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 Glasgow <Literal>fptools</Literal> suite</title>
35 <para>Building the Glasgow tools <Emphasis>can</Emphasis> be
36 complicated, mostly because there are so many permutations of
37 what/why/how, e.g., ``Build Happy with HBC, everything else with
38 GHC, leave out profiling, and test it all on the `real' NoFib
39 programs.'' Yeeps!</para>
41 <para>Happily, such complications don't apply to most people. A
42 few common ``strategies'' serve most purposes. Pick one and
43 proceed as suggested:</para>
48 <Term><IndexTerm><Primary>Binary distribution</Primary></IndexTerm>Binary distribution.</Term>
51 If your only purpose is to install some of the
52 <Literal>fptools</Literal> suite then the easiest thing to do is to
53 get a binary distribution. In the binary distribution everything is
54 pre-compiled for your particular machine architecture and operating
55 system, so all you should have to do is install the binaries and
56 libraries in suitable places. The user guide describes how to do this.
60 A binary distribution may not work for you for two reasons. First, we
61 may not have built the suite for the particular architecture/OS
62 platform you want. That may be due to lack of time and energy (in
63 which case you can get a source distribution and build from it; see
64 below). Alternatively, it may be because we haven't yet ported the
65 suite to your architecture, in which case you are considerably worse
70 The second reason a binary distribution may not be what you want is
71 if you want to read or modify the souce code.
73 </ListItem></VarListEntry>
75 <Term><IndexTerm><Primary>Source distribution</Primary></IndexTerm>Source distribution.</Term>
79 platform, but (a) you like the warm fuzzy feeling of compiling things
80 yourself; (b) you want to build something ``extra''—e.g., a set of
81 libraries with strictness-analysis turned off; or (c) you want to hack
86 A source distribution contains complete sources for one or more
87 projects in the <Literal>fptools</Literal> suite. Not only that, but
88 the more awkward machine-independent steps are done for you. For
89 example, if you don't have
90 <Command>happy</Command><IndexTerm><Primary>happy</Primary></IndexTerm>
91 you'll find it convenient that the source distribution contains the
92 result of running <Command>happy</Command> on the parser
93 specifications. If you don't want to alter the parser then this saves
94 you having to find and install <Command>happy</Command>. You will
95 still need a working version of GHC (preferably version 4.08+) on your
96 machine in order to compile (most of) the sources, however.
99 </ListItem></VarListEntry>
102 <term>The CVS repository.</term>
103 <indexterm><primary>CVS repository</primary>
106 <para>We make releases infrequently. If you want more
107 up-to-the minute (but less tested) source code then you need
108 to get access to our CVS repository.</para>
110 <para>All the <Literal>fptools</Literal> source code is held
111 in a CVS repository. CVS is a pretty good source-code
112 control system, and best of all it works over the
115 <para>The repository holds source code only. It holds no
116 mechanically generated files at all. So if you check out a
117 source tree from CVS you will need to install every utility
118 so that you can build all the derived files from
121 <para>More information about our CVS repository can be found
122 in <xref linkend="sec-cvs">.</para>
128 <term>Build GHC from intermediate C <Filename>.hc</Filename> files<IndexTerm><Primary>hc files</Primary></IndexTerm>:</term>
130 <para>You need a working GHC to use a source distribution.
131 What if you don't have a working GHC? Then you may be able
132 to bootstrap up from the intermediate C
133 (<filename>.hc</filename>) files that we provide. Building
134 GHC on an unsupported platform falls into this category.
135 Beware: this route is not for the faint hearted! Please see
136 <Xref LinkEnd="sec-booting-from-C">.</para>
138 <para>Once you have built GHC, you can build the other
139 Glasgow tools with it.</para>
141 <para>In theory, you can (could?) build GHC with another
142 Haskell compiler (e.g., HBC). We haven't tried to do this
143 for ages and it almost certainly doesn't work any more (for
144 tedious reasons).</para>
149 <para>If you are going to do any building from sources (either
150 from a source distribution or the CVS repository) then you need to
151 read all of this manual in detail.</para>
155 <title>Using the CVS repository</title>
157 <para>We use <ulink url="http://www.cvshome.org/">CVS</ulink> (Concurrent Version System) to keep track of our
158 sources for various software projects. CVS lets several people
159 work on the same software at the same time, allowing changes to be
160 checked in incrementally. </para>
162 <para>This section is a set of guidelines for how to use our CVS
163 repository, and will probably evolve in time. The main thing to
164 remember is that most mistakes can be undone, but if there's
165 anything you're not sure about feel free to bug the local CVS
166 meister (namely Jeff Lewis
167 <email>jlewis@galconn.com</email>). </para>
169 <sect2 id="cvs-access">
170 <title>Getting access to the CVS Repository</title>
172 <para>You can access the repository in one of two ways:
173 read-only (<xref linkend="cvs-read-only">), or read-write (<xref
174 linkend="cvs-read-write">).</para>
176 <sect3 id="cvs-read-only">
177 <title>Remote Read-only CVS Access</title>
179 <para>Read-only access is available to anyone - there's no
180 need to ask us first. With read-only CVS access you can do
181 anything except commit changes to the repository. You can
182 make changes to your local tree, and still use CVS's merge
183 facility to keep your tree up to date, and you can generate
184 patches using 'cvs diff' in order to send to us for
187 <para>To get read-only access to the repository:</para>
191 <para>Make sure that <application>cvs</application> is
192 installed on your machine.</para>
195 <para>Set your <literal>$CVSROOT</literal> environment variable to
196 <literal>:pserver:anoncvs@glass.cse.ogi.edu:/cvs</literal></para>
199 <para>Run the command</para>
203 <para>The password is simply <literal>cvs</literal>. This
204 sets up a file in your home directory called
205 <literal>.cvspass</literal>, which squirrels away the
206 dummy password, so you only need to do this step once.</para>
210 <para>Now go to <xref linkend="cvs-first">.</para>
215 <sect3 id="cvs-read-write">
216 <title>Remote Read-Write CVS Access</title>
218 <para>We generally supply read-write access to folk doing
219 serious development on some part of the source tree, when
220 going through us would be a pain. If you're developing some
221 feature, or think you have the time and inclination to fix
222 bugs in our sources, feel free to ask for read-write
223 access. There is a certain amount of responsibility that goes
224 with commit privileges; we are more likely to grant you access
225 if you've demonstrated your competence by sending us patches
226 via mail in the past.</para>
228 <para>To get remote read-write CVS access, you need to do the
229 following steps.</para>
233 <para>Make sure that <literal>cvs</literal> and
234 <literal>ssh</literal> are both installed on your
239 <para>Generate a DSA private-key/public-key pair, thus:</para>
243 <para>(<literal>ssh-keygen</literal> comes with
244 <literal>ssh</literal>.) Running <literal>ssh-keygen
245 -d</literal> creates the private and public keys in
246 <literal>$HOME/.ssh/id_dsa</literal> and
247 <literal>$HOME/.ssh/id_dsa.pub</literal> respectively
248 (assuming you accept the standard defaults).</para>
250 <para><literal>ssh-keygen -d</literal> will only work if
251 you have Version 2 <literal>ssh</literal> installed; it
252 will fail harmlessly otherwise. If you only have Version
253 1 you can instead generate an RSA key pair using plain</para>
258 <para>Doing so creates the private and public RSA keys in
259 <literal>$HOME/.ssh/identity</literal> and
260 <literal>$HOME/.ssh/identity.pub</literal>
263 <para>[Deprecated.] Incidentally, you can force a Version
264 2 <literal>ssh</literal> to use the Version 1 protocol by
265 creating <literal>$HOME/config</literal> with the
266 following in it:</para>
274 <para>In both cases, <literal>ssh-keygen</literal> will
275 ask for a <firstterm>passphrase</firstterm>. The
276 passphrase is a password that protects your private key.
277 In response to the 'Enter passphrase' question, you can
281 <para>[Recommended.] Enter a passphrase, which you
282 will quote each time you use CVS.
283 <literal>ssh-agent</literal> makes this entirely
287 <para>[Deprecated.] Just hit return (i.e. use an empty
288 passphrase); then you won't need to quote the
289 passphrase when using CVS. The downside is that
290 anyone who can see into your <literal>.ssh</literal>
291 directory, and thereby get your private key, can mess
292 up the repository. So you must keep the
293 <literal>.ssh</literal> directory with draconian
294 no-access permissions.</para>
298 <para>[Windows users. To protect your
299 <literal>.ssh</literal> from access by anyone else,
300 right-click your <literal>.ssh</literal> directory, and
301 select <literal>Properties</literal>. If you are not on
302 the access control list, add yourself, and give yourself
303 full permissions (the second panel). Remove everyone else
304 from the access control list. Don't leave them there but
305 deny them access, because 'they' may be a list that
306 includes you!]</para>
310 <para>Send a message to to the CVS repository
311 administrator (currently Jeff Lewis
312 <email>jeff@galconn.com</email>), containing:</para>
315 <para>Your desired user-name.</para>
318 <para>Your <literal>.ssh/id_dsa.pub</literal> (or
319 <literal>.ssh/identity.pub</literal>).</para>
322 <para>He will set up your account.</para>
326 <para>Set the following environment variables:</para>
329 <para><literal>$CVS_RSH</literal> to
330 <literal>ssh</literal></para>
333 <para><literal>$CVSROOT</literal> to
334 <literal>:ext:</literal><replaceable>your-username</replaceable><literal>@cvs.haskell.org:/home/cvs/root</literal></para>
341 <para>The <literal>CVSROOT</literal> environment variable will
342 be recorded in the checked-out tree, so you don't need to set
343 this every time either. Ignore the instructions for setting
344 <literal>CVSROOT</literal> below. </para>
346 <para>Caveats:</para>
350 <para>Setting your <literal>CVS_RSH</literal> to
351 <literal>ssh</literal> assumes that your CVS client
352 understands how to execute shell script
353 ("#!"s,really), which is what
354 <literal>ssh</literal> is. This may not be the case on
355 some platforms (read: Win32), so in that case set
356 <literal>CVS_RSH</literal> to
357 <literal>ssh1</literal>.</para>
361 <para>[Experts.] Once your account is set up, you can get
362 access from other machines without bothering Jeff, thus:</para>
365 <para>Generate a public/private key pair on the new
369 <para>Use ssh to log in to
370 <literal>cvs.haskell.org</literal>, from your old
374 <para>Add the public key for the new machine to the file
375 <literal>$HOME/ssh/authorized_keys</literal> on
376 <literal>cvs.haskell.org</literal>.
377 (<literal>authorized_keys2</literal>, I think, for Version
381 <para>Make sure that the new version of
382 <literal>authorized_keys</literal> still has 600 file
389 <sect2 id="cvs-first">
390 <title>Checking Out a Source Tree</title>
394 <para>Make sure you set your <literal>CVSROOT</literal>
395 environment variable according to either of the remote
396 methods above. The Approved Way to check out a source tree
397 is as follows:</para>
400 $ cvs checkout fpconfig
403 <para>At this point you have a new directory called
404 <literal>fptools</literal> which contains the basic stuff
405 for the fptools suite, including the configuration files and
406 some other junk. </para>
409 $ mv fptools <replaceable>directory</replaceable>
412 <para>You can call the fptools directory whatever you like,
413 CVS won't mind. </para>
415 <para> NB: after you've read the CVS manual you might be
416 tempted to try</para>
418 $ cvs checkout -d <replaceable>directory</replaceable> fpconfig
421 <para>instead of checking out <literal>fpconfig</literal>
422 and then renaming it. But this doesn't work, and will
423 result in checking out the entire repository instead of just
424 the <literal>fpconfig</literal> bit.</para>
426 $ cd <replaceable>directory</replaceable>
427 $ cvs checkout ghc hslibs
430 <para>The second command here checks out the relevant
431 modules you want to work on. For a GHC build, for instance,
432 you need at least the <literal>ghc</literal> and
433 <literal>hslibs</literal> modules (for a full list of the
434 projects available, see <xref linkend="projects">).</para>
439 <sect2 id="cvs-committing">
440 <title>Committing Changes</title>
442 <para>This is only if you have read-write access to the
443 repository. For anoncvs users, CVS will issue a "read-only
444 repository" error if you try to commit changes.</para>
448 <para>Build the software, if necessary. Unless you're just
449 working on documentation, you'll probably want to build the
450 software in order to test any changes you make.</para>
454 <para>Make changes. Preferably small ones first.</para>
458 <para>Test them. You can see exactly what changes you've
459 made by using the <literal>cvs diff</literal> command:</para>
463 <para>lists all the changes (using the
464 <literal>diff</literal> command) in and below the current
465 directory. In emacs, <literal>C-c C-v =</literal> runs
466 <literal>cvs diff</literal> on the current buffer and shows
467 you the results.</para>
471 <para>Before checking in a change, you need to update your
478 <para>This pulls in any changes that other people have made,
479 and merges them with yours. If there are any conflicts, CVS
480 will tell you, and you'll have to resolve them before you
481 can check your changes in. The documentation describes what
482 to do in the event of a conflict.</para>
484 <para>It's not always necessary to do a full cvs update
485 before checking in a change, since CVS will always tell you
486 if you try to check in a file that someone else has changed.
