doc/en_US.ISO8859-1/books/pmake/shortcuts/chapter.sgml
2008-08-13 09:42:21 +00:00

1229 lines
48 KiB
Text

<!--
$FreeBSD$
-->
<chapter id="shortcuts">
<title>Short-cuts and Other Nice Things</title>
<para>Based on what I have told you so far, you may have gotten the
impression that <application>PMake</application> is just a way of
storing away commands and making sure you do not forget to compile
something. Good. That is just what it is. However, the ways I
have described have been inelegant, at best, and painful, at
worst. This chapter contains things that make the writing of
makefiles easier and the makefiles themselves shorter and easier
to modify (and, occasionally, simpler). In this chapter, I assume
you are somewhat more familiar with Sprite (or &unix;, if that is
what you are using) than I did in <xref linkend="basics">, just so
you are on your toes. So without further ado&hellip;</para>
<section id="rules">
<title>Transformation Rules</title>
<para>As you know, a file's name consists of two parts: a base
name, which gives some hint as to the contents of the file, and
a suffix, which usually indicates the format of the file. Over
the years, as &unix; has developed, naming conventions, with
regard to suffixes, have also developed that have become almost
as incontrovertible as Law. E.g. a file ending in
<filename>.c</filename> is assumed to contain C source code; one
with a <filename>.o</filename> suffix is assumed to be a
compiled, relocatable object file that may be linked into any
program; a file with a <filename>.ms</filename> suffix is
usually a text file to be processed by
<application>Troff</application> with the <literal>-ms</literal>
macro package, and so on. One of the best aspects of both
<application>Make</application> and
<application>PMake</application> comes from their understanding
of how the suffix of a file pertains to its contents and their
ability to do things with a file based solely on its suffix.
This ability comes from something known as a transformation
rule. A transformation rule specifies how to change a file with
one suffix into a file with another suffix.</para>
<para>A transformation rule looks much like a dependency line,
except the target is made of two known suffixes stuck
together. Suffixes are made known to
<application>PMake</application> by placing them
as sources on a dependency line whose target is the special
target <makevar>.SUFFIXES</makevar>. E.g.:</para>
<programlisting>.SUFFIXES : .o .c
.c.o :
$(CC) $(CFLAGS) -c $(.IMPSRC)</programlisting>
<para>The creation script attached to the target is used to
trans form a file with the first suffix (in this case,
<filename>.c</filename>) into a
file with the second suffix (here, <filename>.o</filename>).
In addition, the target inherits whatever attributes have
been applied to the transformation rule.
The simple rule given above says that to transform a C source
file into an object file, you compile it using
<application>cc</application> with the <option>-c</option>
flag. This rule is taken straight from the system makefile.
Many transformation rules (and suffixes) are defined there,
and I refer you to it for more examples
(type <command>pmake -h</command> to find out where it
is).</para>
<para>There are several things to note about the
transformation rule given above:</para>
<orderedlist>
<listitem>
<para>The <makevar>.IMPSRC</makevar> variable.
