for example, a
1857 common requirement is to specify which GHC to use with
1858 <option>--with-ghc=<replaceable>ghc</replaceable></option>.</para>
1862 <para>Create the file <filename>mk/build.mk</filename>,
1863 adding definitions for your desired configuration
1872 <para>You can make subsequent changes to
1873 <filename>mk/build.mk</filename> as often as you like. You do
1874 not have to run any further configuration programs to make these
1875 changes take effect. In theory you should, however, say
1876 <command>gmake clean</command>, <command>gmake all</command>,
1877 because configuration option changes could affect
1878 anything—but in practice you are likely to know what's
1883 <title>Making things</title>
1885 <para>At this point you have made yourself a fully-configured
1886 build tree, so you are ready to start building real
1889 <para>The first thing you need to know is that <emphasis>you
1890 must use GNU <command>make</command>, usually called
1891 <command>gmake</command>, not standard Unix
1892 <command>make</command></emphasis>. If you use standard Unix
1893 <command>make</command> you will get all sorts of error messages
1894 (but no damage) because the <literal>fptools</literal>
1895 <command>Makefiles</command> use GNU <command>make</command>'s
1896 facilities extensively.</para>
1898 <para>To just build the whole thing, <command>cd</command> to
1899 the top of your <literal>fptools</literal> tree and type
1900 <command>gmake</command>. This will prepare the tree and build
1901 the various projects in the correct order.</para>
1904 <sect2 id="sec-bootstrapping">
1905 <title>Bootstrapping GHC</title>
1907 <para>GHC requires a 2-stage bootstrap in order to provide
1908 full functionality, including GHCi. By a 2-stage bootstrap, we
1909 mean that the compiler is built once using the installed GHC,
1910 and then again using the compiler built in the first stage. You
1911 can also build a stage 3 compiler, but this normally isn't
1912 necessary except to verify that the stage 2 compiler is working
1915 <para>Note that when doing a bootstrap, the stage 1 compiler
1916 must be built, followed by the runtime system and libraries, and
1917 then the stage 2 compiler. The correct ordering is implemented
1918 by the top-level fptools <filename>Makefile</filename>, so if
1919 you want everything to work automatically it's best to start
1920 <command>make</command> from the top of the tree. When building
1921 GHC, the top-level fptools <filename>Makefile</filename> is set
1922 up to do a 2-stage bootstrap by default (when you say
1923 <command>make</command>). Some other targets it supports
1930 <para>Build everything as normal, including the stage 1
1938 <para>Build the stage 2 compiler only.</para>
1945 <para>Build the stage 3 compiler only.</para>
1950 <term>bootstrap</term> <term>bootstrap2</term>
1952 <para>Build stage 1 followed by stage 2.</para>
1957 <term>bootstrap3</term>
1959 <para>Build stages 1, 2 and 3.</para>
1964 <term>install</term>
1966 <para>Install everything, including the compiler built in
1967 stage 2. To override the stage, say <literal>make install
1968 stage=<replaceable>n</replaceable></literal> where
1969 <replaceable>n</replaceable> is the stage to install.</para>
1974 <para>The top-level <filename>Makefile</filename> also arranges
1975 to do the appropriate <literal>make boot</literal> steps (see
1976 below) before actually building anything.</para>
1978 <para>The <literal>stage1</literal>, <literal>stage2</literal>
1979 and <literal>stage3</literal> targets also work in the
1980 <literal>ghc/compiler</literal> directory, but don't forget that
1981 each stage requires its own <literal>make boot</literal> step:
1982 for example, you must do</para>
1984 <screen>$ make boot stage=2</screen>
1986 <para>before <literal>make stage2</literal> in
1987 <literal>ghc/compiler</literal>.</para>
1990 <sect2 id="sec-standard-targets">
1991 <title>Standard Targets</title>
1992 <indexterm><primary>targets, standard makefile</primary></indexterm>
1993 <indexterm><primary>makefile targets</primary></indexterm>
1995 <para>In any directory you should be able to make the following:</para>
1999 <term><literal>boot</literal></term>
2001 <para>does the one-off preparation required to get ready
2002 for the real work. Notably, it does <command>gmake
2003 depend</command> in all directories that contain programs.
2004 It also builds the necessary tools for compilation to
2007 <para>Invoking the <literal>boot</literal> target
2008 explicitly is not normally necessary. From the top-level
2009 <literal>fptools</literal> directory, invoking
2010 <literal>gmake</literal> causes <literal>gmake boot
2011 all</literal> to be invoked in each of the project
2012 subdirectories, in the order specified by
2013 <literal>$(AllTargets)</literal> in
2014 <literal>config.mk</literal>.</para>
2016 <para>If you're working in a subdirectory somewhere and
2017 need to update the dependencies, <literal>gmake
2018 boot</literal> is a good way to do it.</para>
2023 <term><literal>all</literal></term>
2025 <para>makes all the final target(s) for this Makefile.
2026 Depending on which directory you are in a “final
2027 target” may be an executable program, a library
2028 archive, a shell script, or a Postscript file. Typing
2029 <command>gmake</command> alone is generally the same as
2030 typing <command>gmake all</command>.</para>
2035 <term><literal>install</literal></term>
2037 <para>installs the things built by <literal>all</literal>
2038 (except for the documentation). Where does it install
2039 them? That is specified by
2040 <filename>mk/config.mk.in</filename>; you can override it
2041 in <filename>mk/build.mk</filename>, or by running
2042 <command>configure</command> with command-line arguments
2043 like <literal>--bindir=/home/simonpj/bin</literal>; see
2044 <literal>./configure --help</literal> for the full
2050 <term><literal>install-docs</literal></term>
2052 <para>installs the documentation. Otherwise behaves just
2053 like <literal>install</literal>.</para>
2058 <term><literal>uninstall</literal></term>
2060 <para>reverses the effect of
2061 <literal>install</literal>.</para>
2066 <term><literal>clean</literal></term>
2068 <para>Delete all files from the current directory that are
2069 normally created by building the program. Don't delete
2070 the files that record the configuration, or files
2071 generated by <command>gmake boot</command>. Also preserve
2072 files that could be made by building, but normally aren't
2073 because the distribution comes with them.</para>
2078 <term><literal>distclean</literal></term>
2080 <para>Delete all files from the current directory that are
2081 created by configuring or building the program. If you
2082 have unpacked the source and built the program without
2083 creating any other files, <literal>make
2084 distclean</literal> should leave only the files that were
2085 in the distribution.</para>
2090 <term><literal>mostlyclean</literal></term>
2092 <para>Like <literal>clean</literal>, but may refrain from
2093 deleting a few files that people normally don't want to
2099 <term><literal>maintainer-clean</literal></term>
2101 <para>Delete everything from the current directory that
2102 can be reconstructed with this Makefile. This typically
2103 includes everything deleted by
2104 <literal>distclean</literal>, plus more: C source files
2105 produced by Bison, tags tables, Info files, and so
2108 <para>One exception, however: <literal>make
2109 maintainer-clean</literal> should not delete
2110 <filename>configure</filename> even if
2111 <filename>configure</filename> can be remade using a rule
2112 in the <filename>Makefile</filename>. More generally,
2113 <literal>make maintainer-clean</literal> should not delete
2114 anything that needs to exist in order to run
2115 <filename>configure</filename> and then begin to build the
2121 <term><literal>check</literal></term>
2123 <para>run the test suite.</para>
2128 <para>All of these standard targets automatically recurse into
2129 sub-directories. Certain other standard targets do not:</para>
2133 <term><literal>configure</literal></term>
2135 <para>is only available in the root directory
2136 <constant>$(FPTOOLS_TOP)</constant>; it has
2137 been discussed in <XRef
2138 LinkEnd="sec-build-config">.</para>
2143 <term><literal>depend</literal></term>
2145 <para>make a <filename>.depend</filename> file in each
2146 directory that needs it. This <filename>.depend</filename>
2147 file contains mechanically-generated dependency
2148 information; for example, suppose a directory contains a
2149 Haskell source module <filename>Foo.lhs</filename> which
2150 imports another module <literal>Baz</literal>. Then the
2151 generated <filename>.depend</filename> file will contain
2152 the dependency:</para>
2158 <para>which says that the object file
2159 <filename>Foo.o</filename> depends on the interface file
2160 <filename>Baz.hi</filename> generated by compiling module
2161 <literal>Baz</literal>. The <filename>.depend</filename>
2162 file is automatically included by every Makefile.</para>
2167 <term><literal>binary-dist</literal></term>
2169 <para>make a binary distribution. This is the target we
2170 use to build the binary distributions of GHC and
2176 <term><literal>dist</literal></term>
2178 <para>make a source distribution. Note that this target
2179 does “make distclean” as part of its work;
2180 don't use it if you want to keep what you've built.</para>
2185 <para>Most <filename>Makefile</filename>s have targets other
2186 than these. You can discover them by looking in the
2187 <filename>Makefile</filename> itself.</para>
2191 <title>Using a project from the build tree</title>
2193 <para>If you want to build GHC (say) and just use it direct from
2194 the build tree without doing <literal>make install</literal>
2195 first, you can run the in-place driver script:
2196 <filename>ghc/compiler/ghc-inplace</filename>.</para>
2198 <para> Do <emphasis>NOT</emphasis> use
2199 <filename>ghc/compiler/ghc</filename>, or
2200 <filename>ghc/compiler/ghc-5.xx</filename>, as these are the
2201 scripts intended for installation, and contain hard-wired paths
2202 to the installed libraries, rather than the libraries in the
2205 <para>Happy can similarly be run from the build tree, using
2206 <filename>happy/src/happy-inplace</filename>.</para>
2210 <title>Fast Making</title>
2212 <indexterm><primary>fastmake</primary></indexterm>
2213 <indexterm><primary>dependencies, omitting</primary></indexterm>
2214 <indexterm><primary>FAST, makefile variable</primary></indexterm>
2216 <para>Sometimes the dependencies get in the way: if you've made
2217 a small change to one file, and you're absolutely sure that it
2218 won't affect anything else, but you know that
2219 <command>make</command> is going to rebuild everything anyway,
2220 the following hack may be useful:</para>
2226 <para>This tells the make system to ignore dependencies and just
2227 build what you tell it to. In other words, it's equivalent to
2228 temporarily removing the <filename>.depend</filename> file in
2229 the current directory (where <command>mkdependHS</command> and
2230 friends store their dependency information).</para>
2232 <para>A bit of history: GHC used to come with a
2233 <command>fastmake</command> script that did the above job, but
2234 GNU make provides the features we need to do it without
2235 resorting to a script. Also, we've found that fastmaking is
2236 less useful since the advent of GHC's recompilation checker (see
2237 the User's Guide section on "Separate Compilation").</para>
2241 <sect1 id="sec-makefile-arch">
2242 <title>The <filename>Makefile</filename> architecture</title>
2243 <indexterm><primary>makefile architecture</primary></indexterm>
2245 <para><command>make</command> is great if everything
2246 works—you type <command>gmake install</command> and lo! the
2247 right things get compiled and installed in the right places. Our
2248 goal is to make this happen often, but somehow it often doesn't;
2249 instead some weird error message eventually emerges from the
2250 bowels of a directory you didn't know existed.</para>
2252 <para>The purpose of this section is to give you a road-map to
2253 help you figure out what is going right and what is going
2257 <title>Debugging</title>
2259 <para>Debugging <filename>Makefile</filename>s is something of a
2260 black art, but here's a couple of tricks that we find
2261 particularly useful. The following command allows you to see
2262 the contents of any make variable in the context of the current
2263 <filename>Makefile</filename>:</para>
2265 <screen>$ make show VALUE=HS_SRCS</screen>
2267 <para>where you can replace <literal>HS_SRCS</literal> with the
2268 name of any variable you wish to see the value of.</para>
2270 <para>GNU make has a <option>-d</option> option which generates
2271 a dump of the decision procedure used to arrive at a conclusion
2272 about which files should be recompiled. Sometimes useful for
2273 tracking down problems with superfluous or missing
2274 recompilations.</para>
2278 <title>A small project</title>
2280 <para>To get started, let us look at the
2281 <filename>Makefile</filename> for an imaginary small
2282 <literal>fptools</literal> project, <literal>small</literal>.