487 However, you should still update at regular intervals to
488 avoid making changes that don't work in conjuction with
489 changes that someone else made. Keeping an eye on what goes
490 by on the mailing list can help here.</para>
494 <para>When you're happy that your change isn't going to
495 break anything, check it in. For a one-file change:</para>
498 $ cvs commit <replaceable>filename</replaceable>
501 <para>CVS will then pop up an editor for you to enter a
502 "commit message", this is just a short description
503 of what your change does, and will be kept in the history of
506 <para>If you're using emacs, simply load up the file into a
507 buffer and type <literal>C-x C-q</literal>, and emacs will
508 prompt for a commit message and then check in the file for
511 <para>For a multiple-file change, things are a bit
512 trickier. There are several ways to do this, but this is the
513 way I find easiest. First type the commit message into a
514 temporary file. Then either</para>
517 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>file_1</replaceable> .... <replaceable>file_n</replaceable>
520 <para>or, if nothing else has changed in this part of the
524 $ cvs commit -F <replaceable>commit-message</replaceable> <replaceable>directory</replaceable>
527 <para>where <replaceable>directory</replaceable> is a common
528 parent directory for all your changes, and
529 <replaceable>commit-message</replaceable> is the name of the
530 file containing the commit message.</para>
532 <para>Shortly afterwards, you'll get some mail from the
533 relevant mailing list saying which files changed, and giving
534 the commit message. For a multiple-file change, you should
535 still get only <emphasis>one</emphasis> message.</para>
540 <sect2 id="cvs-update">
541 <title>Updating Your Source Tree</title>
543 <para>It can be tempting to cvs update just part of a source
544 tree to bring in some changes that someone else has made, or
545 before committing your own changes. This is NOT RECOMMENDED!
546 Quite often changes in one part of the tree are dependent on
547 changes in another part of the tree (the
548 <literal>mk/*.mk</literal> files are a good example where
549 problems crop up quite often). Having an inconsistent tree is a
550 major cause of headaches. </para>
552 <para>So, to avoid a lot of hassle, follow this recipe for
553 updating your tree: </para>
557 $ cvs update -Pd 2>&1 | tee log</screen>
559 <para>Look at the log file, and fix any conflicts (denoted by a
560 <quote>C</quote> in the first column). If you're using multiple
561 build trees, then for every build tree you have pointing at this
562 source tree, you need to update the links in case any new files
563 have appeared: </para>
566 $ cd <replaceable>build-tree</replaceable>
567 $ lndir <replaceable>source-tree</replaceable>
570 <para>Some files might have been removed, so you need to remove
571 the links pointing to these non-existent files:</para>
574 $ find . -xtype l -exec rm '{}' \;
577 <para>To be <emphasis>really</emphasis> safe, you should do
580 <screen>$ gmake all</screen>
582 <para>from the top-level, to update the dependencies and build
583 any changed files. </para>
586 <sect2 id="cvs-tags">
587 <title>GHC Tag Policy</title>
589 <para>If you want to check out a particular version of GHC,
590 you'll need to know how we tag versions in the repository. The
591 policy (as of 4.04) is:</para>
595 <para>The tree is branched before every major release. The
596 branch tag is <literal>ghc-x-xx-branch</literal>, where
597 <literal>x-xx</literal> is the version number of the release
598 with the <literal>'.'</literal> replaced by a
599 <literal>'-'</literal>. For example, the 4.04 release lives
600 on <literal>ghc-4-04-branch</literal>.</para>
604 <para>The release itself is tagged with
605 <literal>ghc-x-xx</literal> (on the branch). eg. 4.06 is
606 called <literal>ghc-4-06</literal>.</para>
610 <para>We didn't always follow these guidelines, so to see
611 what tags there are for previous versions, do <literal>cvs
612 log</literal> on a file that's been around for a while (like
613 <literal>fptools/ghc/README</literal>).</para>
617 <para>So, to check out a fresh GHC 4.06 tree you would
621 $ cvs co -r ghc-4-06 fpconfig
623 $ cvs co -r ghc-4-06 ghc hslibs
627 <sect2 id="cvs-hints">
628 <title>General Hints</title>
632 <para>As a general rule: commit changes in small units,
633 preferably addressing one issue or implementing a single
634 feature. Provide a descriptive log message so that the
635 repository records exactly which changes were required to
636 implement a given feature/fix a bug. I've found this
637 <emphasis>very</emphasis> useful in the past for finding out
638 when a particular bug was introduced: you can just wind back
639 the CVS tree until the bug disappears.</para>
643 <para>Keep the sources at least *buildable* at any given
644 time. No doubt bugs will creep in, but it's quite easy to
645 ensure that any change made at least leaves the tree in a
646 buildable state. We do nightly builds of GHC to keep an eye
647 on what things work/don't work each day and how we're doing
648 in relation to previous verions. This idea is truely wrecked
649 if the compiler won't build in the first place!</para>
653 <para>To check out extra bits into an already-checked-out
654 tree, use the following procedure. Suppose you have a
655 checked-out fptools tree containing just ghc, and you want
656 to add nofib to it:</para>
667 $ cvs update -d nofib
670 <para>(the -d flag tells update to create a new
671 directory). If you just want part of the nofib suite, you
676 $ cvs checkout nofib/spectral
679 <para>This works because <literal>nofib</literal> is a
680 module in its own right, and spectral is a subdirectory of
681 the nofib module. The path argument to checkout must always
682 start with a module name. There's no equivalent form of this
683 command using <literal>update</literal>.</para>
689 <sect1 id="projects">
690 <title>What projects are there?</title>
692 <para>The <literal>fptools</literal> suite consists of several
693 <firstterm>projects</firstterm>, most of which can be downloaded,
694 built and installed individually. Each project corresponds to a
695 subdirectory in the source tree, and if checking out from CVS then
696 each project can be checked out individually by sitting in the top
697 level of your source tree and typing <command>cvs checkout
698 <replaceable>project</replaceable></command>.</para>
700 <para>Here is a list of the projects currently available:</para>
704 <term><literal>ghc</literal></term>
705 <indexterm><primary><literal>ghc</literal></primary>
706 <secondary>project</secondary></indexterm>
708 <para>The <ulink url="http://www.haskell.org/ghc/">Glasgow
709 Haskell Compiler</ulink> (minus libraries). Absolutely
710 required for building GHC.</para>
715 <term><literal>glafp-utils</literal></term>
716 <indexterm><primary><literal>glafp-utils</literal></primary><secondary>project</secondary></indexterm>
718 <para>Utility programs, some of which are used by the
719 build/installation system. Required for pretty much
725 <term><literal>green-card</literal></term>
726 <indexterm><primary><literal>green-card</literal></primary><secondary>project</secondary></indexterm>
729 url="http://www.haskell.org/greencard/">Green Card</ulink>
730 system for generating Haskell foreign function
736 <term><literal>haggis</literal></term>
737 <indexterm><primary><literal>haggis</literal></primary><secondary>project</secondary></indexterm>
740 url="http://www.dcs.gla.ac.uk/fp/software/haggis/">Haggis</ulink>
741 Haskell GUI framework.</para>
746 <term><literal>happy</literal></term>
747 <indexterm><primary><literal>happy</literal></primary><secondary>project</secondary></indexterm>
750 url="http://www.haskell.org/happy/">Happy</ulink> Parser
756 <term><literal>hdirect</literal></term>
757 <indexterm><primary><literal>hdirect</literal></primary><secondary>project</secondary></indexterm>
760 url="http://www.haskell.org/hdirect/">H/Direct</ulink>
761 Haskell interoperability tool.</para>
766 <term><literal>hood</literal></term>
767 <indexterm><primary><literal>hood</literal></primary><secondary>project</secondary></indexterm>
769 <para>The <ulink url="http://www.haskell.org/hood/">Haskell
770 Object Observation Debugger</ulink>.</para>
775 <term><literal>hslibs</literal></term>
776 <indexterm><primary><literal>hslibs</literal></primary><secondary>project</secondary></indexterm>
778 <para>GHC's libraries. Required for building GHC.</para>
783 <term><literal>libraries</literal></term>
784 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
786 <para>Hierarchical Haskell library suite
787 (experimental).</para>
792 <term><literal>mhms</literal></term>
793 <indexterm><primary><literal></literal></primary><secondary>project</secondary></indexterm>
795 <para>The Modular Haskell Metric System.</para>
800 <term><literal>nofib</literal></term>
801 <indexterm><primary><literal>nofib</literal></primary><secondary>project</secondary></indexterm>
803 <para>The NoFib suite: A collection of Haskell programs used
804 primarily for benchmarking.</para>
809 <term><literal>testsuite</literal></term>
810 <indexterm><primary><literal>testsuite</literal></primary><secondary>project</secondary></indexterm>
812 <para>A testing framework, including GHC's regression test
818 <para>So, to build GHC you need at least the
819 <literal>ghc</literal> and <literal>hslibs</literal> projects (a
820 GHC source distribution will already include the bits you
824 <sect1 id="sec-build-checks">
825 <title>Things to check before you start</title>
827 <para>Here's a list of things to check before you get
833 <indexterm><primary>Disk space needed</primary></indexterm>
834 <para>Disk space needed: from about 100Mb for a basic GHC
835 build, up to probably 500Mb for a GHC build with everything
836 included (libraries built several different ways,
841 <para>Use an appropriate machine, compilers, and things.
842 SPARC boxes, PCs running Linux or FreeBSD, and Alphas running
843 OSF/1 are all fully supported. Win32 and HP boxes are in
844 pretty good shape. PCs running Solaris, DEC Alphas running
845 Linux or some BSD variant, MIPS and AIX boxes will need some
846 minimal porting effort before they work (as of 4.06). <xref
847 linkend="sec-port-info"> gives the full run-down on ports or
852 <para>Be sure that the ``pre-supposed'' utilities are
853 installed. <Xref LinkEnd="sec-pre-supposed">
858 <para>If you have any problem when building or installing the
859 Glasgow tools, please check the ``known pitfalls'' (<Xref
860 LinkEnd="sec-build-pitfalls">). Also check the FAQ for the
861 version you're building, which should be available from the
862 relevant download page on the <ULink
863 URL="http://www.haskell.org/ghc/" >GHC web
866 <indexterm><primary>known bugs</primary></indexterm>
867 <indexterm><primary>bugs, known</primary></indexterm>
869 <para>If you feel there is still some shortcoming in our
870 procedure or instructions, please report it.</para>
872 <para>For GHC, please see the bug-reporting section of the GHC
873 Users' Guide (separate document), to maximise the usefulness
874 of your report.</para>
876 <indexterm><primary>bugs</primary><secondary>seporting</secondary></indexterm>
878 <para>If in doubt, please send a message to
879 <email>glasgow-haskell-bugs@haskell.org</email>.
880 <indexterm><primary>bugs</primary><secondary>mailing
881 list</secondary></indexterm></para>
887 <Sect1 id="sec-port-info">
888 <Title>What machines the Glasgow tools run on
892 <IndexTerm><Primary>ports, GHC</Primary></IndexTerm>
893 <IndexTerm><Primary>GHC ports</Primary></IndexTerm>
894 <IndexTerm><Primary>supported platforms</Primary></IndexTerm>
895 <IndexTerm><Primary>platforms, supported</Primary></IndexTerm>
896 The main question is whether or not the Haskell compiler (GHC) runs on
901 A ``platform'' is a architecture/manufacturer/operating-system
902 combination, such as <Literal>sparc-sun-solaris2</Literal>. Other common ones are
903 <Literal>alpha-dec-osf2</Literal>, <Literal>hppa1.1-hp-hpux9</Literal>, <Literal>i386-unknown-linux</Literal>,
904 <Literal>i386-unknown-solaris2</Literal>, <Literal>i386-unknown-freebsd</Literal>,
905 <Literal>i386-unknown-cygwin32</Literal>, <Literal>m68k-sun-sunos4</Literal>, <Literal>mips-sgi-irix5</Literal>,
906 <Literal>sparc-sun-sunos4</Literal>, <Literal>sparc-sun-solaris2</Literal>, <Literal>powerpc-ibm-aix</Literal>.
910 Bear in mind that certain ``bundles'', e.g. parallel Haskell, may not
911 work on all machines for which basic Haskell compiling is supported.
915 Some libraries may only work on a limited number of platforms; for
916 example, a sockets library is of no use unless the operating system
917 supports the underlying BSDisms.
921 <Title>What platforms the Haskell compiler (GHC) runs on</Title>
924 <IndexTerm><Primary>fully-supported platforms</Primary></IndexTerm>
925 <IndexTerm><Primary>native-code generator</Primary></IndexTerm>
926 <IndexTerm><Primary>registerised ports</Primary></IndexTerm>
927 <IndexTerm><Primary>unregisterised ports</Primary></IndexTerm>
928 The GHC hierarchy of Porting Goodness: (a) Best is a native-code
929 generator; (b) next best is a ``registerised''
930 port; (c) the bare minimum is an ``unregisterised'' port.
931 (``Unregisterised'' is so terrible that we won't say more about it).
935 We use Sparcs running Solaris 2.7 and x86 boxes running FreeBSD and
936 Linux, so those are the best supported platforms, unsurprisingly.
940 Here's everything that's known about GHC ports. We identify platforms
941 by their ``canonical'' CPU/Manufacturer/OS triple.