This variable is set to the
<quote>implied source</quote> (the file from which
the target is being created; the one with the first
suffix), which, in this case, is the
<filename>.c</filename> file.</para>
</listitem>
<listitem>
<para>The <envar>CFLAGS</envar> variable. Almost all of
the transformation rules in the system makefile are set
up using variables that you can alter in your makefile
to tailor the rule to your needs. In this case, if you
want all your C files to be compiled with the
<option>-g</option> flag, to provide information for
dbx, you would set the <envar>CFLAGS</envar> variable to
contain <option>-g</option> (<literal>CFLAGS =
-g</literal>) and <application>PMake</application>
would take care of the rest.</para>
</listitem>
</orderedlist>
<para>To give you a quick example, the makefile in
<xref linkend="envvars"> could be changed to this:</para>
<programlisting>OBJS = a.o b.o c.o
program : $(OBJS)
$(CC) -o $(.TARGET) $(.ALLSRC)
$(OBJS) : defs.h</programlisting>
<para>The transformation rule I gave above takes the place of the
6 lines
<footnote>
<para>This is also somewhat cleaner, I think, than
the dynamic source solution presented in
<xref linkend="writeanddebug">.</para>
</footnote>:</para>
<programlisting>a.o : a.c
cc -c a.c
b.o : b.c
cc -c b.c
c.o : c.c
cc -c c.c</programlisting>
<para>Now you may be wondering about the dependency between the
<filename>.o</filename>
and <filename>.c</filename> files &ndash; it is not mentioned
anywhere in the new makefile. This is because it is not needed:
one of the effects of applying a transformation rule is the
target comes to depend on the implied source. That's why it is
called the implied source.</para>
<para>For a more detailed example. Say you have a makefile like
this:</para>
<programlisting>a.out : a.o b.o
$(CC) $(.ALLSRC)</programlisting>
<para>and a directory set up like this:</para>
<screen>total 4
-rw-rw-r-- 1 deboor 34 Sep 7 00:43 Makefile
-rw-rw-r-- 1 deboor 119 Oct 3 19:39 a.c
-rw-rw-r-- 1 deboor 201 Sep 7 00:43 a.o
-rw-rw-r-- 1 deboor 69 Sep 7 00:43 b.c</screen>
<para>While just typing <command>pmake</command> will do
the right thing, it is much more informative to type
<command>pmake -d s</command>. This will
show you what <application>PMake</application> is up
to as it processes the files. In this case,
<application>PMake</application> prints the following:</para>
<screen>Suff_FindDeps (a.out)
using existing source a.o
applying .o -&gt; .out to "a.o"
Suff_FindDeps (a.o)
trying a.c...got it
applying .c -&gt; .o to "a.c"
Suff_FindDeps (b.o)
trying b.c...got it
applying .c -&gt; .o to "b.c"
Suff_FindDeps (a.c)
trying a.y...not there
trying a.l...not there
trying a.c,v...not there
trying a.y,v...not there
trying a.l,v...not there
Suff_FindDeps (b.c)
trying b.y...not there
trying b.l...not there
trying b.c,v...not there
trying b.y,v...not there
trying b.l,v...not there
--- a.o ---
cc -c a.c
--- b.o ---
cc -c b.c
--- a.out ---
cc a.o b.o</screen>
<para><computeroutput>Suff_FindDeps</computeroutput> is the
name of a function in <application>PMake</application> that
is called to check for implied sources for a target using
transformation rules. The transformations it tries are,
naturally enough, limited to the ones that have been defined
(a transformation may be defined multiple times, by the way,
but only the most recent one will be used). You will notice,
however, that there is a definite order to the suffixes that
are tried. This order is set by the relative positions of
the suffixes on the <makevar>.SUFFIXES</makevar> line
&ndash; the earlier a suffix appears, the earlier it is
checked as the source of a transformation. Once a suffix
has been defined, the only way to change its position in the
pecking order is to remove all the suffixes (by having a
<makevar>.SUFFIXES</makevar> dependency line with no sources)
and redefine them in the order you want.
(Previously-defined transformation rules will be
automatically redefined as the suffixes they involve are
re-entered.) Another way to affect the search order is to make
the dependency explicit. In the above example,
<filename>a.out</filename> depends on <filename>a.o</filename>
and <filename>b.o</filename>. Since a transformation exists
from <filename>.o</filename> to <filename>.out</filename>,
<application>PMake</application> uses that, as indicated by
the <computeroutput>using existing source a.o</computeroutput>
message.</para>
<para>The search for a transformation starts from the suffix of
the target and continues through all the defined
transformations, in the order dictated by the suffix ranking,
until an existing file with the same base (the target name
minus the suffix and any leading directories) is found. At that
point, one or more transformation rules will have been found
to change the one existing file into the target.</para>
<para>For example, ignoring what's in the system makefile for
now, say you have a makefile like this:</para>
<screen>.SUFFIXES : .out .o .c .y .l
.l.c :
lex $(.IMPSRC)
mv lex.yy.c $(.TARGET)
.y.c :
yacc $(.IMPSRC)
mv y.tab.c $(.TARGET)
.c.o :
cc -c $(.IMPSRC)
.o.out :
cc -o $(.TARGET) $(.IMPSRC)</screen>
<para>and the single file <filename>jive.l</filename>.