2283 Each project in <literal>fptools</literal> has its own directory
2284 in <constant>FPTOOLS_TOP</constant>, so the
2285 <literal>small</literal> project will have its own directory
2286 <constant>FPOOLS_TOP/small/</constant>. Inside the
2287 <filename>small/</filename> directory there will be a
2288 <filename>Makefile</filename>, looking something like
2291 <indexterm><primary>Makefile, minimal</primary></indexterm>
2294 # Makefile for fptools project "small"
2297 include $(TOP)/mk/boilerplate.mk
2299 SRCS = $(wildcard *.lhs) $(wildcard *.c)
2302 include $(TOP)/target.mk
2305 <para>this <filename>Makefile</filename> has three
2310 <para>The first section includes
2313 One of the most important
2314 features of GNU <command>make</command> that we use is the ability for a <filename>Makefile</filename> to
2315 include another named file, very like <command>cpp</command>'s <literal>#include</literal>
2320 a file of “boilerplate” code from the level
2321 above (which in this case will be
2322 <filename><constant>FPTOOLS_TOP</constant>/mk/boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>).
2323 As its name suggests, <filename>boilerplate.mk</filename>
2324 consists of a large quantity of standard
2325 <filename>Makefile</filename> code. We discuss this
2326 boilerplate in more detail in <XRef LinkEnd="sec-boiler">.
2327 <indexterm><primary>include, directive in
2328 Makefiles</primary></indexterm> <indexterm><primary>Makefile
2329 inclusion</primary></indexterm></para>
2331 <para>Before the <literal>include</literal> statement, you
2332 must define the <command>make</command> variable
2333 <constant>TOP</constant><indexterm><primary>TOP</primary></indexterm>
2334 to be the directory containing the <filename>mk</filename>
2335 directory in which the <filename>boilerplate.mk</filename>
2336 file is. It is <emphasis>not</emphasis> OK to simply say</para>
2339 include ../mk/boilerplate.mk # NO NO NO
2343 <para>Why? Because the <filename>boilerplate.mk</filename>
2344 file needs to know where it is, so that it can, in turn,
2345 <literal>include</literal> other files. (Unfortunately,
2346 when an <literal>include</literal>d file does an
2347 <literal>include</literal>, the filename is treated relative
2348 to the directory in which <command>gmake</command> is being
2349 run, not the directory in which the
2350 <literal>include</literal>d sits.) In general,
2351 <emphasis>every file <filename>foo.mk</filename> assumes
2353 <filename><constant>$(TOP)</constant>/mk/foo.mk</filename>
2354 refers to itself.</emphasis> It is up to the
2355 <filename>Makefile</filename> doing the
2356 <literal>include</literal> to ensure this is the case.</para>
2358 <para>Files intended for inclusion in other
2359 <filename>Makefile</filename>s are written to have the
2360 following property: <emphasis>after
2361 <filename>foo.mk</filename> is <literal>include</literal>d,
2362 it leaves <constant>TOP</constant> containing the same value
2363 as it had just before the <literal>include</literal>
2364 statement</emphasis>. In our example, this invariant
2365 guarantees that the <literal>include</literal> for
2366 <filename>target.mk</filename> will look in the same
2367 directory as that for <filename>boilerplate.mk</filename>.</para>
2371 <para> The second section defines the following standard
2372 <command>make</command> variables:
2373 <constant>SRCS</constant><indexterm><primary>SRCS</primary></indexterm>
2374 (the source files from which is to be built), and
2375 <constant>HS_PROG</constant><indexterm><primary>HS_PROG</primary></indexterm>
2376 (the executable binary to be built). We will discuss in
2377 more detail what the “standard variables” are,
2378 and how they affect what happens, in <XRef
2379 LinkEnd="sec-targets">.</para>
2381 <para>The definition for <constant>SRCS</constant> uses the
2382 useful GNU <command>make</command> construct
2383 <literal>$(wildcard $pat$)</literal><indexterm><primary>wildcard</primary></indexterm>,
2384 which expands to a list of all the files matching the
2385 pattern <literal>pat</literal> in the current directory. In
2386 this example, <constant>SRCS</constant> is set to the list
2387 of all the <filename>.lhs</filename> and
2388 <filename>.c</filename> files in the directory. (Let's
2389 suppose there is one of each, <filename>Foo.lhs</filename>
2390 and <filename>Baz.c</filename>.)</para>
2394 <para>The last section includes a second file of standard
2396 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>.
2397 It contains the rules that tell <command>gmake</command> how
2398 to make the standard targets (<Xref
2399 LinkEnd="sec-standard-targets">). Why, you ask, can't this
2400 standard code be part of
2401 <filename>boilerplate.mk</filename>? Good question. We
2402 discuss the reason later, in <Xref
2403 LinkEnd="sec-boiler-arch">.</para>
2405 <para>You do not <emphasis>have</emphasis> to
2406 <literal>include</literal> the
2407 <filename>target.mk</filename> file. Instead, you can write
2408 rules of your own for all the standard targets. Usually,
2409 though, you will find quite a big payoff from using the
2410 canned rules in <filename>target.mk</filename>; the price
2411 tag is that you have to understand what canned rules get
2412 enabled, and what they do (<Xref
2413 LinkEnd="sec-targets">).</para>
2417 <para>In our example <filename>Makefile</filename>, most of the
2418 work is done by the two <literal>include</literal>d files. When
2419 you say <command>gmake all</command>, the following things
2424 <para><command>gmake</command> figures out that the object
2425 files are <filename>Foo.o</filename> and
2426 <filename>Baz.o</filename>.</para>
2430 <para>It uses a boilerplate pattern rule to compile
2431 <filename>Foo.lhs</filename> to <filename>Foo.o</filename>
2432 using a Haskell compiler. (Which one? That is set in the
2433 build configuration.)</para>
2437 <para>It uses another standard pattern rule to compile
2438 <filename>Baz.c</filename> to <filename>Baz.o</filename>,
2439 using a C compiler. (Ditto.)</para>
2443 <para>It links the resulting <filename>.o</filename> files
2444 together to make <literal>small</literal>, using the Haskell
2445 compiler to do the link step. (Why not use
2446 <command>ld</command>? Because the Haskell compiler knows
2447 what standard libraries to link in. How did
2448 <command>gmake</command> know to use the Haskell compiler to
2449 do the link, rather than the C compiler? Because we set the
2450 variable <constant>HS_PROG</constant> rather than
2451 <constant>C_PROG</constant>.)</para>
2455 <para>All <filename>Makefile</filename>s should follow the above
2456 three-section format.</para>
2460 <title>A larger project</title>
2462 <para>Larger projects are usually structured into a number of
2463 sub-directories, each of which has its own
2464 <filename>Makefile</filename>. (In very large projects, this
2465 sub-structure might be iterated recursively, though that is
2466 rare.) To give you the idea, here's part of the directory
2467 structure for the (rather large) GHC project:</para>
2477 ...source files for documentation...
2480 ...source files for driver...
2483 parser/...source files for parser...
2484 renamer/...source files for renamer...
2488 <para>The sub-directories <filename>docs</filename>,
2489 <filename>driver</filename>, <filename>compiler</filename>, and
2490 so on, each contains a sub-component of GHC, and each has its
2491 own <filename>Makefile</filename>. There must also be a
2492 <filename>Makefile</filename> in
2493 <filename><constant>$(FPTOOLS_TOP)</constant>/ghc</filename>.