947 <term>alpha-dec-{osf,linux,freebsd,openbsd,netbsd}:</term>
948 <indexterm><primary>alpha-dec-osf</primary></indexterm>
949 <indexterm><primary>alpha-dec-linux</primary></indexterm>
950 <indexterm><primary>alpha-dec-freebsd</primary></indexterm>
951 <indexterm><primary>alpha-dec-openbsd</primary></indexterm>
952 <indexterm><primary>alpha-dec-netbsd</primary></indexterm>
955 <para>The OSF port is currently working (as of GHC version
956 5.02.1) and well supported. The native code generator is
957 currently non-working. Other operating systems will
958 require some minor porting.</para>
963 <term>sparc-sun-sunos4</term>
964 <indexterm><primary>sparc-sun-sunos4</primary></indexterm>
966 <para>Probably works with minor tweaks, hasn't been tested
972 <term>sparc-sun-solaris2</term>
973 <indexterm><primary>sparc-sun-solaris2</primary></indexterm>
975 <para>Fully supported, including native-code
981 <term>hppa1.1-hp-hpux (HP-PA boxes running HPUX 9.x)</term>
982 <indexterm><primary>hppa1.1-hp-hpux</primary></indexterm>
984 <para>Works registerised. No native-code
990 <term>i386-unknown-linux (PCs running Linux, ELF binary format)</term>
991 <indexterm><primary>i386-*-linux</primary></indexterm>
993 <para>GHC works registerised and has a native code
994 generator. You <Emphasis>must</Emphasis> have GCC 2.7.x
995 or later. NOTE about <literal>glibc</literal> versions:
996 GHC binaries built on a system running <literal>glibc
997 2.0</literal> won't work on a system running
998 <literal>glibc 2.1</literal>, and vice versa. In general,
999 don't expect compatibility between
1000 <literal>glibc</literal> versions, even if the shared
1001 library version hasn't changed.</para>
1006 <term>i386-unknown-freebsd (PCs running FreeBSD 2.2
1008 <indexterm><primary>i386-unknown-freebsd</primary></indexterm>
1010 <para>GHC works registerised. Pre-built packages are
1011 available in the native package format, so if you just
1012 need binaries you're better off just installing the
1018 <term>i386-unknown-{netbsd,openbsd) (PCs running NetBSD
1020 <indexterm><primary>i386-unknown-netbsd</primary></indexterm>
1021 <indexterm><primary>i386-unknown-openbsd</primary></indexterm>
1023 <para>Will require some minor porting effort, but should
1024 work registerised.</para>
1029 <term>i386-unknown-cygwin32:</term>
1030 <indexterm><primary>i386-unknown-cygwin32</primary></indexterm>
1032 <para>Fully supported under Win9x/NT, including a native
1033 code generator. Requires the <Literal>cygwin32</Literal>
1034 compatibility library and a healthy collection of GNU
1035 tools (i.e., gcc, GNU ld, bash etc.).</para>
1040 <term>mips-sgi-irix5</term>
1041 <indexterm><primary>mips-sgi-irix[5-6]</primary></indexterm>
1043 <para>Port currently doesn't work, needs some minimal
1044 porting effort. As usual, we don't have access to
1045 machines and there hasn't been an overwhelming demand for
1046 this port, but feel free to get in touch.</para>
1051 <term>powerpc-ibm-aix</term>
1052 <indexterm><primary>powerpc-ibm-aix</primary></indexterm>
1054 <para>Port currently doesn't work, needs some minimal
1055 porting effort. As usual, we don't have access to
1056 machines and there hasn't been an overwhelming demand for
1057 this port, but feel free to get in touch.</para>
1062 <term>powerpc-apple-macosx</term>
1063 <indexterm><primary>powerpc-apple-macosx</primary></indexterm>
1065 <para>Works, unregisterised only at the moment.</para>
1070 <para>Various other systems have had GHC ported to them in the
1071 distant past, including various Motorola 68k boxes. The 68k
1072 support still remains, but porting to one of these systems will
1073 certainly be a non-trivial task.</para>
1077 <title>What machines the other tools run on</title>
1079 <para>Unless you hear otherwise, the other tools work if GHC
1085 <Sect1 id="sec-pre-supposed">
1086 <Title>Installing pre-supposed utilities
1088 <IndexTerm><Primary>pre-supposed utilities</Primary></IndexTerm>
1089 <IndexTerm><Primary>utilities, pre-supposed</Primary></IndexTerm></Title>
1092 Here are the gory details about some utility programs you may need;
1093 <Command>perl</Command>, <Command>gcc</Command> and
1094 <command>happy</command> are the only important
1095 ones. (PVM<IndexTerm><Primary>PVM</Primary></IndexTerm> is important
1096 if you're going for Parallel Haskell.) The
1097 <Command>configure</Command><IndexTerm><Primary>configure</Primary></IndexTerm>
1098 script will tell you if you are missing something.
1106 <IndexTerm><Primary>pre-supposed: Perl</Primary></IndexTerm>
1107 <IndexTerm><Primary>Perl, pre-supposed</Primary></IndexTerm>
1110 <Emphasis>You have to have Perl to proceed!</Emphasis>
1111 It is pretty easy to install.
1115 Perl 5 is required. For Win32 platforms, you should use the binary
1116 supplied in the InstallShield (copy it to <filename>/bin</filename>).
1117 The Cygwin-supplied Perl seems not to work.
1121 Perl should be put somewhere so that it can be invoked by the
1122 <Literal>#!</Literal> script-invoking mechanism. The full
1123 pathname may need to be less than 32 characters long on some
1127 </ListItem></VarListEntry>
1129 <Term>GNU C (<Command>gcc</Command>):</Term>
1130 <IndexTerm><Primary>pre-supposed: GCC (GNU C compiler)</Primary></IndexTerm>
1131 <IndexTerm><Primary>GCC (GNU C compiler), pre-supposed</Primary></IndexTerm>
1135 We recommend using GCC version 2.95.2 on all platforms. Failing that,
1136 version 2.7.2 is stable on most platforms. Earlier versions of GCC
1137 can be assumed not to work, and versions in between 2.7.2 and 2.95.2
1138 (including <command>egcs</command>) have varying degrees of stability
1139 depending on the platform.
1143 If your GCC dies with ``internal error'' on some GHC source file,
1144 please let us know, so we can report it and get things improved.
1145 (Exception: on iX86 boxes—you may need to fiddle with GHC's
1146 <Option>-monly-N-regs</Option> option; see the User's Guide)
1148 </ListItem></VarListEntry>
1152 <indexterm><primary>Happy</primary></indexterm>
1154 <para>Happy is a parser generator tool for Haskell, and is used to
1155 generate GHC's parsers. Happy is written in Haskell, and is a project
1156 in the CVS repository (<literal>fptools/happy</literal>). It can be
1157 built from source, but bear in mind that you'll need GHC installed in
1158 order to build it. To avoid the chicken/egg problem, install a binary
1159 distribtion of either Happy or GHC to get started. Happy
1160 distributions are available from <ulink
1161 url="http://www.haskell.org/happy/">Happy's Web Page</ulink>.
1167 <Term>Autoconf:</Term>
1168 <IndexTerm><Primary>pre-supposed: Autoconf</Primary></IndexTerm>
1169 <IndexTerm><Primary>Autoconf, pre-supposed</Primary></IndexTerm>
1172 GNU Autoconf is needed if you intend to build from the CVS sources, it
1173 is <Emphasis>not</Emphasis> needed if you just intend to build a
1174 standard source distribution.
1178 Autoconf builds the <Command>configure</Command> script from
1179 <Filename>configure.in</Filename> and <Filename>aclocal.m4</Filename>.
1180 If you modify either of these files, you'll need
1181 <command>autoconf</command> to rebuild <Filename>configure</Filename>.
1184 </ListItem></VarListEntry>
1186 <Term><Command>sed</Command></Term>
1187 <IndexTerm><Primary>pre-supposed: sed</Primary></IndexTerm>
1188 <IndexTerm><Primary>sed, pre-supposed</Primary></IndexTerm>
1191 You need a working <Command>sed</Command> if you are going to build
1192 from sources. The build-configuration stuff needs it. GNU sed
1193 version 2.0.4 is no good! It has a bug in it that is tickled by the
1194 build-configuration. 2.0.5 is OK. Others are probably OK too
1195 (assuming we don't create too elaborate configure scripts.)
1197 </ListItem></VarListEntry>
1202 One <Literal>fptools</Literal> project is worth a quick note at this
1203 point, because it is useful for all the others:
1204 <Literal>glafp-utils</Literal> contains several utilities which aren't
1205 particularly Glasgow-ish, but Occasionally Indispensable. Like
1206 <Command>lndir</Command> for creating symbolic link trees.
1209 <Sect2 id="pre-supposed-gph-tools">
1210 <Title>Tools for building parallel GHC (GPH)
1217 <Term>PVM version 3:</Term>
1218 <IndexTerm><Primary>pre-supposed: PVM3 (Parallel Virtual Machine)</Primary></IndexTerm>
1219 <IndexTerm><Primary>PVM3 (Parallel Virtual Machine), pre-supposed</Primary></IndexTerm>
1223 PVM is the Parallel Virtual Machine on which Parallel Haskell programs
1224 run. (You only need this if you plan to run Parallel Haskell.
1225 Concurent Haskell, which runs concurrent threads on a uniprocessor
1226 doesn't need it.) Underneath PVM, you can have (for example) a
1227 network of workstations (slow) or a multiprocessor box (faster).
1231 The current version of PVM is 3.3.11; we use 3.3.7. It is readily
1232 available on the net; I think I got it from
1233 <Literal>research.att.com</Literal>, in <Filename>netlib</Filename>.
1237 A PVM installation is slightly quirky, but easy to do. Just follow
1238 the <Filename>Readme</Filename> instructions.
1240 </ListItem></VarListEntry>
1242 <Term><Command>bash</Command>:</Term>
1243 <IndexTerm><Primary>bash, presupposed (Parallel Haskell only)</Primary></IndexTerm>
1246 Sadly, the <Command>gr2ps</Command> script, used to convert ``parallelism profiles''
1247 to PostScript, is written in Bash (GNU's Bourne Again shell).
1248 This bug will be fixed (someday).
1250 </ListItem></VarListEntry>
1256 <Sect2 id="pre-supposed-doc-tools">
1257 <Title>Tools for building the Documentation
1261 The following additional tools are required if you want to format the
1262 documentation that comes with the <Literal>fptools</Literal> projects:
1269 <Term>DocBook:</Term>
1270 <IndexTerm><Primary>pre-supposed: DocBook</Primary></IndexTerm>
1271 <IndexTerm><Primary>DocBook, pre-supposed</Primary></IndexTerm>
1274 All our documentation is written in SGML, using the DocBook DTD.
1275 Instructions on installing and configuring the DocBook tools are in the
1276 installation guide (in the GHC user guide).
1279 </ListItem></VarListEntry>
1282 <IndexTerm><Primary>pre-supposed: TeX</Primary></IndexTerm>
1283 <IndexTerm><Primary>TeX, pre-supposed</Primary></IndexTerm>
1286 A decent TeX distribution is required if you want to produce printable
1287 documentation. We recomment teTeX, which includes just about
1288 everything you need.
1290 </ListItem></VarListEntry>
1295 In order to actually build any documentation, you need to set
1296 <constant>SGMLDocWays</constant> in your
1297 <filename>build.mk</filename>. Valid values to add to this
1298 list are: <literal>dvi</literal>, <literal>ps</literal>,
1299 <literal>pdf</literal>, <literal>html</literal>, and
1300 <literal>rtf</literal>.
1305 <Sect2 id="pre-supposed-other-tools">
1306 <Title>Other useful tools
1312 <IndexTerm><Primary>pre-supposed: flex</Primary></IndexTerm>
1313 <IndexTerm><Primary>flex, pre-supposed</Primary></IndexTerm>
1317 This is a quite-a-bit-better-than-Lex lexer. Used to build a couple
1318 of utilities in <Literal>glafp-utils</Literal>. Depending on your
1319 operating system, the supplied <Command>lex</Command> may or may not
1320 work; you should get the GNU version.
1322 </ListItem></VarListEntry>
1329 <Sect1 id="sec-building-from-source">
1330 <Title>Building from source
1332 <IndexTerm><Primary>Building from source</Primary></IndexTerm>
1333 <IndexTerm><Primary>Source, building from</Primary></IndexTerm></Title>
1336 You've been rash enough to want to build some of
1337 the Glasgow Functional Programming tools (GHC, Happy,
1338 nofib, etc.) from source. You've slurped the source,
1339 from the CVS repository or from a source distribution, and
1340 now you're sitting looking at a huge mound of bits, wondering
1345 Gingerly, you type <Command>make</Command>. Wrong already!
1349 This rest of this guide is intended for duffers like me, who aren't
1350 really interested in Makefiles and systems configurations, but who
1351 need a mental model of the interlocking pieces so that they can make
1352 them work, extend them consistently when adding new software, and lay
1353 hands on them gently when they don't work.
1356 <Sect2 id="sec-source-tree">
1357 <Title>Your source tree
1361 The source code is held in your <Emphasis>source tree</Emphasis>.
1362 The root directory of your source tree <Emphasis>must</Emphasis>
1363 contain the following directories and files:
1372 <Filename>Makefile</Filename>: the root Makefile.
1378 <Filename>mk/</Filename>: the directory that contains the
1379 main Makefile code, shared by all the
1380 <Literal>fptools</Literal> software.
1386 <Filename>configure.in</Filename>, <Filename>config.sub</Filename>, <Filename>config.guess</Filename>:
1387 these files support the configuration process.
1393 <Filename>install-sh</Filename>.
1402 All the other directories are individual <Emphasis>projects</Emphasis> of the
1403 <Literal>fptools</Literal> system—for example, the Glasgow Haskell Compiler
1404 (<Literal>ghc</Literal>), the Happy parser generator (<Literal>happy</Literal>), the <Literal>nofib</Literal> benchmark
1405 suite, and so on. You can have zero or more of these. Needless to
1406 say, some of them are needed to build others.
1410 The important thing to remember is that even if you want only one
1411 project (<Literal>happy</Literal>, say), you must have a source tree whose root
1412 directory contains <Filename>Makefile</Filename>, <Filename>mk/</Filename>, <Filename>configure.in</Filename>, and the
1413 project(s) you want (<Filename>happy/</Filename> in this case). You cannot get by with
1414 just the <Filename>happy/</Filename> directory.
1421 <IndexTerm><Primary>build trees</Primary></IndexTerm>
1422 <IndexTerm><Primary>link trees, for building</Primary></IndexTerm></Title>
1425 While you can build a system in the source tree, we don't recommend it.
1426 We often want to build multiple versions of our software
1427 for different architectures, or with different options (e.g. profiling).
1428 It's very desirable to share a single copy of the source code among
1433 So for every source tree we have zero or more <Emphasis>build trees</Emphasis>. Each
1434 build tree is initially an exact copy of the source tree, except that
1435 each file is a symbolic link to the source file, rather than being a
1436 copy of the source file. There are ``standard'' Unix utilities that
1437 make such copies, so standard that they go by different names:
1438 <Command>lndir</Command><IndexTerm><Primary>lndir</Primary></IndexTerm>, <Command>mkshadowdir</Command><IndexTerm><Primary>mkshadowdir</Primary></IndexTerm> are two (If you
1439 don't have either, the source distribution includes sources for the
1440 X11 <Command>lndir</Command>—check out <Filename>fptools/glafp-utils/lndir</Filename>). See <Xref LinkEnd="sec-storysofar"> for a typical invocation.