If you were to type <command>pmake -rd ms jive.out</command>,
you would get the following output for
<filename>jive.out</filename>:</para>
<screen>Suff_FindDeps (jive.out)
trying jive.o...not there
trying jive.c...not there
trying jive.y...not there
trying jive.l...got it
applying .l -&gt; .c to "jive.l"
applying .c -&gt; .o to "jive.c"
applying .o -&gt; .out to "jive.o"</screen>
<para>and this is why: <application>PMake</application> starts with the
target <filename>jive.out</filename>, figures out its suffix
(<filename>.out</filename>) and looks for things it can
transform to a <filename>.out</filename> file. In this case, it
only finds <filename>.o</filename>, so it looks for the file
<filename>jive.o</filename>. It fails to find it, so it
looks for transformations into a <filename>.o</filename>
file. Again it has only one choice: <filename>.c</filename>.
So it looks for <filename>jive.c</filename> and, as you
know, fails to find it. At this point it has two choices: it can
create the <filename>.c</filename> file from either a
<filename>.y</filename> file or a <filename>.l</filename> file.
Since <filename>.y</filename> came first on the
<makevar>.SUFFIXES</makevar> line, it checks for
<filename>jive.y</filename> first, but can not find it, so it looks
for <filename>jive.l</filename> and, lo and behold, there it is.
At this point, it has defined a transformation path as follows:</para>
<literallayout><filename>.l</filename> -&gt; <filename>.c</filename> -&gt; <filename>.o</filename> -&gt; <filename>.out</filename></literallayout>
<para>and applies the transformation rules accordingly. For completeness,
and to give you a better idea of what <application>PMake</application>
actually did with this three-step transformation, this is what
<application>PMake</application> printed for the rest of the
process:</para>
<screen>Suff_FindDeps (jive.o)
using existing source jive.c
applying .c -&gt; .o to "jive.c"
Suff_FindDeps (jive.c)
using existing source jive.l
applying .l -&gt; .c to "jive.l"
Suff_FindDeps (jive.l)
Examining jive.l...modified 17:16:01 Oct 4, 1987...up-to-date
Examining jive.c...non-existent...out-of-date
--- jive.c ---
lex jive.l
... meaningless lex output deleted ...
mv lex.yy.c jive.c
Examining jive.o...non-existent...out-of-date
--- jive.o ---
cc -c jive.c
Examining jive.out...non-existent...out-of-date
--- jive.out ---
cc -o jive.out jive.o</screen>
<para>One final question remains: what does
<application>PMake</application> do with targets that have no
known suffix? <application>PMake</application> simply pretends
it actually has a known suffix and searches for
transformations accordingly. The suffix it chooses is the
source for the <maketarget>.NULL</maketarget> target mentioned
later. In the system makefile, <filename>.out</filename> is
chosen as the <quote>null suffix</quote> because most people
use <application>PMake</application> to create programs. You
are, however, free and welcome to change it to a suffix
of your own choosing. The null suffix is ignored, however,
when <application>PMake</application> is in compatibility
mode (see <xref linkend="gods">).</para>
</section>
<section id="including">
<title>Including Other Makefiles</title>
<para>Just as for programs, it is often useful to extract certain
parts of a makefile into another file and just include it in
other makefiles somehow. Many compilers allow you say something
like:</para>
<programlisting>#include "defs.h"</programlisting>
<para>to include the contents of <filename>defs.h</filename>
in the source file. <application>PMake</application>
allows you to do the same thing for makefiles, with the
added ability to use variables in the filenames. An include
directive in a makefile looks either like this:</para>
<programlisting>#include &lt;file&gt;</programlisting>
<para>or this:</para>
<programlisting>#include "file"</programlisting>
<para>The difference between the two is where
<application>PMake</application> searches for the file: the first way,
<application>PMake</application> will look for the file only in the
system makefile directory (or directories) (to find out what that
directory is, give <application>PMake</application> the
<filename>-h</filename> flag).