2494 It does most of its work by recursively invoking
2495 <command>gmake</command> on the <filename>Makefile</filename>s
2496 in the sub-directories. We say that
2497 <filename>ghc/Makefile</filename> is a <emphasis>non-leaf
2498 <filename>Makefile</filename></emphasis>, because it does little
2499 except organise its children, while the
2500 <filename>Makefile</filename>s in the sub-directories are all
2501 <emphasis>leaf <filename>Makefile</filename>s</emphasis>. (In
2502 principle the sub-directories might themselves contain a
2503 non-leaf <filename>Makefile</filename> and several
2504 sub-sub-directories, but that does not happen in GHC.)</para>
2506 <para>The <filename>Makefile</filename> in
2507 <filename>ghc/compiler</filename> is considered a leaf
2508 <filename>Makefile</filename> even though the
2509 <filename>ghc/compiler</filename> has sub-directories, because
2510 these sub-directories do not themselves have
2511 <filename>Makefile</filename>s in them. They are just used to
2512 structure the collection of modules that make up GHC, but all
2513 are managed by the single <filename>Makefile</filename> in
2514 <filename>ghc/compiler</filename>.</para>
2516 <para>You will notice that <filename>ghc/</filename> also
2517 contains a directory <filename>ghc/mk/</filename>. It contains
2518 GHC-specific <filename>Makefile</filename> boilerplate code.
2519 More precisely:</para>
2523 <para><filename>ghc/mk/boilerplate.mk</filename> is included
2524 at the top of <filename>ghc/Makefile</filename>, and of all
2525 the leaf <filename>Makefile</filename>s in the
2526 sub-directories. It in turn <literal>include</literal>s the
2527 main boilerplate file
2528 <filename>mk/boilerplate.mk</filename>.</para>
2532 <para><filename>ghc/mk/target.mk</filename> is
2533 <literal>include</literal>d at the bottom of
2534 <filename>ghc/Makefile</filename>, and of all the leaf
2535 <filename>Makefile</filename>s in the sub-directories. It
2536 in turn <literal>include</literal>s the file
2537 <filename>mk/target.mk</filename>.</para>
2541 <para>So these two files are the place to look for GHC-wide
2542 customisation of the standard boilerplate.</para>
2545 <sect2 id="sec-boiler-arch">
2546 <title>Boilerplate architecture</title>
2547 <indexterm><primary>boilerplate architecture</primary></indexterm>
2549 <para>Every <filename>Makefile</filename> includes a
2550 <filename>boilerplate.mk</filename><indexterm><primary>boilerplate.mk</primary></indexterm>
2551 file at the top, and
2552 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>
2553 file at the bottom. In this section we discuss what is in these
2554 files, and why there have to be two of them. In general:</para>
2558 <para><filename>boilerplate.mk</filename> consists of:</para>
2562 <para><emphasis>Definitions of millions of
2563 <command>make</command> variables</emphasis> that
2564 collectively specify the build configuration. Examples:
2565 <constant>HC_OPTS</constant><indexterm><primary>HC_OPTS</primary></indexterm>,
2566 the options to feed to the Haskell compiler;
2567 <constant>NoFibSubDirs</constant><indexterm><primary>NoFibSubDirs</primary></indexterm>,
2568 the sub-directories to enable within the
2569 <literal>nofib</literal> project;
2570 <constant>GhcWithHc</constant><indexterm><primary>GhcWithHc</primary></indexterm>,
2571 the name of the Haskell compiler to use when compiling
2572 GHC in the <literal>ghc</literal> project.</para>
2576 <para><emphasis>Standard pattern rules</emphasis> that
2577 tell <command>gmake</command> how to construct one file
2578 from another.</para>
2582 <para><filename>boilerplate.mk</filename> needs to be
2583 <literal>include</literal>d at the <emphasis>top</emphasis>
2584 of each <filename>Makefile</filename>, so that the user can
2585 replace the boilerplate definitions or pattern rules by
2586 simply giving a new definition or pattern rule in the
2587 <filename>Makefile</filename>. <command>gmake</command>
2588 simply takes the last definition as the definitive one.</para>
2590 <para>Instead of <emphasis>replacing</emphasis> boilerplate
2591 definitions, it is also quite common to
2592 <emphasis>augment</emphasis> them. For example, a
2593 <filename>Makefile</filename> might say:</para>
2599 <para>thereby adding “<option>-O</option>” to
2601 <constant>SRC_HC_OPTS</constant><indexterm><primary>SRC_HC_OPTS</primary></indexterm>.</para>
2605 <para><filename>target.mk</filename> contains
2606 <command>make</command> rules for the standard targets
2607 described in <Xref LinkEnd="sec-standard-targets">. These
2608 rules are selectively included, depending on the setting of
2609 certain <command>make</command> variables. These variables
2610 are usually set in the middle section of the
2611 <filename>Makefile</filename> between the two
2612 <literal>include</literal>s.</para>
2614 <para><filename>target.mk</filename> must be included at the
2615 end (rather than being part of
2616 <filename>boilerplate.mk</filename>) for several tiresome
2622 <para><command>gmake</command> commits target and
2623 dependency lists earlier than it should. For example,
2624 <FIlename>target.mk</FIlename> has a rule that looks
2628 $(HS_PROG) : $(OBJS)
2629 $(HC) $(LD_OPTS) $< -o $@
2632 <para>If this rule was in
2633 <filename>boilerplate.mk</filename> then
2634 <constant>$(HS_PROG)</constant><indexterm><primary>HS_PROG</primary></indexterm>
2636 <constant>$(OBJS)</constant><indexterm><primary>OBJS</primary></indexterm>
2637 would not have their final values at the moment
2638 <command>gmake</command> encountered the rule. Alas,
2639 <command>gmake</command> takes a snapshot of their
2640 current values, and wires that snapshot into the rule.
2641 (In contrast, the commands executed when the rule
2642 “fires” are only substituted at the moment
2643 of firing.) So, the rule must follow the definitions
2644 given in the <filename>Makefile</filename> itself.</para>
2648 <para>Unlike pattern rules, ordinary rules cannot be
2649 overriden or replaced by subsequent rules for the same
2650 target (at least, not without an error message).
2651 Including ordinary rules in
2652 <filename>boilerplate.mk</filename> would prevent the
2653 user from writing rules for specific targets in specific
2658 <para>There are a couple of other reasons I've
2659 forgotten, but it doesn't matter too much.</para>
2666 <sect2 id="sec-boiler">
2667 <title>The main <filename>mk/boilerplate.mk</filename> file</title>
2668 <indexterm><primary>boilerplate.mk</primary></indexterm>
2670 <para>If you look at
2671 <filename><constant>$(FPTOOLS_TOP)</constant>/mk/boilerplate.mk</filename>
2672 you will find that it consists of the following sections, each
2673 held in a separate file:</para>
2677 <term><filename>config.mk</filename></term>
2678 <indexterm><primary>config.mk</primary></indexterm>
2680 <para>is the build configuration file we discussed at
2681 length in <Xref LinkEnd="sec-build-config">.</para>
2686 <term><filename>paths.mk</filename></term>
2687 <indexterm><primary>paths.mk</primary></indexterm>
2689 <para>defines <command>make</command> variables for
2690 pathnames and file lists. This file contains code for
2691 automatically compiling lists of source files and deriving
2692 lists of object files from those. The results can be
2693 overriden in the <filename>Makefile</filename>, but in
2694 most cases the automatic setup should do the right
2697 <para>The following variables may be set in the
2698 <filename>Makefile</filename> to affect how the automatic
2699 source file search is done:</para>
2703 <term><literal>ALL_DIRS</literal></term>
2704 <indexterm><primary><literal>ALL_DIRS</literal></primary>
2707 <para>Set to a list of directories to search in
2708 addition to the current directory for source
2714 <term><literal>EXCLUDE_SRCS</literal></term>
2715 <indexterm><primary><literal>EXCLUDE_SRCS</literal></primary>
2718 <para>Set to a list of source files (relative to the
2719 current directory) to omit from the automatic
2720 search. The source searching machinery is clever
2721 enough to know that if you exclude a source file
2722 from which other sources are derived, then the
2723 derived sources should also be excluded. For
2724 example, if you set <literal>EXCLUDED_SRCS</literal>
2725 to include <filename>Foo.y</filename>, then
2726 <filename>Foo.hs</filename> will also be
2732 <term><literal>EXTRA_SRCS</literal></term>
2733 <indexterm><primary><literal>EXCLUDE_SRCS</literal></primary>
2736 <para>Set to a list of extra source files (perhaps
2737 in directories not listed in
2738 <literal>ALL_DIRS</literal>) that should be
2744 <para>The results of the automatic source file search are
2745 placed in the following make variables:</para>
2749 <term><literal>SRCS</literal></term>
2750 <indexterm><primary><literal>SRCS</literal></primary></indexterm>
2752 <para>All source files found, sorted and without
2753 duplicates, including those which might not exist
2754 yet but will be derived from other existing sources.
2755 <literal>SRCS</literal> <emphasis>can</emphasis> be
2756 overriden if necessary, in which case the variables
2757 below will follow suit.</para>
2762 <term><literal>HS_SRCS</literal></term>
2763 <indexterm><primary><literal>HS_SRCS</literal></primary></indexterm>
2765 <para>all Haskell source files in the current
2766 directory, including those derived from other source
2767 files (eg. Happy sources also give rise to Haskell
2773 <term><literal>HS_OBJS</literal></term>
2774 <indexterm><primary><literal>HS_OBJS</literal></primary></indexterm>
2776 <para>Object files derived from
2777 <literal>HS_SRCS</literal>.</para>
2782 <term><literal>HS_IFACES</literal></term>
2783 <indexterm><primary><literal>HS_IFACES</literal></primary></indexterm>
2785 <para>Interface files (<literal>.hi</literal> files)
2786 derived from <literal>HS_SRCS</literal>.</para>
2791 <term><literal>C_SRCS</literal></term>
2792 <indexterm><primary><literal>C_SRCS</literal></primary></indexterm>
2794 <para>All C source files found.</para>
2799 <term><literal>C_OBJS</literal></term>
2800 <indexterm><primary><literal>C_OBJS</literal></primary></indexterm>
2802 <para>Object files derived from
2803 <literal>C_SRCS</literal>.</para>
2808 <term><literal>SCRIPT_SRCS</literal></term>
2809 <indexterm><primary><literal>SCRIPT_SRCS</literal></primary></indexterm>
2811 <para>All script source files found
2812 (<literal>.lprl</literal> files).</para>
2817 <term><literal>SCRIPT_OBJS</literal></term>
2818 <indexterm><primary><literal>SCRIPT_OBJS</literal></primary></indexterm>
2820 <para><quote>object</quote> files derived from
2821 <literal>SCRIPT_SRCS</literal>
2822 (<literal>.prl</literal> files).</para>
2827 <term><literal>HSC_SRCS</literal></term>
2828 <indexterm><primary><literal>HSC_SRCS</literal></primary></indexterm>
2830 <para>All <literal>hsc2hs</literal> source files
2831 (<literal>.hsc</literal> files).</para>
2836 <term><literal>HAPPY_SRCS</literal></term>
2837 <indexterm><primary><literal>HAPPY_SRCS</literal></primary></indexterm>
2839 <para>All <literal>happy</literal> source files
2840 (<literal>.y</literal> or <literal>.hy</literal> files).</para>
2845 <term><literal>OBJS</literal></term>
2846 <indexterm><primary>OBJS</primary></indexterm>
2848 <para>the concatenation of
2849 <literal>$(HS_OBJS)</literal>,
2850 <literal>$(C_OBJS)</literal>, and
2851 <literal>$(SCRIPT_OBJS)</literal>.</para>