1444 The build tree does not need to be anywhere near the source tree in
1445 the file system. Indeed, one advantage of separating the build tree
1446 from the source is that the build tree can be placed in a
1447 non-backed-up partition, saving your systems support people from
1448 backing up untold megabytes of easily-regenerated, and
1449 rapidly-changing, gubbins. The golden rule is that (with a single
1450 exception—<XRef LinkEnd="sec-build-config">)
1451 <Emphasis>absolutely everything in the build tree is either a symbolic
1452 link to the source tree, or else is mechanically generated</Emphasis>.
1453 It should be perfectly OK for your build tree to vanish overnight; an
1454 hour or two compiling and you're on the road again.
1458 You need to be a bit careful, though, that any new files you create
1459 (if you do any development work) are in the source tree, not a build tree!
1463 Remember, that the source files in the build tree are <Emphasis>symbolic
1464 links</Emphasis> to the files in the source tree. (The build tree soon
1465 accumulates lots of built files like <Filename>Foo.o</Filename>, as well.) You
1466 can <Emphasis>delete</Emphasis> a source file from the build tree without affecting
1467 the source tree (though it's an odd thing to do). On the other hand,
1468 if you <Emphasis>edit</Emphasis> a source file from the build tree, you'll edit the
1469 source-tree file directly. (You can set up Emacs so that if you edit
1470 a source file from the build tree, Emacs will silently create an
1471 edited copy of the source file in the build tree, leaving the source
1472 file unchanged; but the danger is that you think you've edited the
1473 source file whereas actually all you've done is edit the build-tree
1474 copy. More commonly you do want to edit the source file.)
1478 Like the source tree, the top level of your build tree must be (a
1479 linked copy of) the root directory of the <Literal>fptools</Literal> suite. Inside
1480 Makefiles, the root of your build tree is called
1481 <Constant>$(FPTOOLS_TOP)</Constant><IndexTerm><Primary>FPTOOLS_TOP</Primary></IndexTerm>. In the rest of this document path
1482 names are relative to <Constant>$(FPTOOLS_TOP)</Constant> unless otherwise stated. For
1483 example, the file <Filename>ghc/mk/target.mk</Filename> is actually
1484 <Filename><Constant>$(FPTOOLS_TOP)</Constant>/ghc/mk/target.mk</Filename>.
1489 <Sect2 id="sec-build-config">
1490 <Title>Getting the build you want
1494 When you build <Literal>fptools</Literal> you will be compiling code on a particular
1495 <Emphasis>host platform</Emphasis>, to run on a particular <Emphasis>target platform</Emphasis>
1496 (usually the same as the host platform)<IndexTerm><Primary>platform</Primary></IndexTerm>. The
1497 difficulty is that there are minor differences between different
1498 platforms; minor, but enough that the code needs to be a bit different
1499 for each. There are some big differences too: for a different
1500 architecture we need to build GHC with a different native-code
1505 There are also knobs you can turn to control how the <Literal>fptools</Literal>
1506 software is built. For example, you might want to build GHC optimised
1507 (so that it runs fast) or unoptimised (so that you can compile it fast
1508 after you've modified it. Or, you might want to compile it with
1509 debugging on (so that extra consistency-checking code gets included)
1514 All of this stuff is called the <Emphasis>configuration</Emphasis> of your build.
1515 You set the configuration using a three-step process.
1519 <Term>Step 1: get ready for configuration.</Term>
1521 <para>Change directory to
1522 <Constant>$(FPTOOLS_TOP)</Constant> and
1524 <Command>autoconf</Command><IndexTerm><Primary>autoconf</Primary></IndexTerm>
1525 (with no arguments). This GNU program converts
1526 <Filename><Constant>$(FPTOOLS_TOP)</Constant>/configure.in</Filename>
1527 to a shell script called
1528 <Filename><Constant>$(FPTOOLS_TOP)</Constant>/configure</Filename>.
1531 <para>Some projects, including GHC, have their own
1532 configure script. If there's an
1533 <Constant>$(FPTOOLS_TOP)/<project>/configure.in</Constant>,
1534 then you need to run <command>autoconf</command> in that
1535 directory too.</para>
1537 <para>Both these steps are completely
1538 platform-independent; they just mean that the
1539 human-written file (<Filename>configure.in</Filename>)
1540 can be short, although the resulting shell script,
1541 <Command>configure</Command>, and
1542 <Filename>mk/config.h.in</Filename>, are long.</para>
1544 <para>In case you don't have <Command>autoconf</Command>
1545 we distribute the results, <Command>configure</Command>,
1546 and <Filename>mk/config.h.in</Filename>, with the source
1547 distribution. They aren't kept in the repository,
1553 <term>Step 2: system configuration.</term>
1555 <para>Runs the newly-created
1556 <Command>configure</Command> script, thus:</para>
1559 ./configure <optional><parameter>args</parameter></optional>
1562 <para><Command>configure</Command>'s mission is to
1563 scurry round your computer working out what architecture
1564 it has, what operating system, whether it has the
1565 <Function>vfork</Function> system call, where
1566 <Command>yacc</Command> is kept, whether
1567 <Command>gcc</Command> is available, where various
1568 obscure <Literal>#include</Literal> files are,
1569 whether it's a leap year, and what the systems manager
1570 had for lunch. It communicates these snippets of
1571 information in two ways:</para>
1577 <Filename>mk/config.mk.in</Filename><IndexTerm><Primary>config.mk.in</Primary></IndexTerm>
1579 <Filename>mk/config.mk</Filename><IndexTerm><Primary>config.mk</Primary></IndexTerm>,
1580 substituting for things between
1581 ``<Literal>@</Literal>'' brackets. So,
1582 ``<Literal>@HaveGcc@</Literal>'' will be replaced by
1583 ``<Literal>YES</Literal>'' or
1584 ``<Literal>NO</Literal>'' depending on what
1585 <Command>configure</Command> finds.
1586 <Filename>mk/config.mk</Filename> is included by
1587 every Makefile (directly or indirectly), so the
1588 configuration information is thereby communicated to
1589 all Makefiles.</para>
1593 <para> It translates
1594 <Filename>mk/config.h.in</Filename><IndexTerm><Primary>config.h.in</Primary></IndexTerm>
1596 <Filename>mk/config.h</Filename><IndexTerm><Primary>config.h</Primary></IndexTerm>.
1597 The latter is <Literal>#include</Literal>d by
1598 various C programs, which can thereby make use of
1599 configuration information.</para>
1603 <para><command>configure</command> takes some optional
1604 arguments. Use <literal>./configure --help</literal> to
1605 get a list of the available arguments. Here are some of
1606 the ones you might need:</para>
1610 <term><literal>--with-ghc=<parameter>path</parameter></literal></term>
1611 <indexterm><primary><literal>--with-ghc</literal></primary>
1614 <para>Specifies the path to an installed GHC which
1615 you would like to use. This compiler will be used
1616 for compiling GHC-specific code (eg. GHC itself).
1617 This option <emphasis>cannot</emphasis> be
1618 specified using <filename>build.mk</filename> (see
1619 later), because <command>configure</command> needs
1620 to auto-detect the version of GHC you're using.
1621 The default is to look for a compiler named
1622 <literal>ghc</literal> in your path.</para>
1627 <term><literal>--with-hc=<parameter>path</parameter></literal></term>
1628 <indexterm><primary><literal>--with-hc</literal></primary>
1631 <para>Specifies the path to any installed Haskell
1632 compiler. This compiler will be used for
1633 compiling generic Haskell code. The default is to
1634 use <literal>ghc</literal>.</para>
1639 <term><literal>--with-gcc=<parameter>path</parameter></literal></term>
1640 <indexterm><primary><literal>--with-gcc</literal></primary>
1643 <para>Specifies the path to the installed
1644 GCC. This compiler will be used to compile all C
1645 files, <emphasis>except</emphasis> any generated
1646 by the installed Haskell compiler, which will have
1647 its own idea of which C compiler (if any) to use.
1648 The default is to use <literal>gcc</literal>.</para>
1653 <para><command>configure</command> caches the results of
1654 its run in <Filename>config.cache</Filename>. Quite
1655 often you don't want that; you're running
1656 <Command>configure</Command> a second time because
1657 something has changed. In that case, simply delete
1658 <Filename>config.cache</Filename>.</para>
1663 <Term>Step 3: build configuration.</Term>
1666 Next, you say how this build of <Literal>fptools</Literal> is to differ from the
1667 standard defaults by creating a new file <Filename>mk/build.mk</Filename><IndexTerm><Primary>build.mk</Primary></IndexTerm>
1668 <Emphasis>in the build tree</Emphasis>. This file is the one and only file you edit
1669 in the build tree, precisely because it says how this build differs
1670 from the source. (Just in case your build tree does die, you might
1671 want to keep a private directory of <Filename>build.mk</Filename> files, and use a
1672 symbolic link in each build tree to point to the appropriate one.) So
1673 <Filename>mk/build.mk</Filename> never exists in the source tree—you create one in
1674 each build tree from the template. We'll discuss what to put in it
1677 </ListItem></VarListEntry>
1682 And that's it for configuration. Simple, eh?
1685 <para>What do you put in your build-specific configuration file
1686 <filename>mk/build.mk</filename>? <Emphasis>For almost all
1687 purposes all you will do is put make variable definitions that
1688 override those in</Emphasis>
1689 <filename>mk/config.mk.in</filename>. The whole point of
1690 <filename>mk/config.mk.in</filename>—and its derived
1691 counterpart <filename>mk/config.mk</filename>—is to define
1692 the build configuration. It is heavily commented, as you will
1693 see if you look at it. So generally, what you do is look at
1694 <filename>mk/config.mk.in</filename>, and add definitions in
1695 <filename>mk/build.mk</filename> that override any of the
1696 <filename>config.mk</filename> definitions that you want to
1697 change. (The override occurs because the main boilerplate file,
1698 <filename>mk/boilerplate.mk</filename><IndexTerm><Primary>boilerplate.mk</Primary></IndexTerm>,
1699 includes <filename>build.mk</filename> after
1700 <filename>config.mk</filename>.)</para>
1702 <para>For example, <filename>config.mk.in</filename> contains
1703 the definition:</para>
1706 GhcHcOpts=-O -Rghc-timing
1709 <para>The accompanying comment explains that this is the list of
1710 flags passed to GHC when building GHC itself. For doing
1711 development, it is wise to add <literal>-DDEBUG</literal>, to
1712 enable debugging code. So you would add the following to
1713 <filename>build.mk</filename>:</para>
1715 <para>or, if you prefer,</para>
1718 GhcHcOpts += -DDEBUG
1721 <para>GNU <Command>make</Command> allows existing definitions to
1722 have new text appended using the ``<Literal>+=</Literal>''
1723 operator, which is quite a convenient feature.)</para>
1725 <para>If you want to remove the <literal>-O</literal> as well (a
1726 good idea when developing, because the turn-around cycle gets a
1727 lot quicker), you can just override
1728 <literal>GhcLibHcOpts</literal> altogether:</para>
1731 GhcHcOpts=-DDEBUG -Rghc-timing
1734 <para>When reading <filename>config.mk.in</filename>, remember
1735 that anything between ``@...@'' signs is going to be substituted
1736 by <Command>configure</Command> later. You
1737 <Emphasis>can</Emphasis> override the resulting definition if
1738 you want, but you need to be a bit surer what you are doing.