The system makefile directory search path can be overridden via the
<option>-m</option> option. For files in double-quotes, the search
is more complex:</para>
<orderedlist>
<listitem>
<para>The directory of the makefile that's including the
file.</para>
</listitem>
<listitem>
<para>The current directory (the one in which you
invoked <application>PMake</application>).</para>
</listitem>
<listitem>
<para>The directories given by you using
<option>-I</option> flags, in the order in which you
gave them.</para>
</listitem>
<listitem>
<para>Directories given by
<makevar>.PATH</makevar> dependency lines (see
<xref linkend="gods">).</para>
</listitem>
<listitem>
<para>The system makefile directory.</para>
</listitem>
</orderedlist>
<para>in that order.</para>
<para>You are free to use <application>PMake</application> variables
in the filename &ndash; <application>PMake</application>
will expand them before searching for the file. You must
specify the searching method with either angle brackets or
double-quotes outside of a variable expansion. I.e. the following:</para>
<programlisting>SYSTEM = &lt;command.mk&gt;
#include $(SYSTEM)</programlisting>
<para>will not work.</para>
</section>
<section id="savingcmds">
<title>Saving Commands</title>
<para>There may come a time when you will want to save certain
commands to be executed when everything else is done. For
instance: you are making several different libraries at one
time and you want to create the members in parallel. Problem is,
<application>ranlib</application> is another one of those
programs that can not be run more than once in the same directory
at the same time (each one creates a file called
<filename>__.SYMDEF</filename> into which it stuffs information
for the linker to use. Two of them running at once will
overwrite each other's file and the result will be garbage for
both parties). You might want a way to save the ranlib
commands til the end so they can be run one after the other,
thus keeping them from trashing each other's file.
<application>PMake</application> allows you to do this by
inserting an ellipsis (<quote>...</quote>) as a command between
commands to be run at once and those to be run later.</para>
<para>So for the <application>ranlib</application> case above,
you might do this:</para>
<programlisting>lib1.a : $(LIB1OBJS)
rm -f $(.TARGET)
ar cr $(.TARGET) $(.ALLSRC)
...
ranlib $(.TARGET)
lib2.a : $(LIB2OBJS)
rm -f $(.TARGET)
ar cr $(.TARGET) $(.ALLSRC)
...
ranlib $(.TARGET)</programlisting>
<para>This would save both</para>
<programlisting>ranlib $(.TARGET)</programlisting>
<para>commands until the end, when they would run one after the
other (using the correct value for the
<makevar>.TARGET</makevar> variable, of course).</para>
<para>Commands saved in this manner are only executed if
<application>PMake</application> manages to re-create
everything without an error.</para>
</section>
<section id="targetattr">
<title>Target Attributes</title>
<para><application>PMake</application> allows you to give
attributes to targets by means of special sources. Like
everything else <application>PMake</application> uses, these
sources begin with a period and are made up of all upper-case
letters. There are various reasons for using them, and I will
try to give examples for most of them. Others you will have to
find uses for yourself. Think of it as <quote>an exercise for
the reader</quote>. By placing one (or more) of these as a
source on a dependency line, you are <quote>marking the
target(s) with that attribute</quote>. That is just the way I
phrase it, so you know.</para>
<para>Any attributes given as sources for a transformation
rule are applied to the target of the transformation rule
when the rule is applied.</para>
<informaltable frame="none">
<tgroup cols="2">
<colspec colwidth="1*">
<colspec colwidth="10*">
<tbody>
<row valign="top">
<entry><literal>.DONTCARE</literal></entry>
<entry>If a target is marked with this attribute and
<application>PMake</application> can not figure out
how to create it, it will ignore this fact and assume
the file is not really needed or actually exists and
<application>PMake</application> just can not find
it. This may prove wrong, but the error will be
noted later on, not when <application>PMake</application>
tries to create the target so marked. This attribute also
prevents <application>PMake</application> from attempting
to touch the target if it is given the
<option>-t</option> flag.</entry>
</row>
<row valign="top">
<entry><literal>.EXEC</literal></entry>
<entry><para>This attribute causes its shell script to be
executed while having no effect on targets that depend
on it. This makes the target into a sort of subroutine.