2856 <para>Any or all of these definitions can easily be
2857 overriden by giving new definitions in your
2858 <filename>Makefile</filename>.</para>
2860 <para>What, exactly, does <filename>paths.mk</filename>
2861 consider a <quote>source file</quote> to be? It's based
2862 on the file's suffix (e.g. <filename>.hs</filename>,
2863 <filename>.lhs</filename>, <filename>.c</filename>,
2864 <filename>.hy</filename>, etc), but this is the kind of
2865 detail that changes, so rather than enumerate the source
2866 suffices here the best thing to do is to look in
2867 <filename>paths.mk</filename>.</para>
2872 <term><filename>opts.mk</filename></term>
2873 <indexterm><primary>opts.mk</primary></indexterm>
2875 <para>defines <command>make</command> variables for option
2876 strings to pass to each program. For example, it defines
2877 <constant>HC_OPTS</constant><indexterm><primary>HC_OPTS</primary></indexterm>,
2878 the option strings to pass to the Haskell compiler. See
2879 <Xref LinkEnd="sec-suffix">.</para>
2884 <term><filename>suffix.mk</filename></term>
2885 <indexterm><primary>suffix.mk</primary></indexterm>
2887 <para>defines standard pattern rules—see <Xref
2888 LinkEnd="sec-suffix">.</para>
2893 <para>Any of the variables and pattern rules defined by the
2894 boilerplate file can easily be overridden in any particular
2895 <filename>Makefile</filename>, because the boilerplate
2896 <literal>include</literal> comes first. Definitions after this
2897 <literal>include</literal> directive simply override the default
2898 ones in <filename>boilerplate.mk</filename>.</para>
2901 <sect2 id="sec-suffix">
2902 <title>Pattern rules and options</title>
2903 <indexterm><primary>Pattern rules</primary></indexterm>
2906 <filename>suffix.mk</filename><indexterm><primary>suffix.mk</primary></indexterm>
2907 defines standard <emphasis>pattern rules</emphasis> that say how
2908 to build one kind of file from another, for example, how to
2909 build a <filename>.o</filename> file from a
2910 <filename>.c</filename> file. (GNU <command>make</command>'s
2911 <emphasis>pattern rules</emphasis> are more powerful and easier
2912 to use than Unix <command>make</command>'s <emphasis>suffix
2913 rules</emphasis>.)</para>
2915 <para>Almost all the rules look something like this:</para>
2920 $(CC) $(CC_OPTS) -c $< -o $@
2923 <para>Here's how to understand the rule. It says that
2924 <emphasis>something</emphasis><filename>.o</filename> (say
2925 <filename>Foo.o</filename>) can be built from
2926 <emphasis>something</emphasis><filename>.c</filename>
2927 (<filename>Foo.c</filename>), by invoking the C compiler (path
2928 name held in <constant>$(CC)</constant>), passing to it
2929 the options <constant>$(CC_OPTS)</constant> and
2930 the rule's dependent file of the rule
2931 <literal>$<</literal> (<filename>Foo.c</filename> in
2932 this case), and putting the result in the rule's target
2933 <literal>$@</literal> (<filename>Foo.o</filename> in this
2936 <para>Every program is held in a <command>make</command>
2937 variable defined in <filename>mk/config.mk</filename>—look
2938 in <filename>mk/config.mk</filename> for the complete list. One
2939 important one is the Haskell compiler, which is called
2940 <constant>$(HC)</constant>.</para>
2942 <para>Every program's options are are held in a
2943 <command>make</command> variables called
2944 <constant><prog>_OPTS</constant>. the
2945 <constant><prog>_OPTS</constant> variables are
2946 defined in <filename>mk/opts.mk</filename>. Almost all of them
2947 are defined like this:</para>
2950 CC_OPTS = $(SRC_CC_OPTS) $(WAY$(_way)_CC_OPTS) $($*_CC_OPTS) $(EXTRA_CC_OPTS)
2953 <para>The four variables from which
2954 <constant>CC_OPTS</constant> is built have the following
2959 <term><constant>SRC_CC_OPTS</constant><indexterm><primary>SRC_CC_OPTS</primary></indexterm>:</term>
2961 <para>options passed to all C compilations.</para>
2966 <term><constant>WAY_<way>_CC_OPTS</constant>:</term>
2968 <para>options passed to C compilations for way
2969 <literal><way></literal>. For example,
2970 <constant>WAY_mp_CC_OPTS</constant>
2971 gives options to pass to the C compiler when compiling way
2972 <literal>mp</literal>. The variable
2973 <constant>WAY_CC_OPTS</constant> holds
2974 options to pass to the C compiler when compiling the
2975 standard way. (<Xref LinkEnd="sec-ways"> dicusses
2976 multi-way compilation.)</para>
2981 <term><constant><module>_CC_OPTS</constant>:</term>
2983 <para>options to pass to the C compiler that are specific
2984 to module <literal><module></literal>. For example,
2985 <constant>SMap_CC_OPTS</constant> gives the
2986 specific options to pass to the C compiler when compiling
2987 <filename>SMap.c</filename>.</para>
2992 <term><constant>EXTRA_CC_OPTS</constant><indexterm><primary>EXTRA_CC_OPTS</primary></indexterm>:</term>
2994 <para>extra options to pass to all C compilations. This
2995 is intended for command line use, thus:</para>
2998 gmake libHS.a EXTRA_CC_OPTS="-v"
3005 <sect2 id="sec-targets">
3006 <title>The main <filename>mk/target.mk</filename> file</title>
3007 <indexterm><primary>target.mk</primary></indexterm>
3009 <para><filename>target.mk</filename> contains canned rules for
3010 all the standard targets described in <Xref
3011 LinkEnd="sec-standard-targets">. It is complicated by the fact
3012 that you don't want all of these rules to be active in every
3013 <filename>Makefile</filename>. Rather than have a plethora of
3014 tiny files which you can include selectively, there is a single
3015 file, <filename>target.mk</filename>, which selectively includes
3016 rules based on whether you have defined certain variables in
3017 your <filename>Makefile</filename>. This section explains what
3018 rules you get, what variables control them, and what the rules
3019 do. Hopefully, you will also get enough of an idea of what is
3020 supposed to happen that you can read and understand any weird
3021 special cases yourself.</para>
3025 <term><constant>HS_PROG</constant><indexterm><primary>HS_PROG</primary></indexterm>.</term>
3027 <para>If <constant>HS_PROG</constant> is defined,
3028 you get rules with the following targets:</para>
3032 <term><filename>HS_PROG</filename><indexterm><primary>HS_PROG</primary></indexterm></term>
3034 <para>itself. This rule links
3035 <constant>$(OBJS)</constant> with the Haskell
3036 runtime system to get an executable called
3037 <constant>$(HS_PROG)</constant>.</para>
3042 <term><literal>install</literal><indexterm><primary>install</primary></indexterm></term>
3045 <constant>$(HS_PROG)</constant> in
3046 <constant>$(bindir)</constant>.</para>
3055 <term><constant>C_PROG</constant><indexterm><primary>C_PROG</primary></indexterm></term>
3057 <para>is similar to <constant>HS_PROG</constant>,
3058 except that the link step links
3059 <constant>$(C_OBJS)</constant> with the C
3060 runtime system.</para>
3065 <term><constant>LIBRARY</constant><indexterm><primary>LIBRARY</primary></indexterm></term>
3067 <para>is similar to <constant>HS_PROG</constant>,
3068 except that it links
3069 <constant>$(LIB_OBJS)</constant> to make the
3070 library archive <constant>$(LIBRARY)</constant>,
3071 and <literal>install</literal> installs it in
3072 <constant>$(libdir)</constant>.</para>
3077 <term><constant>LIB_DATA</constant><indexterm><primary>LIB_DATA</primary></indexterm></term>
3079 <para>…</para>
3084 <term><constant>LIB_EXEC</constant><indexterm><primary>LIB_EXEC</primary></indexterm></term>
3086 <para>…</para>
3091 <term><constant>HS_SRCS</constant><indexterm><primary>HS_SRCS</primary></indexterm>, <constant>C_SRCS</constant><indexterm><primary>C_SRCS</primary></indexterm>.</term>
3093 <para>If <constant>HS_SRCS</constant> is defined
3094 and non-empty, a rule for the target
3095 <literal>depend</literal> is included, which generates
3096 dependency information for Haskell programs. Similarly
3097 for <constant>C_SRCS</constant>.</para>
3102 <para>All of these rules are “double-colon” rules,
3106 install :: $(HS_PROG)
3107 ...how to install it...
3110 <para>GNU <command>make</command> treats double-colon rules as
3111 separate entities. If there are several double-colon rules for
3112 the same target it takes each in turn and fires it if its
3113 dependencies say to do so. This means that you can, for
3114 example, define both <constant>HS_PROG</constant> and
3115 <constant>LIBRARY</constant>, which will generate two rules for
3116 <literal>install</literal>. When you type <command>gmake
3117 install</command> both rules will be fired, and both the program
3118 and the library will be installed, just as you wanted.</para>
3121 <sect2 id="sec-subdirs">
3122 <title>Recursion</title>
3123 <indexterm><primary>recursion, in makefiles</primary></indexterm>
3124 <indexterm><primary>Makefile, recursing into subdirectories</primary></indexterm>
3126 <para>In leaf <filename>Makefile</filename>s the variable
3127 <constant>SUBDIRS</constant><indexterm><primary>SUBDIRS</primary></indexterm>
3128 is undefined. In non-leaf <filename>Makefile</filename>s,
3129 <constant>SUBDIRS</constant> is set to the list of
3130 sub-directories that contain subordinate
3131 <filename>Makefile</filename>s. <emphasis>It is up to you to
3132 set <constant>SUBDIRS</constant> in the
3133 <filename>Makefile</filename>.</emphasis> There is no automation
3134 here—<constant>SUBDIRS</constant> is too important to
3137 <para>When <constant>SUBDIRS</constant> is defined,
3138 <filename>target.mk</filename> includes a rather neat rule for
3139 the standard targets (<Xref LinkEnd="sec-standard-targets"> that
3140 simply invokes <command>make</command> recursively in each of
3141 the sub-directories.</para>
3143 <para><emphasis>These recursive invocations are guaranteed to
3144 occur in the order in which the list of directories is specified
3145 in <constant>SUBDIRS</constant>. </emphasis>This guarantee can
3146 be important. For example, when you say <command>gmake
3147 boot</command> it can be important that the recursive invocation
3148 of <command>make boot</command> is done in one sub-directory
3149 (the include files, say) before another (the source files).
3150 Generally, put the most independent sub-directory first, and the
3151 most dependent last.</para>
3154 <sect2 id="sec-ways">
3155 <title>Way management</title>
3156 <indexterm><primary>way management</primary></indexterm>
3158 <para>We sometimes want to build essentially the same system in
3159 several different “ways”. For example, we want to build GHC's
3160 <literal>Prelude</literal> libraries with and without profiling,
3161 so that there is an appropriately-built library archive to link
3162 with when the user compiles his program. It would be possible
3163 to have a completely separate build tree for each such “way”,
3164 but it would be horribly bureaucratic, especially since often
3165 only parts of the build tree need to be constructed in multiple
3169 <filename>target.mk</filename><indexterm><primary>target.mk</primary></indexterm>
3170 contains some clever magic to allow you to build several
3171 versions of a system; and to control locally how many versions
3172 are built and how they differ. This section explains the
3175 <para>The files for a particular way are distinguished by
3176 munging the suffix. The <quote>normal way</quote> is always
3177 built, and its files have the standard suffices
3178 <filename>.o</filename>, <filename>.hi</filename>, and so on.