1739 For example, there's a line that says:</para>
1745 <para>This defines the Make variables <Constant>YACC</Constant>
1746 to the pathname for a <Command>yacc</Command> that
1747 <Command>configure</Command> finds somewhere. If you have your
1748 own pet <Command>yacc</Command> you want to use instead, that's
1749 fine. Just add this line to <filename>mk/build.mk</filename>:</para>
1755 <para>You do not <Emphasis>have</Emphasis> to have a
1756 <filename>mk/build.mk</filename> file at all; if you don't,
1757 you'll get all the default settings from
1758 <filename>mk/config.mk.in</filename>.</para>
1760 <para>You can also use <filename>build.mk</filename> to override
1761 anything that <Command>configure</Command> got wrong. One place
1762 where this happens often is with the definition of
1763 <Constant>FPTOOLS_TOP_ABS</Constant>: this
1764 variable is supposed to be the canonical path to the top of your
1765 source tree, but if your system uses an automounter then the
1766 correct directory is hard to find automatically. If you find
1767 that <Command>configure</Command> has got it wrong, just put the
1768 correct definition in <filename>build.mk</filename>.</para>
1772 <sect2 id="sec-storysofar">
1773 <title>The story so far</title>
1775 <para>Let's summarise the steps you need to carry to get
1776 yourself a fully-configured build tree from scratch.</para>
1780 <para> Get your source tree from somewhere (CVS repository
1781 or source distribution). Say you call the root directory
1782 <filename>myfptools</filename> (it does not have to be
1783 called <filename>fptools</filename>). Make sure that you
1784 have the essential files (see <XRef
1785 LinkEnd="sec-source-tree">).</para>
1790 <para>(Optional) Use <Command>lndir</Command> or
1791 <Command>mkshadowdir</Command> to create a build tree.</para>
1795 $ mkshadowdir . /scratch/joe-bloggs/myfptools-sun4
1798 <para>(N.B. <Command>mkshadowdir</Command>'s first argument
1799 is taken relative to its second.) You probably want to give
1800 the build tree a name that suggests its main defining
1801 characteristic (in your mind at least), in case you later
1806 <para>Change directory to the build tree. Everything is
1807 going to happen there now.</para>
1810 $ cd /scratch/joe-bloggs/myfptools-sun4
1816 <para>Prepare for system configuration:</para>
1822 <para>(You can skip this step if you are starting from a
1823 source distribution, and you already have
1824 <filename>configure</filename> and
1825 <filename>mk/config.h.in</filename>.)</para>
1827 <para>Some projects, including GHC itself, have their own
1828 configure scripts, so it is necessary to run autoconf again
1829 in the appropriate subdirectories. eg:</para>
1832 $ (cd ghc; autoconf)
1837 <para>Do system configuration:</para>
1845 <para>Create the file <filename>mk/build.mk</filename>,
1846 adding definitions for your desired configuration options.</para>
1854 <para>You can make subsequent changes to
1855 <filename>mk/build.mk</filename> as often as you like. You do
1856 not have to run any further configuration programs to make these
1857 changes take effect. In theory you should, however, say
1858 <Command>gmake clean</Command>, <Command>gmake all</Command>,
1859 because configuration option changes could affect
1860 anything—but in practice you are likely to know what's
1865 <Title>Making things</Title>
1867 <para>At this point you have made yourself a fully-configured
1868 build tree, so you are ready to start building real
1871 <para>The first thing you need to know is that <Emphasis>you
1872 must use GNU <Command>make</Command>, usually called
1873 <Command>gmake</Command>, not standard Unix
1874 <Command>make</Command></Emphasis>. If you use standard Unix
1875 <Command>make</Command> you will get all sorts of error messages
1876 (but no damage) because the <Literal>fptools</Literal>
1877 <Command>Makefiles</Command> use GNU <Command>make</Command>'s
1878 facilities extensively.</para>
1880 <para>To just build the whole thing, <command>cd</command> to
1881 the top of your <literal>fptools</literal> tree and type
1882 <command>gmake</command>. This will prepare the tree and build
1883 the various projects in the correct order.</para>
1887 <Sect2 id="sec-standard-targets">
1888 <Title>Standard Targets</title>
1889 <IndexTerm><Primary>targets, standard makefile</Primary></IndexTerm>
1890 <IndexTerm><Primary>makefile targets</Primary></IndexTerm>
1892 <para>In any directory you should be able to make the following:
1897 <Term><Literal>boot</Literal>:</Term>
1899 <para>does the one-off preparation required to get ready for the real
1900 work. Notably, it does <Command>gmake depend</Command> in all
1901 directories that contain programs. It also builds the necessary tools
1902 for compilation to proceed.</para>
1904 <para>Invoking the <literal>boot</literal> target explicitly is not
1905 normally necessary. From the top-level <literal>fptools</literal>
1906 directory, invoking <literal>gmake</literal> causes <literal>gmake
1907 boot all</literal> to be invoked in each of the project
1908 subdirectories, in the order specified by
1909 <literal>$(AllTargets)</literal> in
1910 <literal>config.mk</literal>.</para>
1912 <para>If you're working in a subdirectory somewhere and need to update
1913 the dependencies, <literal>gmake boot</literal> is a good way to do it.</para>
1915 </ListItem></VarListEntry>
1917 <Term><Literal>all</Literal>:</Term>
1920 makes all the final target(s) for this Makefile.
1921 Depending on which directory you are in a ``final target'' may be an
1922 executable program, a library archive, a shell script, or a Postscript
1923 file. Typing <Command>gmake</Command> alone is generally the same as typing <Command>gmake all</Command>.
1925 </ListItem></VarListEntry>
1927 <Term><Literal>install</Literal>:</Term>
1930 installs the things built by <Literal>all</Literal> (except for the documentation). Where does it
1931 install them? That is specified by
1932 <filename>mk/config.mk.in</filename>; you can override it in
1933 <filename>mk/build.mk</filename>, or by running
1934 <command>configure</command> with command-line arguments like
1935 <literal>--bindir=/home/simonpj/bin</literal>; see <literal>./configure
1936 --help</literal> for the full details.
1938 </ListItem></VarListEntry>
1940 <Term><Literal>install-docs</Literal>:</Term>
1943 installs the documentation. Otherwise behaves just like <literal>install</literal>.
1945 </ListItem></VarListEntry>
1947 <Term><Literal>uninstall</Literal>:</Term>
1950 reverses the effect of <Literal>install</Literal>.
1952 </ListItem></VarListEntry>
1955 <Term><Literal>clean</Literal>:</Term>
1958 Delete all files from the current directory that are normally created
1959 by building the program. Don't delete the files that record the
1960 configuration, or files generated by <Command>gmake boot</Command>.
1961 Also preserve files that could be made by building, but normally
1962 aren't because the distribution comes with them.</para>
1963 </ListItem></VarListEntry>
1966 <term><literal>distclean</literal>:</term>
1968 <para>Delete all files from the current directory that are created by
1969 configuring or building the program. If you have unpacked the source
1970 and built the program without creating any other files, <literal>make
1971 distclean</literal> should leave only the files that were in the
1972 distribution.</para>
1977 <term><literal>mostlyclean</literal>:</term>
1979 <para>Like <literal>clean</literal>, but may refrain from deleting a
1980 few files that people normally don't want to recompile.</para>
1985 <Term><Literal>maintainer-clean</Literal>:</Term>
1988 Delete everything from the current directory that can be reconstructed
1989 with this Makefile. This typically includes everything deleted by
1990 <literal>distclean</literal>, plus more: C source files produced by
1991 Bison, tags tables, Info files, and so on.</para>
1993 <para>One exception, however: <literal>make maintainer-clean</literal>
1994 should not delete <filename>configure</filename> even if
1995 <filename>configure</filename> can be remade using a rule in the
1996 <filename>Makefile</filename>. More generally, <literal>make
1997 maintainer-clean</literal> should not delete anything that needs to
1998 exist in order to run <filename>configure</filename> and then begin to
1999 build the program.</para>
2004 <Term><Literal>check</Literal>:</Term>
2009 </ListItem></VarListEntry>
2014 All of these standard targets automatically recurse into
2015 sub-directories. Certain other standard targets do not:
2022 <Term><Literal>configure</Literal>:</Term>
2025 is only available in the root directory
2026 <Constant>$(FPTOOLS_TOP)</Constant>; it has been discussed in <XRef LinkEnd="sec-build-config">.
2028 </ListItem></VarListEntry>
2030 <Term><Literal>depend</Literal>:</Term>
2033 make a <filename>.depend</filename> file in each directory that needs
2034 it. This <filename>.depend</filename> file contains mechanically-generated dependency
2035 information; for example, suppose a directory contains a Haskell
2036 source module <filename>Foo.lhs</filename> which imports another module <Literal>Baz</Literal>.
2037 Then the generated <filename>.depend</filename> file will contain the dependency:
2049 which says that the object file <filename>Foo.o</filename> depends on the interface file
2050 <filename>Baz.hi</filename> generated by compiling module <Literal>Baz</Literal>. The <filename>.depend</filename> file is
2051 automatically included by every Makefile.
2053 </ListItem></VarListEntry>
2055 <Term><Literal>binary-dist</Literal>:</Term>
2058 make a binary distribution. This is the
2059 target we use to build the binary distributions of GHC and Happy.
2061 </ListItem></VarListEntry>
2063 <Term><Literal>dist</Literal>:</Term>
2066 make a source distribution. Note that this target does “make
2067 distclean” as part of its work; don't use it if you want to keep
2070 </ListItem></VarListEntry>
2075 Most <filename>Makefile</filename>s have targets other than these. You can discover them by looking in the <filename>Makefile</filename> itself.
2081 <title>Using a project from the build tree</title>
2083 If you want to build GHC (say) and just use it direct from the build
2084 tree without doing <literal>make install</literal> first, you can run
2085 the in-place driver script:
2086 <filename>ghc/compiler/ghc-inplace</filename>.
2089 <para> Do <emphasis>NOT</emphasis> use
2090 <filename>ghc/compiler/ghc</filename>, or
2091 <filename>ghc/compiler/ghc-5.xx</filename>, as these are the scripts
2092 intended for installation, and contain hard-wired paths to the
2093 installed libraries, rather than the libraries in the build tree.
2097 Happy can similarly be run from the build tree, using
2098 <filename>happy/src/happy-inplace</filename>.
2103 <Title>Fast Making <IndexTerm><Primary>fastmake</Primary></IndexTerm>
2104 <IndexTerm><Primary>dependencies, omitting</Primary></IndexTerm>
2105 <IndexTerm><Primary>FAST, makefile
2106 variable</Primary></IndexTerm></Title>
2109 Sometimes the dependencies get in the way: if you've made a small
2110 change to one file, and you're absolutely sure that it won't affect
2111 anything else, but you know that <Command>make</Command> is going to rebuild everything
2112 anyway, the following hack may be useful:
2124 This tells the make system to ignore dependencies and just build what
2125 you tell it to. In other words, it's equivalent to temporarily
2126 removing the <filename>.depend</filename> file in the current directory (where
2127 <Command>mkdependHS</Command> and friends store their dependency information).
2131 A bit of history: GHC used to come with a <Command>fastmake</Command> script that did
2132 the above job, but GNU make provides the features we need to do it
2133 without resorting to a script. Also, we've found that fastmaking is
2134 less useful since the advent of GHC's recompilation checker (see the
2135 User's Guide section on "Separate Compilation").
2142 <Sect1 id="sec-makefile-arch">
2143 <Title>The <filename>Makefile</filename> architecture
2144 <IndexTerm><Primary>makefile architecture</Primary></IndexTerm></Title>
2147 <Command>make</Command> is great if everything works—you type <Command>gmake install</Command> and
2148 lo! the right things get compiled and installed in the right places.
2149 Our goal is to make this happen often, but somehow it often doesn't;
2150 instead some weird error message eventually emerges from the bowels of
2151 a directory you didn't know existed.
2155 The purpose of this section is to give you a road-map to help you figure
2156 out what is going right and what is going wrong.
2160 <Title>A small project</Title>
2163 To get started, let us look at the <filename>Makefile</filename> for an imaginary small
2164 <Literal>fptools</Literal> project, <Literal>small</Literal>. Each project in <Literal>fptools</Literal> has its own
2165 directory in <Constant>FPTOOLS_TOP</Constant>, so the <Literal>small</Literal> project will have its own
2166 directory <Constant>FPOOLS_TOP/small/</Constant>. Inside the <filename>small/</filename> directory there
2167 will be a <filename>Makefile</filename>, looking something like this:
2171 <IndexTerm><Primary>Makefile, minimal</Primary></IndexTerm>
2174 # Makefile for fptools project "small"
2177 include $(TOP)/mk/boilerplate.mk
2179 SRCS = $(wildcard *.lhs) $(wildcard *.c)
2182 include $(TOP)/target.mk
2188 This <filename>Makefile</filename> has three sections:
2197 The first section includes
2201 One of the most important
2202 features of GNU <Command>make</Command> that we use is the ability for a <filename>Makefile</filename> to
2203 include another named file, very like <Command>cpp</Command>'s <Literal>#include</Literal>
2208 a file of ``boilerplate'' code from the level
2209 above (which in this case will be
2210 <filename><Constant>FPTOOLS_TOP</Constant>/mk/boilerplate.mk</filename><IndexTerm><Primary>boilerplate.mk</Primary></IndexTerm>). As its name
2211 suggests, <filename>boilerplate.mk</filename> consists of a large quantity of standard
2212 <filename>Makefile</filename> code. We discuss this boilerplate in more detail in
2213 <XRef LinkEnd="sec-boiler">.
2214 <IndexTerm><Primary>include, directive in Makefiles</Primary></IndexTerm>
2215 <IndexTerm><Primary>Makefile inclusion</Primary></IndexTerm>
2217 Before the <Literal>include</Literal> statement, you must define the <Command>make</Command> variable
2218 <Constant>TOP</Constant><IndexTerm><Primary>TOP</Primary></IndexTerm> to be the directory containing the <filename>mk</filename> directory in
2219 which the <filename>boilerplate.mk</filename> file is. It is <Emphasis>not</Emphasis> OK to simply say
2223 include ../mk/boilerplate.mk # NO NO NO
2227 Why? Because the <filename>boilerplate.mk</filename> file needs to know where it is, so
2228 that it can, in turn, <Literal>include</Literal> other files. (Unfortunately, when an
2229 <Literal>include</Literal>d file does an <Literal>include</Literal>, the filename is treated relative to
2230 the directory in which <Command>gmake</Command> is being run, not the directory in
2231 which the <Literal>include</Literal>d sits.) In general, <Emphasis>every file <filename>foo.mk</filename>
2232 assumes that <filename><Constant>$(TOP)</Constant>/mk/foo.mk</filename> refers to itself.</Emphasis> It is up to the
2233 <filename>Makefile</filename> doing the <Literal>include</Literal> to ensure this is the case.
2235 Files intended for inclusion in other <filename>Makefile</filename>s are written to have
2236 the following property: <Emphasis>after <filename>foo.mk</filename> is <Literal>include</Literal>d, it leaves
2237 <Constant>TOP</Constant> containing the same value as it had just before the <Literal>include</Literal>
2238 statement</Emphasis>. In our example, this invariant guarantees that the
2239 <Literal>include</Literal> for <filename>target.mk</filename> will look in the same directory as that for
2240 <filename>boilerplate.mk</filename>.
2247 The second section defines the following standard <Command>make</Command>
2248 variables: <Constant>SRCS</Constant><IndexTerm><Primary>SRCS</Primary></IndexTerm> (the source files from which is to be
2249 built), and <Constant>HS_PROG</Constant><IndexTerm><Primary>HS_PROG</Primary></IndexTerm> (the executable binary to be
2250 built). We will discuss in more detail what the ``standard
2251 variables'' are, and how they affect what happens, in <XRef LinkEnd="sec-targets">.
2253 The definition for <Constant>SRCS</Constant> uses the useful GNU <Command>make</Command> construct
2254 <Literal>$(wildcard $pat$)</Literal><IndexTerm><Primary>wildcard</Primary></IndexTerm>, which expands to a list of all
2255 the files matching the pattern <Literal>pat</Literal> in the current directory. In
2256 this example, <Constant>SRCS</Constant> is set to the list of all the <filename>.lhs</filename> and <filename>.c</filename>
2257 files in the directory. (Let's suppose there is one of each,
2258 <filename>Foo.lhs</filename> and <filename>Baz.c</filename>.)