An example. Say you have some LISP files that need to
be compiled and loaded into a LISP process. To do this,
you echo LISP commands into a file and execute a LISP
with this file as its input when everything is
done. Say also that you have to load
other files from another system before you can compile
your files and further, that you do not want to go
through the loading and dumping unless one of your
files has changed. Your makefile might look a little
bit like this (remember, this is an educational example,
and do not worry about the <maketarget>COMPILE</maketarget>
rule, all will soon become clear, grasshopper):
<programlisting>system : init a.fasl b.fasl c.fasl
for i in $(.ALLSRC);
do
echo -n '(load "' &gt;&gt; input
echo -n ${i} &gt;&gt; input
echo '")' &gt;&gt; input
done
echo '(dump "$(.TARGET)")' &gt;&gt; input
lisp &lt; input
a.fasl : a.l init COMPILE
b.fasl : b.l init COMPILE
c.fasl : c.l init COMPILE
COMPILE : .USE
echo '(compile "$(.ALLSRC)")' &gt;&gt; input
init : .EXEC
echo '(load-system)' &gt; input</programlisting>
<literal>.EXEC</literal> sources, do not appear in the
local variables of targets that depend on them (nor are
they touched if <application>PMake</application> is
given the <option>-t</option>
flag). Note that all the rules, not just that for
system, include init as a source. This is because
none of the other targets can be made until init
has been made, thus they depend on it.</para></entry>
</row>
<row valign="top">
<entry><literal>.EXPORT</literal></entry>
<entry>This is used to mark those targets whose
creation should be sent to another machine if at
all possible. This may be used by some exportation
schemes if the exportation is expensive. You
should ask your system administrator if it is
necessary.</entry>
</row>
<row valign="top">
<entry><literal>.EXPORTSAME</literal></entry>
<entry>Tells the export system that the job
should be exported to a machine of the same
architecture as the current one. Certain
operations (e.g. running text through nroff) can be
performed the same on any architecture (CPU and
operating system type), while others (e.g. compiling
a program with cc) must be performed on a
machine with the same architecture. Not all export
systems will support this attribute.</entry>
</row>
<row valign="top">
<entry><literal>.IGNORE</literal></entry>
<entry>Giving a target the
<literal>.IGNORE</literal> attribute causes
<application>PMake</application> to ignore errors
from any of the target's commands, as if they all
had <literal>-</literal> before them.</entry>
</row>
<row valign="top">
<entry><literal>.INVISIBLE</literal></entry>
<entry><para>This allows you to specify one target as a
source for another without the one affecting the
other's local variables. Useful if, say, you
have a makefile that creates two programs, one
of which is used to create the other, so it must
exist before the other is created. You could say
<programlisting>prog1 : $(PROG1OBJS) prog2 MAKEINSTALL
prog2 : $(PROG2OBJS) .INVISIBLE MAKEINSTALL</programlisting>
where <literal>MAKEINSTALL</literal>
is some complex <literal>.USE</literal> rule (see
below) that depends on the <makevar>.ALLSRC</makevar>
variable containing the right things. Without the
<literal>.INVISIBLE</literal>
attribute for <maketarget>prog2</maketarget>,
the <literal>MAKEINSTALL</literal> rule
could not be applied. This is not as useful as it
should be, and the semantics may change (or the
whole thing go away) in the not-too-distant
future.</para></entry>
</row>
<row valign="top">
<entry><literal>.JOIN</literal></entry>
<entry><para>This is another way to avoid performing some
operations in parallel while permitting
everything else to be done so. Specifically it forces
the target's shell script to be executed only if
one or more of the sources was out-of-date. In
addition, the target's name, in both its