3179 In addition, you can build one or more extra ways, each
3180 distinguished by a <emphasis>way tag</emphasis>. The object
3181 files and interface files for one of these extra ways are
3182 distinguished by their suffix. For example, way
3183 <literal>mp</literal> has files
3184 <filename>.mp_o</filename> and
3185 <filename>.mp_hi</filename>. Library archives have their
3186 way tag the other side of the dot, for boring reasons; thus,
3187 <filename>libHS_mp.a</filename>.</para>
3189 <para>A <command>make</command> variable called
3190 <constant>way</constant> holds the current way tag.
3191 <emphasis><constant>way</constant> is only ever set on the
3192 command line of <command>gmake</command></emphasis> (usually in
3193 a recursive invocation of <command>gmake</command> by the
3194 system). It is never set inside a
3195 <filename>Makefile</filename>. So it is a global constant for
3196 any one invocation of <command>gmake</command>. Two other
3197 <command>make</command> variables,
3198 <constant>way_</constant> and
3199 <constant>_way</constant> are immediately derived from
3200 <constant>$(way)</constant> and never altered. If
3201 <constant>way</constant> is not set, then neither are
3202 <constant>way_</constant> and
3203 <constant>_way</constant>, and the invocation of
3204 <command>make</command> will build the <quote>normal
3205 way</quote>. If <constant>way</constant> is set, then the other
3206 two variables are set in sympathy. For example, if
3207 <constant>$(way)</constant> is “<literal>mp</literal>”,
3208 then <constant>way_</constant> is set to
3209 “<literal>mp_</literal>” and
3210 <constant>_way</constant> is set to
3211 “<literal>_mp</literal>”. These three variables are
3212 then used when constructing file names.</para>
3214 <para>So how does <command>make</command> ever get recursively
3215 invoked with <constant>way</constant> set? There are two ways
3216 in which this happens:</para>
3220 <para>For some (but not all) of the standard targets, when
3221 in a leaf sub-directory, <command>make</command> is
3222 recursively invoked for each way tag in
3223 <constant>$(WAYS)</constant>. You set
3224 <constant>WAYS</constant> in the
3225 <filename>Makefile</filename> to the list of way tags you
3226 want these targets built for. The mechanism here is very
3227 much like the recursive invocation of
3228 <command>make</command> in sub-directories (<Xref
3229 LinkEnd="sec-subdirs">). It is up to you to set
3230 <constant>WAYS</constant> in your
3231 <filename>Makefile</filename>; this is how you control what
3232 ways will get built.</para>
3236 <para>For a useful collection of targets (such as
3237 <filename>libHS_mp.a</filename>,
3238 <filename>Foo.mp_o</filename>) there is a rule which
3239 recursively invokes <command>make</command> to make the
3240 specified target, setting the <constant>way</constant>
3241 variable. So if you say <command>gmake
3242 Foo.mp_o</command> you should see a recursive
3243 invocation <command>gmake Foo.mp_o way=mp</command>,
3244 and <emphasis>in this recursive invocation the pattern rule
3245 for compiling a Haskell file into a <filename>.o</filename>
3246 file will match</emphasis>. The key pattern rules (in
3247 <filename>suffix.mk</filename>) look like this:
3251 $(HC) $(HC_OPTS) $< -o $@
3258 <para>You can invoke <command>make</command> with a
3259 particular <literal>way</literal> setting yourself, in order
3260 to build files related to a particular
3261 <literal>way</literal> in the current directory. eg.
3267 will build files for the profiling way only in the current
3274 <title>When the canned rule isn't right</title>
3276 <para>Sometimes the canned rule just doesn't do the right thing.
3277 For example, in the <literal>nofib</literal> suite we want the
3278 link step to print out timing information. The thing to do here
3279 is <emphasis>not</emphasis> to define
3280 <constant>HS_PROG</constant> or
3281 <constant>C_PROG</constant>, and instead define a special
3282 purpose rule in your own <filename>Makefile</filename>. By
3283 using different variable names you will avoid the canned rules
3284 being included, and conflicting with yours.</para>
3288 <sect1 id="building-docs">
3289 <title>Building the documentation</title>
3291 <sect2 id="pre-supposed-doc-tools">
3292 <title>Tools for building the Documentation</title>
3294 <para>The following additional tools are required if you want to
3295 format the documentation that comes with the
3296 <literal>fptools</literal> projects:</para>
3300 <term>DocBook</term>
3301 <indexterm><primary>pre-supposed: DocBook</primary></indexterm>
3302 <indexterm><primary>DocBook, pre-supposed</primary></indexterm>
3304 <para>Much of our documentation is written in SGML, using
3305 the DocBook DTD. Instructions on installing and
3306 configuring the DocBook tools are below.</para>
3312 <indexterm><primary>pre-supposed: TeX</primary></indexterm>
3313 <indexterm><primary>TeX, pre-supposed</primary></indexterm>
3315 <para>A decent TeX distribution is required if you want to
3316 produce printable documentation. We recomment teTeX,
3317 which includes just about everything you need.</para>
3322 <term>Haddock</term>
3323 <indexterm><primary>Haddock</primary>
3326 <para>Haddock is a Haskell documentation tool that we use
3327 for automatically generating documentation from the
3328 library source code. It is an <literal>fptools</literal>
3329 project in itself. To build documentation for the
3330 libraries (<literal>fptools/libraries</literal>) you
3331 should check out and build Haddock in
3332 <literal>fptools/haddock</literal>. Haddock requires GHC
3340 <title>Installing the DocBook tools</title>
3343 <title>Installing the DocBook tools on Linux</title>
3345 <para>If you're on a recent RedHat system (7.0+), you probably
3346 have working DocBook tools already installed. The configure
3347 script should detect your setup and you're away.</para>
3349 <para>If you don't have DocBook tools installed, and you are
3350 using a system that can handle RedHat RPM packages, you can
3351 probably use the <ULink
3352 URL="http://sourceware.cygnus.com/docbook-tools/">Cygnus
3353 DocBook tools</ULink>, which is the most shrink-wrapped SGML
3354 suite that we could find. You need all the RPMs except for
3355 psgml (i.e. <Filename>docbook</Filename>,
3356 <Filename>jade</Filename>, <Filename>jadetex</Filename>,
3357 <Filename>sgmlcommon</Filename> and
3358 <Filename>stylesheets</Filename>). Note that most of these
3359 RPMs are architecture neutral, so are likely to be found in a
3360 <Filename>noarch</Filename> directory. The SuSE RPMs also
3361 work; the RedHat ones <Emphasis>don't</Emphasis> in RedHat 6.2
3362 (7.0 and later should be OK), but they are easy to fix: just
3364 <Filename>/usr/lib/sgml/stylesheets/nwalsh-modular/lib/dblib.dsl</Filename>
3365 to <Filename>/usr/lib/sgml/lib/dblib.dsl</Filename>. </para>
3369 <title>Installing DocBook on FreeBSD</title>
3371 <para>On FreeBSD systems, the easiest way to get DocBook up
3372 and running is to install it from the ports tree or a
3373 pre-compiled package (packages are available from your local
3374 FreeBSD mirror site).</para>
3376 <para>To use the ports tree, do this:
3378 $ cd /usr/ports/textproc/docproj
3381 This installs the FreeBSD documentation project tools, which
3382 includes everything needed to format the GHC
3383 documentation.</para>
3387 <title>Installing from binaries on Windows</title>
3389 <Para>It's a good idea to use Norman Walsh's <ULink
3390 URL="http://nwalsh.com/docbook/dsssl/doc/install.html">installation
3391 notes</ULink> as a guide. You should get version 3.1 of
3392 DocBook, and note that his file <Filename>test.sgm</Filename>
3393 won't work, as it needs version 3.0. You should unpack Jade
3394 into <Filename>\Jade</Filename>, along with the entities,
3395 DocBook into <Filename>\docbook</Filename>, and the DocBook
3396 stylesheets into <Filename>\docbook\stylesheets</Filename> (so
3397 they actually end up in
3398 <Filename>\docbook\stylesheets\docbook</Filename>).</para>
3403 <title>Installing the DocBook tools from source</title>
3408 <para>Install <ULink
3409 URL="http://openjade.sourceforge.net/">OpenJade</ULink>
3410 (Windows binaries are available as well as sources). If you
3411 want DVI, PS, or PDF then install JadeTeX from the
3412 <Filename>dsssl</Filename> subdirectory. (If you get the
3416 ! LaTeX Error: Unknown option implicit=false' for package hyperref'.
3419 your version of <Command>hyperref</Command> is out of date;
3420 download it from CTAN
3421 (<Filename>macros/latex/contrib/supported/hyperref</Filename>),
3422 and make it, ensuring that you have first removed or renamed
3423 your old copy. If you start getting file not found errors
3424 when making the test for <Command>hyperref</Command>, you
3425 can abort at that point and proceed straight to
3426 <Command>make install</Command>, or enter them as
3427 <Filename>../</Filename><Emphasis>filename</Emphasis>.)</para>
3429 <para>Make links from <Filename>virtex</Filename> to
3430 <Filename>jadetex</Filename> and
3431 <Filename>pdfvirtex</Filename> to
3432 <Filename>pdfjadetex</Filename> (otherwise DVI, PostScript
3433 and PDF output will not work). Copy
3434 <Filename>dsssl/*.{dtd,dsl}</Filename> and
3435 <Filename>catalog</Filename> to
3436 <Filename>/usr/[local/]lib/sgml</Filename>.</para>
3440 <title>DocBook and the DocBook stylesheets</title>
3442 <para>Get a Zip of <ULink
3443 URL="http://www.oasis-open.org/docbook/sgml/3.1/index.html">DocBook</ULink>
3444 and install the contents in
3445 <Filename>/usr/[local/]/lib/sgml</Filename>.</para>
3447 <para>Get the <ULink
3448 URL="http://nwalsh.com/docbook/dsssl/">DocBook
3449 stylesheets</ULink> and install in
3450 <Filename>/usr/[local/]lib/sgml/stylesheets</Filename>
3451 (thereby creating a subdirectory docbook). For indexing,
3452 copy or link <Filename>collateindex.pl</Filename> from the
3453 DocBook stylesheets archive in <Filename>bin</Filename> into
3454 a directory on your <Constant>PATH</Constant>.</para>
3456 <para>Download the <ULink
3457 URL="http://www.oasis-open.org/cover/ISOEnts.zip">ISO
3458 entities</ULink> into
3459 <Filename>/usr/[local/]lib/sgml</Filename>.</para>
3465 <title>Configuring the DocBook tools</title>
3467 <Para>Once the DocBook tools are installed, the configure script
3468 will detect them and set up the build system accordingly. If you
3469 have a system that isn't supported, let us know, and we'll try
3474 <title>Remaining problems</title>
3476 <para>If you install from source, you'll get a pile of warnings
3479 <Screen>DTDDECL catalog entries are not supported</Screen>
3481 every time you build anything. These can safely be ignored, but
3482 if you find them tedious you can get rid of them by removing all
3483 the <Constant>DTDDECL</Constant> entries from
3484 <Filename>docbook.cat</Filename>.</para>
3488 <title>Building the documentation</title>
3490 <para>To build documentation in a certain format, you can
3491 say, for example,</para>
3497 <para>to build HTML documentation below the current directory.