2265 The last section includes a second file of standard code,
2266 called <filename>target.mk</filename><IndexTerm><Primary>target.mk</Primary></IndexTerm>. It contains the rules that tell
2267 <Command>gmake</Command> how to make the standard targets (<Xref LinkEnd="sec-standard-targets">). Why, you ask,
2268 can't this standard code be part of <filename>boilerplate.mk</filename>? Good question.
2269 We discuss the reason later, in <Xref LinkEnd="sec-boiler-arch">.
2271 You do not <Emphasis>have</Emphasis> to <Literal>include</Literal> the <filename>target.mk</filename> file. Instead, you
2272 can write rules of your own for all the standard targets. Usually,
2273 though, you will find quite a big payoff from using the canned rules
2274 in <filename>target.mk</filename>; the price tag is that you have to understand what
2275 canned rules get enabled, and what they do (<Xref LinkEnd="sec-targets">).
2285 In our example <filename>Makefile</filename>, most of the work is done by the two
2286 <Literal>include</Literal>d files. When you say <Command>gmake all</Command>, the following things
2296 <Command>gmake</Command> figures out that the object files are <filename>Foo.o</filename> and
2297 <filename>Baz.o</filename>.
2304 It uses a boilerplate pattern rule to compile <filename>Foo.lhs</filename> to
2305 <filename>Foo.o</filename> using a Haskell compiler. (Which one? That is set in the
2306 build configuration.)
2313 It uses another standard pattern rule to compile <filename>Baz.c</filename> to
2314 <filename>Baz.o</filename>, using a C compiler. (Ditto.)
2321 It links the resulting <filename>.o</filename> files together to make <Literal>small</Literal>,
2322 using the Haskell compiler to do the link step. (Why not use <Command>ld</Command>?
2323 Because the Haskell compiler knows what standard libraries to link in.
2324 How did <Command>gmake</Command> know to use the Haskell compiler to do the link,
2325 rather than the C compiler? Because we set the variable <Constant>HS_PROG</Constant>
2326 rather than <Constant>C_PROG</Constant>.)
2336 All <filename>Makefile</filename>s should follow the above three-section format.
2342 <Title>A larger project</Title>
2345 Larger projects are usually structured into a number of sub-directories,
2346 each of which has its own <filename>Makefile</filename>. (In very large projects, this
2347 sub-structure might be iterated recursively, though that is rare.)
2348 To give you the idea, here's part of the directory structure for
2349 the (rather large) GHC project:
2362 ...source files for documentation...
2365 ...source files for driver...
2368 parser/...source files for parser...
2369 renamer/...source files for renamer...
2376 The sub-directories <filename>docs</filename>, <filename>driver</filename>, <filename>compiler</filename>, and so on, each
2377 contains a sub-component of GHC, and each has its own <filename>Makefile</filename>.
2378 There must also be a <filename>Makefile</filename> in <filename><Constant>$(FPTOOLS_TOP)</Constant>/ghc</filename>. It does most
2379 of its work by recursively invoking <Command>gmake</Command> on the <filename>Makefile</filename>s in the
2380 sub-directories. We say that <filename>ghc/Makefile</filename> is a <Emphasis>non-leaf
2381 <filename>Makefile</filename></Emphasis>, because it does little except organise its children,
2382 while the <filename>Makefile</filename>s in the sub-directories are all <Emphasis>leaf
2383 <filename>Makefile</filename>s</Emphasis>. (In principle the sub-directories might themselves
2384 contain a non-leaf <filename>Makefile</filename> and several sub-sub-directories, but
2385 that does not happen in GHC.)
2389 The <filename>Makefile</filename> in <filename>ghc/compiler</filename> is considered a leaf <filename>Makefile</filename> even
2390 though the <filename>ghc/compiler</filename> has sub-directories, because these sub-directories
2391 do not themselves have <filename>Makefile</filename>s in them. They are just used to structure
2392 the collection of modules that make up GHC, but all are managed by the
2393 single <filename>Makefile</filename> in <filename>ghc/compiler</filename>.
2397 You will notice that <filename>ghc/</filename> also contains a directory <filename>ghc/mk/</filename>. It
2398 contains GHC-specific <filename>Makefile</filename> boilerplate code. More precisely:
2407 <filename>ghc/mk/boilerplate.mk</filename> is included at the top of
2408 <filename>ghc/Makefile</filename>, and of all the leaf <filename>Makefile</filename>s in the
2409 sub-directories. It in turn <Literal>include</Literal>s the main boilerplate file
2410 <filename>mk/boilerplate.mk</filename>.
2418 <filename>ghc/mk/target.mk</filename> is <Literal>include</Literal>d at the bottom of
2419 <filename>ghc/Makefile</filename>, and of all the leaf <filename>Makefile</filename>s in the
2420 sub-directories. It in turn <Literal>include</Literal>s the file <filename>mk/target.mk</filename>.
2430 So these two files are the place to look for GHC-wide customisation
2431 of the standard boilerplate.
2436 <Sect2 id="sec-boiler-arch">
2437 <Title>Boilerplate architecture
2438 <IndexTerm><Primary>boilerplate architecture</Primary></IndexTerm>
2442 Every <filename>Makefile</filename> includes a <filename>boilerplate.mk</filename><IndexTerm><Primary>boilerplate.mk</Primary></IndexTerm> file
2443 at the top, and <filename>target.mk</filename><IndexTerm><Primary>target.mk</Primary></IndexTerm> file at the bottom. In
2444 this section we discuss what is in these files, and why there have to
2445 be two of them. In general:
2454 <filename>boilerplate.mk</filename> consists of:
2460 <Emphasis>Definitions of millions of <Command>make</Command> variables</Emphasis> that
2461 collectively specify the build configuration. Examples:
2462 <Constant>HC_OPTS</Constant><IndexTerm><Primary>HC_OPTS</Primary></IndexTerm>, the options to feed to the Haskell compiler;
2463 <Constant>NoFibSubDirs</Constant><IndexTerm><Primary>NoFibSubDirs</Primary></IndexTerm>, the sub-directories to enable within the
2464 <Literal>nofib</Literal> project; <Constant>GhcWithHc</Constant><IndexTerm><Primary>GhcWithHc</Primary></IndexTerm>, the name of the Haskell
2465 compiler to use when compiling GHC in the <Literal>ghc</Literal> project.
2471 <Emphasis>Standard pattern rules</Emphasis> that tell <Command>gmake</Command> how to construct one
2479 <filename>boilerplate.mk</filename> needs to be <Literal>include</Literal>d at the <Emphasis>top</Emphasis>
2480 of each <filename>Makefile</filename>, so that the user can replace the
2481 boilerplate definitions or pattern rules by simply giving a new
2482 definition or pattern rule in the <filename>Makefile</filename>. <Command>gmake</Command>
2483 simply takes the last definition as the definitive one.
2485 Instead of <Emphasis>replacing</Emphasis> boilerplate definitions, it is also quite
2486 common to <Emphasis>augment</Emphasis> them. For example, a <filename>Makefile</filename> might say:
2494 thereby adding ``<Option>-O</Option>'' to the end of <Constant>SRC_HC_OPTS</Constant><IndexTerm><Primary>SRC_HC_OPTS</Primary></IndexTerm>.
2501 <filename>target.mk</filename> contains <Command>make</Command> rules for the standard
2502 targets described in <Xref LinkEnd="sec-standard-targets">. These rules are selectively included,
2503 depending on the setting of certain <Command>make</Command> variables. These
2504 variables are usually set in the middle section of the
2505 <filename>Makefile</filename> between the two <Literal>include</Literal>s.
2507 <filename>target.mk</filename> must be included at the end (rather than being part of
2508 <filename>boilerplate.mk</filename>) for several tiresome reasons:
2515 <Command>gmake</Command> commits target and dependency lists earlier than
2516 it should. For example, <FIlename>target.mk</FIlename> has a rule that looks like
2521 $(HS_PROG) : $(OBJS)
2522 $(HC) $(LD_OPTS) $< -o $@
2526 If this rule was in <filename>boilerplate.mk</filename> then <Constant>$(HS_PROG)</Constant><IndexTerm><Primary>HS_PROG</Primary></IndexTerm>
2527 and <Constant>$(OBJS)</Constant><IndexTerm><Primary>OBJS</Primary></IndexTerm> would not have their final values at the
2528 moment <Command>gmake</Command> encountered the rule. Alas, <Command>gmake</Command> takes a snapshot
2529 of their current values, and wires that snapshot into the rule. (In
2530 contrast, the commands executed when the rule ``fires'' are only
2531 substituted at the moment of firing.) So, the rule must follow the
2532 definitions given in the <filename>Makefile</filename> itself.
2539 Unlike pattern rules, ordinary rules cannot be overriden or
2540 replaced by subsequent rules for the same target (at least, not without an
2541 error message). Including ordinary rules in <filename>boilerplate.mk</filename> would
2542 prevent the user from writing rules for specific targets in specific cases.
2549 There are a couple of other reasons I've forgotten, but it doesn't
2565 <Sect2 id="sec-boiler">
2566 <Title>The main <filename>mk/boilerplate.mk</filename> file
2568 <IndexTerm><Primary>boilerplate.mk</Primary></IndexTerm></Title>
2571 If you look at <filename><Constant>$(FPTOOLS_TOP)</Constant>/mk/boilerplate.mk</filename> you will find
2572 that it consists of the following sections, each held in a separate
2580 <Term><filename>config.mk</filename><IndexTerm><Primary>config.mk</Primary></IndexTerm></Term>
2583 is the build configuration file we
2584 discussed at length in <Xref LinkEnd="sec-build-config">.
2586 </ListItem></VarListEntry>
2588 <Term><filename>paths.mk</filename><IndexTerm><Primary>paths.mk</Primary></IndexTerm></Term>
2591 defines <Command>make</Command> variables for
2592 pathnames and file lists. In particular, it gives definitions for:
2599 <Term><Constant>SRCS</Constant><IndexTerm><Primary>SRCS</Primary></IndexTerm>:</Term>
2602 all source files in the current directory.
2604 </ListItem></VarListEntry>
2606 <Term><Constant>HS_SRCS</Constant><IndexTerm><Primary>HS_SRCS</Primary></IndexTerm>:</Term>
2609 all Haskell source files in the current directory.
2610 It is derived from <Constant>$(SRCS)</Constant>, so if you override <Constant>SRCS</Constant> with a new value
2611 <Constant>HS_SRCS</Constant> will follow suit.
2613 </ListItem></VarListEntry>
2615 <Term><Constant>C_SRCS</Constant><IndexTerm><Primary>C_SRCS</Primary></IndexTerm>:</Term>
2618 similarly for C source files.
2620 </ListItem></VarListEntry>
2622 <Term><Constant>HS_OBJS</Constant><IndexTerm><Primary>HS_OBJS</Primary></IndexTerm>:</Term>
2625 the <filename>.o</filename> files derived from <Constant>$(HS_SRCS)</Constant>.
2627 </ListItem></VarListEntry>
2629 <Term><Constant>C_OBJS</Constant><IndexTerm><Primary>C_OBJS</Primary></IndexTerm>:</Term>
2632 similarly for <Constant>$(C_SRCS)</Constant>.
2634 </ListItem></VarListEntry>
2636 <Term><Constant>OBJS</Constant><IndexTerm><Primary>OBJS</Primary></IndexTerm>:</Term>
2639 the concatenation of <Constant>$(HS_OBJS)</Constant> and <Constant>$(C_OBJS)</Constant>.
2641 </ListItem></VarListEntry>
2646 Any or all of these definitions can easily be overriden by giving new
2647 definitions in your <filename>Makefile</filename>. For example, if there are things in
2648 the current directory that look like source files but aren't, then
2649 you'll need to set <Constant>SRCS</Constant> manually in your <filename>Makefile</filename>. The other
2650 definitions will then work from this new definition.
2654 What, exactly, does <filename>paths.mk</filename> consider a ``source file'' to be? It's
2655 based on the file's suffix (e.g. <filename>.hs</filename>, <filename>.lhs</filename>, <filename>.c</filename>, <filename>.lc</filename>, etc), but
2656 this is the kind of detail that changes, so rather than
2657 enumerate the source suffices here the best thing to do is to look in
2658 <filename>paths.mk</filename>.
2660 </ListItem></VarListEntry>
2662 <Term><filename>opts.mk</filename><IndexTerm><Primary>opts.mk</Primary></IndexTerm></Term>
2665 defines <Command>make</Command> variables for option
2666 strings to pass to each program. For example, it defines
2667 <Constant>HC_OPTS</Constant><IndexTerm><Primary>HC_OPTS</Primary></IndexTerm>, the option strings to pass to the Haskell
2668 compiler. See <Xref LinkEnd="sec-suffix">.
2670 </ListItem></VarListEntry>
2672 <Term><filename>suffix.mk</filename><IndexTerm><Primary>suffix.mk</Primary></IndexTerm></Term>
2675 defines standard pattern rules—see <Xref LinkEnd="sec-suffix">.
2677 </ListItem></VarListEntry>
2682 Any of the variables and pattern rules defined by the boilerplate file
2683 can easily be overridden in any particular <filename>Makefile</filename>, because the
2684 boilerplate <Literal>include</Literal> comes first. Definitions after this <Literal>include</Literal>
2685 directive simply override the default ones in <filename>boilerplate.mk</filename>.
2690 <Sect2 id="sec-suffix">
2691 <Title>Pattern rules and options
2693 <IndexTerm><Primary>Pattern rules</Primary></IndexTerm></Title>
2696 The file <filename>suffix.mk</filename><IndexTerm><Primary>suffix.mk</Primary></IndexTerm> defines standard <Emphasis>pattern
2697 rules</Emphasis> that say how to build one kind of file from another, for
2698 example, how to build a <filename>.o</filename> file from a <filename>.c</filename> file. (GNU <Command>make</Command>'s
2699 <Emphasis>pattern rules</Emphasis> are more powerful and easier to use than Unix
2700 <Command>make</Command>'s <Emphasis>suffix rules</Emphasis>.)