<makevar>.TARGET</makevar>
variable and all the local variables of any
target that depends on it, is replaced by the value
of its <makevar>.ALLSRC</makevar> variable. As an
example, suppose you have a program that has
four libraries that
compile in the same directory along with, and at
the same time as, the program. You again have
the problem with ranlib that I mentioned
earlier, only this time it is more severe: you can not
just put the ranlib off to the end since the
program will need those libraries before it can
be re-created. You can do something like this:
<programlisting>program : $(OBJS) libraries
cc -o $(.TARGET) $(.ALLSRC)
libraries : lib1.a lib2.a lib3.a lib4.a .JOIN
ranlib $(.OODATE)</programlisting>
In this case, <application>PMake</application> will re-create
the <literal>$(OBJS)</literal>
as necessary, along with <filename>lib1.a</filename>,
<filename>lib2.a</filename>, <filename>lib3.a</filename>
and <filename>lib4.a</filename>. It will then
execute ranlib on any library that was changed and set
program's <makevar>.ALLSRC</makevar> variable to contain
what's in <literal>$(OBJS)</literal>
followed by <quote><filename>lib1.a</filename>
<filename>lib2.a</filename>
<filename>lib3.a</filename>
<filename>lib4.a</filename>.</quote> In
case you are wondering, it is called
<literal>.JOIN</literal> because
it joins together different threads of the
<quote>input graph</quote> at the target marked
with the attribute. Another aspect of the
<literal>.JOIN</literal>
attribute is it keeps the target from being
created if the <option>-t</option> flag was
given.</para></entry>
</row>
<row valign="top">
<entry><literal>.MAKE</literal></entry>
<entry><para>The <literal>.MAKE</literal>
attribute marks its target as being a
recursive invocation of PMake. This forces
<application>PMake</application> to execute the
script associated with the
target (if it is out-of-date) even if you gave
the <option>-n</option> or
<option>-t</option> flag. By doing this, you can start
at the top of a system and type
<literallayout><command>pmake -n</command></literallayout>
and have it descend the directory tree (if your
makefiles are set up correctly), printing what
it would have executed if you had not included
the <option>-n</option> flag.</para></entry>
</row>
<row valign="top">
<entry><literal>.NOEXPORT</literal></entry>
<entry>If possible,
<application>PMake</application> will attempt to
export the creation of all targets to another machine
(this depends on how <application>PMake</application>
was configured). Sometimes, the creation is so
simple, it is pointless to send it to another machine.
If you give the target the
<literal>.NOEXPORT</literal> attribute, it will be run
loally, even if you have given
<application>PMake</application> the <option>-L
0</option> flag.</entry>
</row>
<row valign="top">
<entry><literal>.NOTMAIN</literal></entry>
<entry>Normally, if you do not specify a target to
make in any other way,
<application>PMake</application> will take the first
target on the first dependency line of a makefile as
the target to create. That target is known as the
<quote>Main Target</quote> and is labeled as such if
you print the dependencies out using the
<option>-p</option> flag. Giving a target this
attribute tells <application>PMake</application> that
the target is definitely not the Main Target. This
allows you to place targets in an included makefile
and have <application>PMake</application> create
something else by default.</entry>
</row>
<row valign="top">
<entry><literal>.PRECIOUS</literal></entry>
<entry>When <application>PMake</application> is
interrupted (you type control-C at the keyboard), it will
attempt to clean up after itself by removing any
half-made targets. If a target has the
<literal>.PRECIOUS</literal> attribute, however,
<application>PMake</application> will leave it alone.