3498 The available formats are: <literal>dvi</literal>,
3499 <literal>ps</literal>, <literal>pdf</literal>,
3500 <literal>html</literal>, and <literal>rtf</literal>. Note that
3501 not all documentation can be built in all of these formats: HTML
3502 documentation is generally supported everywhere, and DocBook
3503 documentation might support the other formats (depending on what
3504 other tools you have installed).</para>
3506 <para>All of these targets are recursive; that is, saying
3507 <literal>make html</literal> will make HTML docs for all the
3508 documents recursively below the current directory.</para>
3510 <para>Because there are many different formats that the DocBook
3511 documentation can be generated in, you have to select which ones
3512 you want by setting the <literal>SGMLDocWays</literal> variable
3513 to a list of them. For example, in
3514 <filename>build.mk</filename> you might have a line:</para>
3517 SGMLDocWays = html ps
3520 <para>This will cause the documentation to be built in the requested
3521 formats as part of the main build (the default is not to build
3522 any documentation at all).</para>
3526 <title>Installing the documentation</title>
3528 <para>To install the documentation, use:</para>
3534 <para>This will install the documentation into
3535 <literal>$(datadir)</literal> (which defaults to
3536 <literal>$(prefix)/share</literal>). The exception is HTML
3537 documentation, which goes into
3538 <literal>$(datadir)/html</literal>, to keep things tidy.</para>
3540 <para>Note that unless you set <literal>$(SGMLDocWays)</literal>
3541 to a list of formats, the <literal>install-docs</literal> target
3542 won't do anything for SGML documentation.</para>
3548 <sect1 id="sec-porting-ghc">
3549 <title>Porting GHC</title>
3551 <para>This section describes how to port GHC to a currenly
3552 unsupported platform. There are two distinct
3553 possibilities:</para>
3557 <para>The hardware architecture for your system is already
3558 supported by GHC, but you're running an OS that isn't
3559 supported (or perhaps has been supported in the past, but
3560 currently isn't). This is the easiest type of porting job,
3561 but it still requires some careful bootstrapping. Proceed to
3562 <xref linkend="sec-booting-from-hc">.</para>
3566 <para>Your system's hardware architecture isn't supported by
3567 GHC. This will be a more difficult port (though by comparison
3568 perhaps not as difficult as porting gcc). Proceed to <xref
3569 linkend="unregisterised-porting">.</para>
3573 <sect2 id="sec-booting-from-hc">
3574 <title>Booting/porting from C (<filename>.hc</filename>) files</title>
3576 <indexterm><primary>building GHC from .hc files</primary></indexterm>
3577 <indexterm><primary>booting GHC from .hc files</primary></indexterm>
3578 <indexterm><primary>porting GHC</primary></indexterm>
3580 <para>Bootstrapping GHC on a system without GHC already
3581 installed is achieved by taking the intermediate C files (known
3582 as HC files) from a GHC compilation on a supported system to the
3583 target machine, and compiling them using gcc to get a working
3586 <para><emphasis>NOTE: GHC version 5.xx is significantly harder
3587 to bootstrap from C than previous versions. We recommend
3588 starting from version 4.08.2 if you need to bootstrap in this
3589 way.</emphasis></para>
3591 <para>HC files are architecture-dependent (but not
3592 OS-dependent), so you have to get a set that were generated on
3593 similar hardware. There may be some supplied on the GHC
3594 download page, otherwise you'll have to compile some up
3595 yourself, or start from <emphasis>unregisterised</emphasis> HC
3596 files - see <xref linkend="unregisterised-porting">.</para>
3598 <para>The following steps should result in a working GHC build
3599 with full libraries:</para>
3603 <para>Unpack the HC files on top of a fresh source tree
3604 (make sure the source tree version matches the version of
3605 the HC files <emphasis>exactly</emphasis>!). This will
3606 place matching <filename>.hc</filename> files next to the
3607 corresponding Haskell source (<filename>.hs</filename> or
3608 <filename>.lhs</filename>) in the compiler subdirectory
3609 <filename>ghc/compiler</filename> and in the libraries
3610 (subdirectories of <filename>hslibs</filename> and
3611 <literal>libraries</literal>).</para>
3615 <para>The actual build process is fully automated by the
3616 <filename>hc-build</filename> script located in the
3617 <filename>distrib</filename> directory. If you eventually
3618 want to install GHC into the directory
3619 <replaceable>dir</replaceable>, the following
3620 command will execute the whole build process (it won't
3621 install yet):</para>
3624 foo% distrib/hc-build --prefix=<replaceable>dir</replaceable>
3626 <indexterm><primary>--hc-build</primary></indexterm>
3628 <para>By default, the installation directory is
3629 <filename>/usr/local</filename>. If that is what you want,
3630 you may omit the argument to <filename>hc-build</filename>.
3631 Generally, any option given to <filename>hc-build</filename>
3632 is passed through to the configuration script
3633 <filename>configure</filename>. If
3634 <filename>hc-build</filename> successfully completes the
3635 build process, you can install the resulting system, as
3645 <sect2 id="unregisterised-porting">
3646 <title>Porting GHC to a new architecture</title>
3648 <para>The first step in porting to a new architecture is to get
3649 an <firstterm>unregisterised</firstterm> build working. An
3650 unregisterised build is one that compiles via vanilla C only.
3651 By contrast, a registerised build uses the following
3652 architecture-specific hacks for speed:</para>
3656 <para>Global register variables: certain abstract machine
3657 <quote>registers</quote> are mapped to real machine
3658 registers, depending on how many machine registers are
3660 <filename>ghc/includes/MachRegs.h</filename>).</para>
3664 <para>Assembly-mangling: when compiling via C, we feed the
3665 assembly generated by gcc though a Perl script known as the
3666 <firstterm>mangler</firstterm> (see
3667 <filename>ghc/driver/mangler/ghc-asm.lprl</filename>). The
3668 mangler rearranges the assembly to support tail-calls and
3669 various other optimisations.</para>
3673 <para>In an unregisterised build, neither of these hacks are
3674 used — the idea is that the C code generated by the
3675 compiler should compile using gcc only. The lack of these
3676 optimisations costs about a factor of two in performance, but
3677 since unregisterised compilation is usually just a step on the
3678 way to a full registerised port, we don't mind too much.</para>
3681 <title>Building an unregisterised port</title>
3683 <para>The first step is to get some unregisterised HC files.
3684 Either (a) download them from the GHC site (if there are
3685 some available for the right version of GHC), or
3686 (b) build them yourself on any machine with a working
3687 GHC. If at all possible this should be a machine with the
3688 same word size as the target.</para>
3690 <para>There is a script available which should automate the
3691 process of doing the 2-stage bootstrap necessary to get the
3692 unregisterised HC files - it's available in <ulink
3693 url="http://cvs.haskell.org/cgi-bin/cvsweb.cgi/fptools/distrib/cross-port"><filename>fptools/distrib/cross-port</filename></ulink>
3696 <para>Now take these unregisterised HC files to the target
3697 platform and bootstrap a compiler from them as per the
3698 instructions in <xref linkend="sec-booting-from-hc">. In
3699 <filename>build.mk</filename>, you need to tell the build
3700 system that the compiler you're building is
3701 (a) unregisterised itself, and (b) builds
3702 unregisterised binaries. This varies depending on the GHC
3703 version you're bootstraping:</para>
3706 # build.mk for GHC 4.08.x
3707 GhcWithRegisterised=NO
3711 # build.mk for GHC 5.xx
3712 GhcUnregisterised=YES
3715 <para>Version 5.xx only: use the option
3716 <option>--enable-hc-boot-unregisterised</option> instead of
3717 <option>--enable-hc-boot</option> when running
3718 <filename>./configure</filename>.</para>
3720 <para>The build may not go through cleanly. We've tried to
3721 stick to writing portable code in most parts of the compiler,
3722 so it should compile on any POSIXish system with gcc, but in
3723 our experience most systems differ from the standards in one
3724 way or another. Deal with any problems as they arise - if you
3725 get stuck, ask the experts on
3726 <email>glasgow-haskell-users@haskell.org</email>.</para>
3728 <para>Once you have the unregisterised compiler up and
3729 running, you can use it to start a registerised port. The
3730 following sections describe the various parts of the system
3731 that will need architecture-specific tweaks in order to get a
3732 registerised build going.</para>
3734 <para>Lots of useful information about the innards of GHC is
3735 available in the <ulink
3736 url="http://www.cse.unsw.edu.au/~chak/haskell/ghc/comm/">GHC
3737 Commentary</ulink>, which might be helpful if you run into
3738 some code which needs tweaking for your system.</para>
3742 <title>Porting the RTS</title>
3744 <para>The following files need architecture-specific code for a
3745 registerised build:</para>
3749 <term><filename>ghc/includes/MachRegs.h</filename></term>
3750 <indexterm><primary><filename>MachRegs.h</filename></primary>
3753 <para>Defines the STG-register to machine-register
3754 mapping. You need to know your platform's C calling
3755 convention, and which registers are generally available
3756 for mapping to global register variables. There are
3757 plenty of useful comments in this file.</para>
3761 <term><filename>ghc/includes/TailCalls.h</filename></term>
3762 <indexterm><primary><filename>TailCalls.h</filename></primary>
3765 <para>Macros that cooperate with the mangler (see <xref
3766 linkend="sec-mangler">) to make proper tail-calls
3771 <term><filename>ghc/rts/Adjustor.c</filename></term>
3772 <indexterm><primary><filename>Adjustor.c</filename></primary>
3776 <literal>foreign import "wrapper"</literal>
3778 <literal>foreign export dynamic</literal>).