2704 Almost all the rules look something like this:
2712 $(CC) $(CC_OPTS) -c $< -o $@
2718 Here's how to understand the rule. It says that
2719 <Emphasis>something</Emphasis><filename>.o</filename> (say <filename>Foo.o</filename>) can be built from
2720 <Emphasis>something</Emphasis><filename>.c</filename> (<filename>Foo.c</filename>), by invoking the C compiler
2721 (path name held in <Constant>$(CC)</Constant>), passing to it the options
2722 <Constant>$(CC_OPTS)</Constant> and the rule's dependent file of the rule
2723 <Literal>$<</Literal> (<filename>Foo.c</filename> in this case), and putting the result in
2724 the rule's target <Literal>$@</Literal> (<filename>Foo.o</filename> in this case).
2728 Every program is held in a <Command>make</Command> variable defined in
2729 <filename>mk/config.mk</filename>—look in <filename>mk/config.mk</filename> for the
2730 complete list. One important one is the Haskell compiler, which is
2731 called <Constant>$(HC)</Constant>.
2735 Every program's options are are held in a <Command>make</Command> variables called
2736 <Constant><prog>_OPTS</Constant>. the <Constant><prog>_OPTS</Constant> variables are defined in
2737 <filename>mk/opts.mk</filename>. Almost all of them are defined like this:
2743 CC_OPTS = $(SRC_CC_OPTS) $(WAY$(_way)_CC_OPTS) $($*_CC_OPTS) $(EXTRA_CC_OPTS)
2749 The four variables from which <Constant>CC_OPTS</Constant> is built have the following meaning:
2756 <Term><Constant>SRC_CC_OPTS</Constant><IndexTerm><Primary>SRC_CC_OPTS</Primary></IndexTerm>:</Term>
2759 options passed to all C
2762 </ListItem></VarListEntry>
2764 <Term><Constant>WAY_<way>_CC_OPTS</Constant>:</Term>
2768 compilations for way <Literal><way></Literal>. For example,
2769 <Constant>WAY_mp_CC_OPTS</Constant> gives options to pass to the C compiler when
2770 compiling way <Literal>mp</Literal>. The variable <Constant>WAY_CC_OPTS</Constant> holds
2771 options to pass to the C compiler when compiling the standard way.
2772 (<Xref LinkEnd="sec-ways"> dicusses multi-way
2775 </ListItem></VarListEntry>
2777 <Term><Constant><module>_CC_OPTS</Constant>:</Term>
2781 pass to the C compiler that are specific to module <Literal><module></Literal>. For example, <Constant>SMap_CC_OPTS</Constant> gives the specific options
2782 to pass to the C compiler when compiling <filename>SMap.c</filename>.
2784 </ListItem></VarListEntry>
2786 <Term><Constant>EXTRA_CC_OPTS</Constant><IndexTerm><Primary>EXTRA_CC_OPTS</Primary></IndexTerm>:</Term>
2789 extra options to pass to all
2790 C compilations. This is intended for command line use, thus:
2796 gmake libHS.a EXTRA_CC_OPTS="-v"
2800 </ListItem></VarListEntry>
2806 <Sect2 id="sec-targets">
2807 <Title>The main <filename>mk/target.mk</filename> file
2809 <IndexTerm><Primary>target.mk</Primary></IndexTerm></Title>
2812 <filename>target.mk</filename> contains canned rules for all the standard targets
2813 described in <Xref LinkEnd="sec-standard-targets">. It is complicated by the fact that you don't want all of
2814 these rules to be active in every <filename>Makefile</filename>. Rather than have a
2815 plethora of tiny files which you can include selectively, there is a
2816 single file, <filename>target.mk</filename>, which selectively includes rules based on
2817 whether you have defined certain variables in your <filename>Makefile</filename>. This
2818 section explains what rules you get, what variables control them, and
2819 what the rules do. Hopefully, you will also get enough of an idea of
2820 what is supposed to happen that you can read and understand any weird
2821 special cases yourself.
2828 <Term><Constant>HS_PROG</Constant><IndexTerm><Primary>HS_PROG</Primary></IndexTerm>.</Term>
2831 If <Constant>HS_PROG</Constant> is defined, you get
2832 rules with the following targets:
2836 <Term><filename>HS_PROG</filename><IndexTerm><Primary>HS_PROG</Primary></IndexTerm></Term>
2839 itself. This rule links <Constant>$(OBJS)</Constant>
2840 with the Haskell runtime system to get an executable called
2841 <Constant>$(HS_PROG)</Constant>.
2843 </ListItem></VarListEntry>
2845 <Term><Literal>install</Literal><IndexTerm><Primary>install</Primary></IndexTerm></Term>
2848 installs <Constant>$(HS_PROG)</Constant>
2849 in <Constant>$(bindir)</Constant>.
2851 </ListItem></VarListEntry>
2854 </ListItem></VarListEntry>
2856 <Term><Constant>C_PROG</Constant><IndexTerm><Primary>C_PROG</Primary></IndexTerm></Term>
2859 is similar to <Constant>HS_PROG</Constant>, except that
2860 the link step links <Constant>$(C_OBJS)</Constant> with the C runtime system.
2862 </ListItem></VarListEntry>
2864 <Term><Constant>LIBRARY</Constant><IndexTerm><Primary>LIBRARY</Primary></IndexTerm></Term>
2867 is similar to <Constant>HS_PROG</Constant>, except that
2868 it links <Constant>$(LIB_OBJS)</Constant> to make the library archive <Constant>$(LIBRARY)</Constant>, and
2869 <Literal>install</Literal> installs it in <Constant>$(libdir)</Constant>.
2871 </ListItem></VarListEntry>
2873 <Term><Constant>LIB_DATA</Constant><IndexTerm><Primary>LIB_DATA</Primary></IndexTerm></Term>
2878 </ListItem></VarListEntry>
2880 <Term><Constant>LIB_EXEC</Constant><IndexTerm><Primary>LIB_EXEC</Primary></IndexTerm></Term>
2885 </ListItem></VarListEntry>
2887 <Term><Constant>HS_SRCS</Constant><IndexTerm><Primary>HS_SRCS</Primary></IndexTerm>, <Constant>C_SRCS</Constant><IndexTerm><Primary>C_SRCS</Primary></IndexTerm>.</Term>
2890 If <Constant>HS_SRCS</Constant>
2891 is defined and non-empty, a rule for the target <Literal>depend</Literal> is included,
2892 which generates dependency information for Haskell programs.
2893 Similarly for <Constant>C_SRCS</Constant>.
2895 </ListItem></VarListEntry>
2900 All of these rules are ``double-colon'' rules, thus
2906 install :: $(HS_PROG)
2907 ...how to install it...
2913 GNU <Command>make</Command> treats double-colon rules as separate entities. If there
2914 are several double-colon rules for the same target it takes each in
2915 turn and fires it if its dependencies say to do so. This means that
2916 you can, for example, define both <Constant>HS_PROG</Constant> and <Constant>LIBRARY</Constant>, which will
2917 generate two rules for <Literal>install</Literal>. When you type <Command>gmake install</Command> both
2918 rules will be fired, and both the program and the library will be
2919 installed, just as you wanted.
2924 <Sect2 id="sec-subdirs">
2927 <IndexTerm><Primary>recursion, in makefiles</Primary></IndexTerm>
2928 <IndexTerm><Primary>Makefile, recursing into subdirectories</Primary></IndexTerm></Title>
2931 In leaf <filename>Makefile</filename>s the variable <Constant>SUBDIRS</Constant><IndexTerm><Primary>SUBDIRS</Primary></IndexTerm> is undefined.
2932 In non-leaf <filename>Makefile</filename>s, <Constant>SUBDIRS</Constant> is set to the list of
2933 sub-directories that contain subordinate <filename>Makefile</filename>s. <Emphasis>It is up to
2934 you to set <Constant>SUBDIRS</Constant> in the <filename>Makefile</filename>.</Emphasis> There is no automation here—<Constant>SUBDIRS</Constant> is too important to automate.
2938 When <Constant>SUBDIRS</Constant> is defined, <filename>target.mk</filename> includes a rather
2939 neat rule for the standard targets (<Xref LinkEnd="sec-standard-targets"> that simply invokes
2940 <Command>make</Command> recursively in each of the sub-directories.
2944 <Emphasis>These recursive invocations are guaranteed to occur in the order
2945 in which the list of directories is specified in <Constant>SUBDIRS</Constant>. </Emphasis>This
2946 guarantee can be important. For example, when you say <Command>gmake boot</Command> it
2947 can be important that the recursive invocation of <Command>make boot</Command> is done
2948 in one sub-directory (the include files, say) before another (the
2949 source files). Generally, put the most independent sub-directory
2950 first, and the most dependent last.
2955 <Sect2 id="sec-ways">
2956 <Title>Way management
2958 <IndexTerm><Primary>way management</Primary></IndexTerm></Title>
2961 We sometimes want to build essentially the same system in several
2962 different ``ways''. For example, we want to build GHC's <Literal>Prelude</Literal>
2963 libraries with and without profiling, with and without concurrency,
2964 and so on, so that there is an appropriately-built library archive to
2965 link with when the user compiles his program. It would be possible to
2966 have a completely separate build tree for each such ``way'', but it
2967 would be horribly bureaucratic, especially since often only parts of
2968 the build tree need to be constructed in multiple ways.
2972 Instead, the <filename>target.mk</filename><IndexTerm><Primary>target.mk</Primary></IndexTerm> contains some clever magic to
2973 allow you to build several versions of a system; and to control
2974 locally how many versions are built and how they differ. This section
2979 The files for a particular way are distinguished by munging the
2980 suffix. The ``normal way'' is always built, and its files have the
2981 standard suffices <filename>.o</filename>, <filename>.hi</filename>, and so on. In addition, you can build
2982 one or more extra ways, each distinguished by a <Emphasis>way tag</Emphasis>. The
2983 object files and interface files for one of these extra ways are
2984 distinguished by their suffix. For example, way <Literal>mp</Literal> has files
2985 <filename>.mp_o</filename> and <filename>.mp_hi</filename>. Library archives have their way tag the other
2986 side of the dot, for boring reasons; thus, <filename>libHS_mp.a</filename>.
2990 A <Command>make</Command> variable called <Constant>way</Constant> holds the current way tag. <Emphasis><Constant>way</Constant>
2991 is only ever set on the command line of a recursive invocation of
2992 <Command>gmake</Command>.</Emphasis> It is never set inside a <filename>Makefile</filename>. So it is a global
2993 constant for any one invocation of <Command>gmake</Command>. Two other <Command>make</Command>
2994 variables, <Constant>way_</Constant> and <Constant>_way</Constant> are immediately derived from <Constant>$(way)</Constant> and
2995 never altered. If <Constant>way</Constant> is not set, then neither are <Constant>way_</Constant> and
2996 <Constant>_way</Constant>, and the invocation of <Command>make</Command> will build the ``normal way''.
2997 If <Constant>way</Constant> is set, then the other two variables are set in sympathy.
2998 For example, if <Constant>$(way)</Constant> is ``<Literal>mp</Literal>'', then <Constant>way_</Constant> is set to ``<Literal>mp_</Literal>''
2999 and <Constant>_way</Constant> is set to ``<Literal>_mp</Literal>''. These three variables are then used
3000 when constructing file names.
3004 So how does <Command>make</Command> ever get recursively invoked with <Constant>way</Constant> set? There
3005 are two ways in which this happens:
3014 For some (but not all) of the standard targets, when in a leaf
3015 sub-directory, <Command>make</Command> is recursively invoked for each way tag in
3016 <Constant>$(WAYS)</Constant>. You set <Constant>WAYS</Constant> to the list of way tags you want these
3017 targets built for. The mechanism here is very much like the recursive
3018 invocation of <Command>make</Command> in sub-directories (<Xref LinkEnd="sec-subdirs">).
3020 It is up to you to set <Constant>WAYS</Constant> in your <filename>Makefile</filename>; this is how you
3021 control what ways will get built.
3027 For a useful collection of
3028 targets (such as <filename>libHS_mp.a</filename>, <filename>Foo.mp_o</filename>) there is a rule which
3029 recursively invokes <Command>make</Command> to make the specified target, setting the
3030 <Constant>way</Constant> variable. So if you say <Command>gmake Foo.mp_o</Command> you should see a
3031 recursive invocation <Command>gmake Foo.mp_o way=mp</Command>, and <Emphasis>in this
3032 recursive invocation the pattern rule for compiling a Haskell file
3033 into a <filename>.o</filename> file will match</Emphasis>. The key pattern rules (in <filename>suffix.mk</filename>)
3039 $(HC) $(HC_OPTS) $< -o $@
3054 <Title>When the canned rule isn't right</Title>
3057 Sometimes the canned rule just doesn't do the right thing. For
3058 example, in the <Literal>nofib</Literal> suite we want the link step to print out
3059 timing information. The thing to do here is <Emphasis>not</Emphasis> to define
3060 <Constant>HS_PROG</Constant> or <Constant>C_PROG</Constant>, and instead define a special purpose rule in
3061 your own <filename>Makefile</filename>. By using different variable names you will avoid
3062 the canned rules being included, and conflicting with yours.
3069 <Sect1 id="sec-booting-from-C">
3070 <Title>Booting/porting from C (<filename>.hc</filename>) files
3072 <IndexTerm><Primary>building GHC from .hc files</Primary></IndexTerm>
3073 <IndexTerm><Primary>booting GHC from .hc files</Primary></IndexTerm>
3074 <IndexTerm><Primary>porting GHC</Primary></IndexTerm></Title>
3077 This section is for people trying to get GHC going by using the supplied
3078 intermediate C (<filename>.hc</filename>) files. This would probably be
3079 because no binaries have been provided, or because the machine is not ``fully
3084 The intermediate C files are normally made available together with a source
3085 release, please check the announce message for exact directions of where to
3086 find them. If we haven't made them available or you can't find them, please
3091 Assuming you've got them, unpack them on top of a fresh source tree. This
3092 will place matching <filename>.hc</filename> files next to the corresponding
3093 Haskell source in the compiler subdirectory <filename>ghc</filename> and in
3094 the language package of hslibs (i.e., in <filename>hslibs/lang</filename>).
3095 Then follow the `normal' instructions in <Xref
3096 LinkEnd="sec-building-from-source"> for setting up a build tree.
3100 The actual build process is fully automated by the
3101 <filename>hc-build</filename> script located in the
3102 <filename>distrib</filename> directory. If you eventually want to install GHC
3103 into the directory <filename>INSTALL_DIRECTORY</filename>, the following
3104 command will execute the whole build process (it won't install yet):
3107 foo% distrib/hc-build --prefix=INSTALL_DIRECTORY
3109 <IndexTerm><Primary>--hc-build</Primary></IndexTerm>
3111 By default, the installation directory is <filename>/usr/local</filename>. If
3112 that is what you want, you may omit the argument to
3113 <filename>hc-build</filename>. Generally, any option given to
3114 <filename>hc-build</filename> is passed through to the configuration script
3115 <filename>configure</filename>. If <filename>hc-build</filename>
3116 successfully completes the build process, you can install the resulting
3117 system, as normal, with
3124 That's the mechanics of the boot process, but, of course, if you're
3125 trying to boot on a platform that is not supported and significantly
3126 `different' from any of the supported ones, this is only the start of
3127 the adventure…(ToDo: porting tips—stuff to look out for, etc.)
3132 <Sect1 id="sec-build-pitfalls">
3133 <Title>Known pitfalls in building Glasgow Haskell
3135 <IndexTerm><Primary>problems, building</Primary></IndexTerm>
3136 <IndexTerm><Primary>pitfalls, in building</Primary></IndexTerm>
3137 <IndexTerm><Primary>building pitfalls</Primary></IndexTerm></Title>
3140 WARNINGS about pitfalls and known ``problems'':
3149 One difficulty that comes up from time to time is running out of space
3150 in <literal>TMPDIR</literal>. (It is impossible for the configuration stuff to
3151 compensate for the vagaries of different sysadmin approaches to temp
3153 <IndexTerm><Primary>tmp, running out of space in</Primary></IndexTerm>
3155 The quickest way around it is <Command>setenv TMPDIR /usr/tmp</Command><IndexTerm><Primary>TMPDIR</Primary></IndexTerm> or
3156 even <Command>setenv TMPDIR .</Command> (or the equivalent incantation with your shell
3159 The best way around it is to say
3162 export TMPDIR=<dir>
3165 in your <filename>build.mk</filename> file.
3166 Then GHC and the other <Literal>fptools</Literal> programs will use the appropriate directory
3175 In compiling some support-code bits, e.g., in <filename>ghc/rts/gmp</filename> and even
3176 in <filename>ghc/lib</filename>, you may get a few C-compiler warnings. We think these
3184 When compiling via C, you'll sometimes get ``warning: assignment from
3185 incompatible pointer type'' out of GCC. Harmless.
3192 Similarly, <Command>ar</Command>chiving warning messages like the following are not
3196 ar: filename GlaIOMonad__1_2s.o truncated to GlaIOMonad_
3197 ar: filename GlaIOMonad__2_2s.o truncated to GlaIOMonad_
3207 In compiling the compiler proper (in <filename>compiler/</filename>), you <Emphasis>may</Emphasis>
3208 get an ``Out of heap space'' error message. These can vary with the
3209 vagaries of different systems, it seems. The solution is simple:
3216 If you're compiling with GHC 4.00 or later, then the
3217 <Emphasis>maximum</Emphasis> heap size must have been reached. This
3218 is somewhat unlikely, since the maximum is set to 64M by default.
3219 Anyway, you can raise it with the
3220 <Option>-optCrts-M<size></Option> flag (add this flag to
3221 <Constant><module>_HC_OPTS</Constant>
3222 <Command>make</Command> variable in the appropriate
3223 <filename>Makefile</filename>).
3230 For GHC < 4.00, add a suitable <Option>-H</Option> flag to the <filename>Makefile</filename>, as
3239 and try again: <Command>gmake</Command>. (see <Xref LinkEnd="sec-suffix"> for information about
3240 <Constant><module>_HC_OPTS</Constant>.)
3242 Alternatively, just cut to the chase:
3246 % make EXTRA_HC_OPTS=-optCrts-M128M
3255 If you try to compile some Haskell, and you get errors from GCC about
3256 lots of things from <filename>/usr/include/math.h</filename>, then your GCC was
3257 mis-installed. <Command>fixincludes</Command> wasn't run when it should've been.
3259 As <Command>fixincludes</Command> is now automagically run as part of GCC installation,
3260 this bug also suggests that you have an old GCC.
3268 You <Emphasis>may</Emphasis> need to re-<Command>ranlib</Command><IndexTerm><Primary>ranlib</Primary></IndexTerm> your libraries (on Sun4s).
3272 % cd $(libdir)/ghc-x.xx/sparc-sun-sunos4
3273 % foreach i ( `find . -name '*.a' -print` ) # or other-shell equiv...
3275 ? # or, on some machines: ar s $i
3280 We'd be interested to know if this is still necessary.
3288 GHC's sources go through <Command>cpp</Command> before being compiled, and <Command>cpp</Command> varies
3289 a bit from one Unix to another. One particular gotcha is macro calls
3294 SLIT("Hello, world")
3298 Some <Command>cpp</Command>s treat the comma inside the string as separating two macro
3299 arguments, so you get
3303 :731: macro `SLIT' used with too many (2) args
3307 Alas, <Command>cpp</Command> doesn't tell you the offending file!
3309 Workaround: don't put weird things in string args to <Command>cpp</Command> macros.
3320 <Sect1 id="winbuild"><Title>Notes for building under Windows</Title>
3323 This section summarises how to get the utilities you need on your
3324 Win95/98/NT/2000 machine to use CVS and build GHC. Similar notes for
3325 installing and running GHC may be found in the user guide. In general,
3326 Win95/Win98 behave the same, and WinNT/Win2k behave the same.
3327 You should read the GHC installation guide sections on Windows (in the user
3328 guide) before continuing to read these notes.
3332 Because of various hard-wired infelicities, you need to copy
3333 <Filename>bash.exe</Filename> (from GHC's <Filename>extra-bin</Filename>
3334 directory), and <Filename>perl.exe</Filename> and
3335 <Filename>cat.exe</Filename> (from GHC's <Filename>bin</Filename> directory)
3336 to <Filename>/bin</Filename> (discover where your Cygwin root directory is
3337 by typign <Command>mount</Command>).
3341 Before you start, you need to make sure that the user environment variable
3342 <Constant>MAKE_MODE</Constant> is set to <Literal>UNIX</Literal>. If you
3343 don't do this you get very weird messages when you type
3344 <Command>make</Command>, such as:
3347 /c: /c: No such file or directory</Screen>
3349 <Sect2><Title>Configuring ssh</Title>
3355 Generate a key, by running <filename>c:/user/local/bin/ssh-keygen1</filename>.
3356 This generates a public key in <filename>.ssh/identity.pub</filename>, and a
3357 private key in <filename>.ssh/identity</filename>
3361 In response to the 'Enter passphrase' question, just hit
3362 return (i.e. use an empty passphrase). The passphrase is
3363 a password that protects your private key. But it's a pain
3364 to type this passphrase everytime you use <Command>ssh</Command>, so the best
3365 thing to do is simply to protect your <filename>.ssh</filename> directory, and
3366 <filename>.ssh/identity</filename> from access by anyone else. To do this
3367 right-click your <filename>.ssh</filename> directory, and select Properties.
3368 If you are not on the access control list, add yourself, and
3369 give yourself full permissions (the second panel).
3370 Remove everyone else from the access control list. (Don't
3371 leave them there but deny them access, because 'they' may be
3372 a list that includes you!)
3376 If you have problems running <Command>ssh-keygen1</Command>
3377 from within <Command>bash</Command>, start up <filename>cmd.exe</filename> and run it as follows:
3381 c:\tmp> set CYGWIN32=tty
3382 c:\tmp> c:/user/local/bin/ssh-keygen1
3388 If you don't have an account on <Literal>cvs.haskell.org</Literal>, send
3389 your <filename>.ssh/identity.pub</filename> to the CVS repository administrator
3390 (currently Jeff Lewis <Email>jlewis@galconn.com</Email>). He will set up
3395 If you do have an account on <Literal>cvs.haskell.org</Literal>, use TeraTerm
3396 to logon to it. Once in, copy the
3397 key that <Command>ssh-keygen1</Command> deposited in <filename>/.ssh/identity.pub</filename> into
3398 your <filename>~/.ssh/authorized_keys</filename>. Make sure that the new version
3399 of <filename>authorized_keys</filename> still has 600 file permission.
3408 <Sect2><Title>Configuring CVS</Title>
3414 From the System control panel,
3415 set the following <Emphasis>user</Emphasis> environment variables (see the GHC user guide)
3421 <Constant>HOME</Constant>: points to your home directory. This is where CVS
3422 will look for its <filename>.cvsrc</filename> file.
3428 <Constant>CVS_RSH</Constant>: <filename>c:/path_to_ghc/extra-bin/ssh</filename>
3434 <Constant>CVSROOT</Constant>: <Literal>:ext:username@cvs.haskell.org:/home/cvs/root</Literal>,
3435 where <Literal>username</Literal> is your userid
3441 <Constant>CVSEDITOR</Constant>: <filename>bin/gnuclient.exe</filename> if you want to use an Emacs buffer for typing in those long commit messages.
3447 <Constant>SHELL</Constant>: To use bash as the shell in Emacs, you need to
3448 set this to point to <Filename>bash.exe</Filename>.
3457 Put the following in <filename>$HOME/.cvsrc</filename>:
3468 These are the default options for the specified CVS commands,
3469 and represent better defaults than the usual ones. (Feel
3470 free to change them.)
3474 Filenames starting with <filename>.</filename> were illegal in
3475 the 8.3 DOS filesystem, but that restriction should have
3476 been lifted by now (i.e., you're using VFAT or later filesystems.) If
3477 you're still having problems creating it, don't worry; <filename>.cvsrc</filename> is entirely
3484 Try doing <Command>cvs co fpconfig</Command>. All being well, bytes should
3485 start to trickle through, leaving a directory <filename>fptools</filename>
3486 in your current directory. (You can <Command>rm</Command> it if you don't
3487 want to keep it.) The following messages appear to be harmless:
3491 setsockopt IPTOS_LOWDELAY: Invalid argument
3492 setsockopt IPTOS_THROUGHPUT: Invalid argument
3496 At this point I found that CVS tried to invoke a little dialogue with
3497 me (along the lines of `do you want to talk to this host?'), but
3498 for some reason bombed out. This was from a bash shell running in Emacs.
3499 I solved this by invoking a Cygnus shell, and running CVS from there.
3500 Once things are dialogue free, it seems to work OK from within Emacs.
3506 If you want to check out part of large tree, proceed as follows:
3510 cvs -f checkout -l papers
3516 This sequence checks out the <Literal>papers</Literal> module, but none
3517 of its sub-directories.
3518 The "<Option>-l</Option>" flag says not to check out sub-directories.
3519 The "<Option>-f</Option>" flag says not to read the <filename>.cvsrc</filename> file
3520 whose <Option>-P</Option> default (don't check out empty directories) is
3525 The <Command>cvs update</Command> command sucks in a named sub-directory.
3532 There is a very nice graphical front-end to CVS for Win32 platforms,
3533 with a UI that people will be familiar with, at
3534 <ULink URL="http://www.wincvs.org/">wincvs.org</ULink>.
3535 I have not tried it yet.
3541 <Sect2><Title>Building GHC</Title>
3547 You should give the option
3548 <option>--host=i386-unknown-mingw32</option> to
3549 <command>autoconf</command>, so that it doesn't try to
3550 build for Cygwin (unless that's what you really
3551 want). Since it's possible to get into trouble using the
3552 wrong C compiler, it's wise either to avoid installing
3553 Cygwin gcc, or to explicitly specify
3554 <option>--with-gcc=/mingw/bin/gcc</option>.
3560 You have to run <Command>autoconf</Command> both in <filename>fptools</filename>
3561 and in <filename>fptools/ghc</filename>. If you omit the latter step you'll
3562 get an error when you run <filename>./configure</filename>:
3567 creating mk/config.h
3568 mk/config.h is unchanged
3570 running /bin/sh ./configure --cache-file=.././config.cache --srcdir=.
3571 ./configure: ./configure: No such file or directory
3572 configure: error: ./configure failed for ghc
3578 You need <filename>ghc</filename> to be in your <Constant>PATH</Constant> before you run
3579 <Command>configure</Command>. The InstallShield tells you the path
3580 when you install it.
3589 <title>Building the Windows InstallShield® Installer</title>
3592 This section is intended for GHC developers only; no-one else
3593 should need to build an InstallShield.
3597 Having built a second-stage tree and done <command>make
3598 install</command> on it, open the InstallShield
3599 (<filename>.ism</filename>) file. Open the Project screen, and
3600 then the Project subfolder of the Path variables folder, and
3601 set <literal>SourceFiles</literal> to the top of your
3602 tree. You might also need to set <literal>GHCBITS</literal> to
3603 point to the tree of various external bits that are added into
3604 the IS mix. You should then be able to build an InstallShield.
3608 <title>Extra features of the InstallShield</title>
3611 The InstallShield has some IS-specific twiddles:
3616 Two registry entries are set under
3617 <literal>HKEY_LOCAL_MACHINE\SOFTWARE\GHC</literal>:
3618 <literal>Path</literal> and
3619 <literal>Version</literal>, which record respectively
3620 the directory in which GHC was installed, and the
3626 The InstallShield adds some entries to the Program
3627 menu, for GHCi and for the documentation. See under
3628 Setup Design and the individual components (each
3629 component can add entries to the menu).
3637 <title>External add-ins</title>
3640 The external add-ins consist of Mingwin gcc and Mingwin
3641 Perl. The layout of the add-ins tree is as follows:
3646 perl.exe (Mingwin perl)
3649 Mingwin gcc binaries, libraries and headers
3652 imports for HDirect's com library