3779 Not essential for getting GHC bootstrapped, so this file
3780 can be deferred until later if necessary.</para>
3784 <term><filename>ghc/rts/StgCRun.c</filename></term>
3785 <indexterm><primary><filename>StgCRun.c</filename></primary>
3788 <para>The little assembly layer between the C world and
3789 the Haskell world. See the comments and code for the
3790 other architectures in this file for pointers.</para>
3794 <term><filename>ghc/rts/MBlock.h</filename></term>
3795 <term><filename>ghc/rts/MBlock.c</filename></term>
3796 <indexterm><primary><filename>MBlock.h</filename></primary>
3798 <indexterm><primary><filename>MBlock.c</filename></primary>
3801 <para>These files are really OS-specific rather than
3802 architecture-specific. In <filename>MBlock.h</filename>
3803 is specified the absolute location at which the RTS
3804 should try to allocate memory on your platform (try to
3805 find an area which doesn't conflict with code or dynamic
3806 libraries). In <filename>Mblock.c</filename> you might
3807 need to tweak the call to <literal>mmap()</literal> for
3814 <sect3 id="sec-mangler">
3815 <title>The mangler</title>
3817 <para>The mangler is an evil Perl-script that rearranges the
3818 assembly code output from gcc to do two main things:</para>
3822 <para>Remove function prologues and epilogues, and all
3823 movement of the C stack pointer. This is to support
3824 tail-calls: every code block in Haskell code ends in an
3825 explicit jump, so we don't want the C-stack overflowing
3826 while we're jumping around between code blocks.</para>
3829 <para>Move the <firstterm>info table</firstterm> for a
3830 closure next to the entry code for that closure. In
3831 unregisterised code, info tables contain a pointer to the
3832 entry code, but in registerised compilation we arrange
3833 that the info table is shoved right up against the entry
3834 code, and addressed backwards from the entry code pointer
3835 (this saves a word in the info table and an extra
3836 indirection when jumping to the closure entry
3841 <para>The mangler is abstracted to a certain extent over some
3842 architecture-specific things such as the particular assembler
3843 directives used to herald symbols. Take a look at the
3844 definitions for other architectures and use these as a
3845 starting point.</para>
3849 <title>The native code generator</title>
3851 <para>The native code generator isn't essential to getting a
3852 registerised build going, but it's a desirable thing to have
3853 because it can cut compilation times in half. The native code
3854 generator is described in some detail in the <ulink
3855 url="http://www.cse.unsw.edu.au/~chak/haskell/ghc/comm/">GHC
3856 commentary</ulink>.</para>
3862 <para>To support GHCi, you need to port the dynamic linker
3863 (<filename>fptools/ghc/rts/Linker.c</filename>). The linker
3864 currently supports the ELF and PEi386 object file formats - if
3865 your platform uses one of these then you probably don't have
3866 to do anything except fiddle with the
3867 <literal>#ifdef</literal>s at the top of
3868 <filename>Linker.c</filename> to tell it about your OS.</para>
3870 <para>If your system uses a different object file format, then
3871 you have to write a linker — good luck!</para>
3877 <sect1 id="sec-build-pitfalls">
3878 <title>Known pitfalls in building Glasgow Haskell
3880 <indexterm><primary>problems, building</primary></indexterm>
3881 <indexterm><primary>pitfalls, in building</primary></indexterm>
3882 <indexterm><primary>building pitfalls</primary></indexterm></title>
3885 WARNINGS about pitfalls and known “problems”:
3894 One difficulty that comes up from time to time is running out of space
3895 in <literal>TMPDIR</literal>. (It is impossible for the configuration stuff to
3896 compensate for the vagaries of different sysadmin approaches to temp
3898 <indexterm><primary>tmp, running out of space in</primary></indexterm>
3900 The quickest way around it is <command>setenv TMPDIR /usr/tmp</command><indexterm><primary>TMPDIR</primary></indexterm> or
3901 even <command>setenv TMPDIR .</command> (or the equivalent incantation with your shell
3904 The best way around it is to say
3907 export TMPDIR=<dir>
3910 in your <filename>build.mk</filename> file.
3911 Then GHC and the other <literal>fptools</literal> programs will use the appropriate directory
3920 In compiling some support-code bits, e.g., in <filename>ghc/rts/gmp</filename> and even
3921 in <filename>ghc/lib</filename>, you may get a few C-compiler warnings. We think these
3929 When compiling via C, you'll sometimes get “warning: assignment from
3930 incompatible pointer type” out of GCC. Harmless.
3937 Similarly, <command>ar</command>chiving warning messages like the following are not
3941 ar: filename GlaIOMonad__1_2s.o truncated to GlaIOMonad_
3942 ar: filename GlaIOMonad__2_2s.o truncated to GlaIOMonad_
3952 In compiling the compiler proper (in <filename>compiler/</filename>), you <emphasis>may</emphasis>
3953 get an “Out of heap space” error message. These can vary with the
3954 vagaries of different systems, it seems. The solution is simple:
3961 If you're compiling with GHC 4.00 or later, then the
3962 <emphasis>maximum</emphasis> heap size must have been reached. This
3963 is somewhat unlikely, since the maximum is set to 64M by default.
3964 Anyway, you can raise it with the
3965 <option>-optCrts-M<size></option> flag (add this flag to
3966 <constant><module>_HC_OPTS</constant>
3967 <command>make</command> variable in the appropriate
3968 <filename>Makefile</filename>).
3975 For GHC < 4.00, add a suitable <option>-H</option> flag to the <filename>Makefile</filename>, as
3984 and try again: <command>gmake</command>. (see <Xref LinkEnd="sec-suffix"> for information about
3985 <constant><module>_HC_OPTS</constant>.)
3987 Alternatively, just cut to the chase:
3991 % make EXTRA_HC_OPTS=-optCrts-M128M
4000 If you try to compile some Haskell, and you get errors from GCC about
4001 lots of things from <filename>/usr/include/math.h</filename>, then your GCC was
4002 mis-installed. <command>fixincludes</command> wasn't run when it should've been.
4004 As <command>fixincludes</command> is now automagically run as part of GCC installation,
4005 this bug also suggests that you have an old GCC.
4013 You <emphasis>may</emphasis> need to re-<command>ranlib</command><indexterm><primary>ranlib</primary></indexterm> your libraries (on Sun4s).
4017 % cd $(libdir)/ghc-x.xx/sparc-sun-sunos4
4018 % foreach i ( `find . -name '*.a' -print` ) # or other-shell equiv...
4020 ? # or, on some machines: ar s $i
4025 We'd be interested to know if this is still necessary.
4033 GHC's sources go through <command>cpp</command> before being compiled, and <command>cpp</command> varies
4034 a bit from one Unix to another. One particular gotcha is macro calls
4039 SLIT("Hello, world")
4043 Some <command>cpp</command>s treat the comma inside the string as separating two macro
4044 arguments, so you get
4048 :731: macro `SLIT' used with too many (2) args
4052 Alas, <command>cpp</command> doesn't tell you the offending file!
4054 Workaround: don't put weird things in string args to <command>cpp</command> macros.
4065 <Sect1 id="winbuild"><Title>Notes for building under Windows</Title>
4068 This section summarises how to get the utilities you need on your
4069 Win95/98/NT/2000 machine to use CVS and build GHC. Similar notes for
4070 installing and running GHC may be found in the user guide. In general,
4071 Win95/Win98 behave the same, and WinNT/Win2k behave the same.
4072 You should read the GHC installation guide sections on Windows (in the user
4073 guide) before continuing to read these notes.
4077 <sect2><Title>Cygwin and MinGW</Title>
4079 <para> The Windows situation for building GHC is rather confusing. This section
4080 tries to clarify, and to establish terminology.</para>
4082 <sect3 id="ghc-mingw"><title>GHC-mingw</title>
4084 <para> <ulink url="http://www.mingw.org">MinGW (Minimalist GNU for Windows)</ulink>
4085 is a collection of header
4086 files and import libraries that allow one to use <command>gcc</command> and produce
4087 native Win32 programs that do not rely on any third-party DLLs. The
4088 current set of tools include GNU Compiler Collection (<command>gcc</command>), GNU Binary
4089 Utilities (Binutils), GNU debugger (Gdb), GNU make, and a assorted
4092 <para>The GHC that we distribute includes, inside the distribution itself, the MinGW <command>gcc</command>,
4093 <command>as</command>, <command>ld</command>, and a bunch of input/output libraries.
4094 GHC compiles Haskell to C (or to
4095 assembly code), and then invokes these MinGW tools to generate an executable binary.
4096 The resulting binaries can run on any Win32 system.
4098 <para> We will call a GHC that targets MinGW in this way <emphasis>GHC-mingw</emphasis>.</para>
4100 <para> The down-side of GHC-mingw is that the MinGW libraries do not support anything like the full
4101 Posix interface. So programs compiled with GHC-mingw cannot import the (Haskell) Posix
4102 library; they have to do
4103 their input output using standard Haskell I/O libraries, or native Win32 bindings.
4107 <sect3 id="ghc-cygwin"><title>GHC-cygwin</title>
4109 <para>There <emphasis>is</emphasis> a way to get the full Posix interface, which is to use Cygwin.
4110 <ulink url="http://www.cygwin.com">Cygwin</ulink> is a complete Unix simulation that runs on Win32.
4111 Cygwin comes with a shell, and all the usual Unix commands: <command>mv</command>, <command>rm</command>,
4112 <command>ls</command>, plus of course <command>gcc</command>, <command>ld</command> and so on.
4113 A C program compiled with the Cygwin <command>gcc</command> certainly can use all of Posix.
4115 <para>So why doesn't GHC use the Cygwin <command>gcc</command> and libraries? Because
4116 Cygwin comes with a DLL <emphasis>that must be linked with every runnable Cygwin-compiled program</emphasis>.
4117 A program compiled by the Cygwin tools cannot run at all unless Cygwin is installed.
4118 If GHC targeted Cygwin, users would have to install Cygwin just to run the Haskell programs
4119 that GHC compiled; and the Cygwin DLL would have to be in the DLL load path.
4120 Worse, Cygwin is a moving target. The name of the main DLL, <literal>cygwin1.dll</literal>
4121 does not change, but the implementation certainly does. Even the interfaces to functions
4122 it exports seem to change occasionally. So programs compiled by GHC might only run with
4123 particular versions of Cygwin. All of this seems very undesirable.
4126 Nevertheless, it is certainly possible to build a version of GHC that targets Cygwin;
4127 we will call that <emphasis>GHC-cygwin</emphasis>. The up-side of GHC-cygwin is
4128 that Haskell programs compiled by GHC-cygwin can import the (Haskell) Posix library.
4132 <sect3><title>HOST_OS vs TARGET_OS</title>
4135 In the source code you'll find various ifdefs looking like:
4137 #ifdef mingw32_HOST_OS
4143 #ifdef mingw32_TARGET_OS
4147 These macros are set by the configure script (via the file config.h).
4148 Which is which? The criterion is this. In the ifdefs in GHC's source code:
4151 The "host" system is the one on which GHC itself will be run.
4154 The "target" system is the one for which the program compiled by GHC will be run.
4157 For a stage-2 compiler, in which GHCi is available, the "host" and "target" systems must be the same.
4158 So then it doesn't really matter whether you use the HOST_OS or TARGET_OS cpp macros.
4163 <sect3><title>Summary</title>
4165 <para>Notice that "GHC-mingw" means "GHC that <emphasis>targets</emphasis> MinGW". It says nothing about
4166 how that GHC was <emphasis>built</emphasis>. It is entirely possible to have a GHC-mingw that was built
4167 by compiling GHC's Haskell sources with a GHC-cygwin, or vice versa.</para>
4169 <para>We distribute only a GHC-mingw built by a GHC-mingw; supporting
4170 GHC-cygwin too is beyond our resources. The GHC we distribute
4171 therefore does not require Cygwin to run, nor do the programs it
4172 compiles require Cygwin.</para>
4174 <para>The instructions that follow describe how to build GHC-mingw. It is
4175 possible to build GHC-cygwin, but it's not a supported route, and the build system might
4178 <para>In your build tree, you build a compiler called <Command>ghc-inplace</Command>. It
4179 uses the <Command>gcc</Command> that you specify using the
4180 <option>--with-gcc</option> flag when you run
4181 <Command>configure</Command> (see below).
4182 The makefiles are careful to use <Command>ghc-inplace</Command> (not <Command>gcc</Command>)
4183 to compile any C files, so that it will in turn invoke the right <Command>gcc</Command> rather that
4184 whatever one happens to be in your path. However, the makefiles do use whatever <Command>ld</Command>
4185 and <Command>ar</Command> happen to be in your path. This is a bit naughty, but (a) they are only
4186 used to glom together .o files into a bigger .o file, or a .a file,
4187 so they don't ever get libraries (which would be bogus; they might be the wrong libraries), and (b)
4188 Cygwin and Mingw use the same .o file format. So its ok.
4193 <Sect2><Title>Installing and configuring Cygwin</Title>
4195 <para>You don't need Cygwin to <emphasis>use</emphasis> GHC,
4196 but you do need it to <emphasis>build</emphasis> GHC.</para>
4198 <para> Install Cygwin from <ulink url="http://www.cygwin.com/">http://www.cygwin.com/</ulink>.
4199 The installation process is straightforward; we install it in <Filename>c:/cygwin</Filename>.
4200 Both <command>cvs</command> and <command>ssh</command>
4201 come with Cygwin, but you'll need them, so make sure you select them when running
4202 the Cygwin installer.
4205 <para> Now set the following user environment variables:
4208 <listitem><para> Add <filename>c:/cygwin/bin</filename> and <filename>c:/cygwin/usr/bin</filename> to your
4209 <constant>PATH</constant></para></listitem>
4213 Set <constant>MAKE_MODE</constant> to <Literal>UNIX</Literal>. If you
4214 don't do this you get very weird messages when you type
4215 <Command>make</Command>, such as:
4217 /c: /c: No such file or directory
4222 <listitem><para> Set <constant>SHELL</constant> to
4223 <Filename>c:/cygwin/bin/sh</Filename>. When you invoke a shell in Emacs, this
4224 <constant>SHELL</constant> is what you get.
4227 <listitem><para> Set <constant>HOME</constant> to point to your
4228 home directory. This is where, for example,
4229 <command>bash</command> will look for your <filename>.bashrc</filename>
4230 file. Ditto <command>emacs</command> looking for <filename>.emacsrc</filename>
4236 There are a few other things to do:
4240 Some script files used in the make system start with "<Command>#!/bin/perl</Command>",
4241 (and similarly for <Command>bash</Command>). Notice the hardwired path!
4242 So you need to ensure that your <Filename>/bin</Filename> directory has the following
4245 <listitem> <para><Command>sh</Command></para></listitem>
4246 <listitem> <para><Command>perl</Command></para></listitem>
4247 <listitem> <para><Command>cat</Command></para></listitem>
4249 All these come in Cygwin's <Filename>bin</Filename> directory, which you probably have
4250 installed as <Filename>c:/cygwin/bin</Filename>. By default Cygwin mounts "<Filename>/</Filename>" as
4251 <Filename>c:/cygwin</Filename>, so if you just take the defaults it'll all work ok.
4252 (You can discover where your Cygwin
4253 root directory <Filename>/</Filename> is by typing <Command>mount</Command>).
4254 Provided <Filename>/bin</Filename> points to the Cygwin <Filename>bin</Filename>
4255 directory, there's no need to copy anything.
4261 By default, cygwin provides the command shell <filename>ash</filename>
4262 as <filename>sh.exe</filename>. It has a couple of 'issues' (to do with quoting
4263 and length of command lines), so
4264 in your <filename>/bin</filename> directory, make sure that <filename>
4265 bash.exe</filename> is also provided as <filename>sh.exe</filename>
4266 (i.e. overwrite the old <filename>sh.exe</filename> with a copy of
4267 <filename>bash.exe</filename>).
4273 <para>Finally, here are some things to be aware of when using Cygwin:
4275 <listitem> <para>Cygwin doesn't deal well with filenames that include
4276 spaces. "<filename>Program Files</filename>" and "<filename>Local files</filename>" are
4280 <listitem> <para> Cygwin implements a symbolic link as a text file with some
4281 magical text in it. So other programs that don't use Cygwin's
4282 I/O libraries won't recognise such files as symlinks.
4283 In particular, programs compiled by GHC are meant to be runnable
4284 without having Cygwin, so they don't use the Cygwin library, so
4285 they don't recognise symlinks.
4289 Win32 has a <command>find</command> command which is not the same as Cygwin's find.
4290 You will probably discover that the Win32 <command>find</command> appears in your <constant>PATH</constant>
4291 before the Cygwin one, because it's in the <emphasis>system</emphasis> <constant>PATH</constant>
4292 environment variable, whereas you have probably modified the <emphasis>user</emphasis> <constant>PATH</constant>
4293 variable. You can always invoke <command>find</command> with an absolute path, or rename it.
4300 <Sect2><Title>Other things you need to install</Title>
4302 <para>You have to install the following other things to build GHC:
4306 Install an executable GHC, from <ulink url="http://www.haskell.org/ghc">http://www.haskell.org/ghc</ulink>.
4307 This is what you will use to compile GHC. Add it in your
4308 <constant>PATH</constant>: the installer tells you the path element
4309 you need to add upon completion.
4315 Install an executable Happy, from <ulink url="http://www.haskell.org/happy">http://www.haskell.org/happy</ulink>.
4316 Happy is a parser generator used to compile the Haskell grammar. Add it in your
4317 <constant>PATH</constant>.
4323 <para>GHC uses the <emphasis>mingw</emphasis> C compiler to
4324 generate code, so you have to install that. Just pick up a mingw bundle at
4325 <ulink url="http://www.mingw.org/">http://www.mingw.org/</ulink>.
4326 We install it in <filename>c:/mingw</filename>.
4328 <para>Do <emphasis>not</emphasis> add any of the <emphasis>mingw</emphasis> binaries to your path.
4329 They are only going to get used by explicit access (via the --with-gcc flag you
4330 give to <Command>configure</Command> later). If you do add them to your path
4331 you are likely to get into a mess because their names overlap with Cygwin binaries.
4337 <para> Finally, check out a copy of GHC sources from
4338 the CVS repository, following the instructions above (<xref linkend="cvs-access">).
4345 <Sect2><Title>Building GHC</Title>
4348 Now go read the documentation above on building from source (<xref linkend="sec-building-from-source">);
4349 the bullets below only tell
4350 you about Windows-specific wrinkles.</para>
4354 Run <Command>autoconf</Command> both in <filename>fptools</filename>
4355 and in <filename>fptools/ghc</filename>. If you omit the latter step you'll
4356 get an error when you run <filename>./configure</filename>:
4359 creating mk/config.h
4360 mk/config.h is unchanged
4362 running /bin/sh ./configure --cache-file=.././config.cache --srcdir=.
4363 ./configure: ./configure: No such file or directory
4364 configure: error: ./configure failed for ghc
4369 <listitem> <para><command>autoconf</command> seems to create the file <filename>configure</filename>
4370 read-only. So if you need to run autoconf again (which I sometimes do for safety's sake),
4373 /usr/bin/autoconf: cannot create configure: permission denied
4375 Solution: delete <filename>configure</filename> first.
4380 You either need to add <filename>ghc</filename> to your
4381 <constant>PATH</constant> before you invoke
4382 <Command>configure</Command>, or use the <Command>configure</Command>
4383 option <option>--with-ghc=c:/ghc/ghc-some-version/bin/ghc</option>.
4388 If you are paranoid, delete <filename>config.cache</filename> if it exists.
4389 This file occasionally remembers out-of-date configuration information, which
4390 can be really confusing.
4396 After <command>autoconf</command> run <command>./configure</command> in
4397 <filename>fptools/</filename> thus:
4400 ./configure --host=i386-unknown-mingw32 --with-gcc=c:/mingw/bin/gcc
4402 This is the point at which you specify that you are building GHC-mingw
4403 (see <xref linkend="ghc-mingw">). </para>
4405 <para> Both these options are important! It's possible to get into
4406 trouble using the wrong C compiler!
4407 Furthermore, it's very important that you specify a
4408 full mingw path for <command>gcc</command>, not a cygwin path, because GHC (which
4409 uses this path to invoke <command>gcc</command>) is a Mingw program and won't
4410 understand a cygwin path.. For example, if you
4411 say <literal>--with-gcc=/mingw/bin/gcc</literal>, it'll be interpreted as
4412 <filename>/cygdrive/c/mingw/bin/gcc</filename>, and GHC will fail the first
4413 time it tries to invoke it. (Worse, the failure does not come with
4414 a helpful error message, unfortunately.)
4418 If you want to build GHC-cygwin (<xref linkend="ghc-cygwin">)
4419 you'll have to do something more like:
4421 ./configure --with-gcc=...the Cygwin gcc...
4426 <listitem><para> Do not attempt to build the documentation.
4427 It needs all kinds of wierd Jade stuff that we haven't worked out for
4428 Win32.</para></listitem>