An additional side effect of the <literal>::</literal>
operator is to mark the targets as
<literal>.PRECIOUS</literal>.</entry>
</row>
<row valign="top">
<entry><literal>.SILENT</literal></entry>
<entry>Marking a target with this attribute keeps its
commands from being printed when they are
executed, just as if they had an <literal>@</literal>
in front of them.</entry>
</row>
<row valign="top">
<entry><literal>.USE</literal></entry>
<entry><para>By giving a target this attribute, you turn it
into <application></application>PMake's equivalent of
a macro. When the target is
used as a source for another target, the other target
acquires the commands, sources and attributes (except
<literal>.USE</literal>) of the source. If the target
already has commands, the <literal>.USE</literal> target's
commands are added to the end. If more than one
<literal>.USE</literal>-marked source is given to a
target, the rules are applied sequentially. The typical
<literal>.USE</literal> rule (as I call them) will use
the sources of the target to which it is applied (as
stored in the <makevar>.ALLSRC</makevar> variable for
the target) as its <quote>arguments,</quote> if you
will. For example, you probably noticed that the
commands for creating <filename>lib1.a</filename> and
<filename>lib2.a</filename> in the example in section
<xref linkend="savingcmds"> were exactly the same.
You can use the <literal>.USE</literal> attribute to
eliminate the repetition, like so:
<programlisting>lib1.a : $(LIB1OBJS) MAKELIB
lib2.a : $(LIB2OBJS) MAKELIB
MAKELIB : .USE
rm -f $(.TARGET)
ar cr $(.TARGET) $(.ALLSRC)
...
ranlib $(.TARGET)</programlisting>
Several system makefiles (not to be confused
with The System Makefile) make use of these
<literal>.USE</literal> rules to make your life
easier (they are in the default, system makefile
directory...take a look). Note that the
<literal>.USE</literal> rule source itself
(<maketarget>MAKELIB</maketarget>) does not appear in
any of the targets's local variables. There is no limit
to the number of times I could use the
<maketarget>MAKELIB</maketarget> rule. If there were
more libraries, I could continue with
<literal>lib3.a : $(LIB3OBJS) MAKELIB</literal>
and so on and so forth.</para></entry>
</row>
</tbody>
</tgroup>
</informaltable>
</section>
<section id="specialtargets">
<title>Special Targets</title>
<para>As there were in <application>Make</application>, so there
are certain targets that have special meaning to
<application>PMake</application>. When you use one on a
dependency line,
it is the only target that may appear on the left-hand-side of the
operator. As for the attributes and variables, all the special
targets begin with a period and consist of upper-case letters
only. I will not describe them all in detail because some of them
are rather complex and I will describe them in more detail than you
will want in <xref linkend="gods">. The targets are as follows:</para>
<informaltable frame="none">
<tgroup cols="2">
<colspec colwidth="1*">
<colspec colwidth="10*">
<tbody>
<row valign="top">
<entry><maketarget>.BEGIN</maketarget></entry>
<entry>Any commands attached to this target are
executed before anything else is done. You can use
it for any initialization that needs
doing.</entry>
</row>
<row valign="top">
<entry><maketarget>.DEFAULT</maketarget></entry>
<entry>This is sort of a <literal>.USE</literal>
rule for any target (that was used only as a source)
that <application>PMake</application> can not figure
out any other way to create. It is only <quote>sort
of</quote> a <literal>.USE</literal> rule because
only the shell script attached to the
<maketarget>.DEFAULT</maketarget> target is used.
The <makevar>.IMPSRC</makevar> variable of a target
that inherits <maketarget>.DEFAULT</maketarget>'s
commands is set to the target's own
name.</entry>
</row>
<row valign="top">
<entry><maketarget>.END</maketarget></entry>
<entry>This serves a function similar to
<maketarget>.BEGIN</maketarget>, in that commands
attached to it are executed once everything
has been re-created (so long as no errors
occurred). It also serves the extra function of
being a place on which <application>PMake</application>
can hang commands you put off to the end. Thus the script
for this target will be executed before any of the
commands you save with the
<quote>...</quote>.</entry>
</row>
<row valign="top">
<entry><maketarget>.EXPORT</maketarget></entry>
<entry>The sources for this target are passed
to the exportation system compiled into
<application>PMake</application>. Some systems will use
these sources to configure themselves. You should ask
your system administrator about this.</entry>
</row>
<row valign="top">
<entry><maketarget>.IGNORE</maketarget></entry>
<entry>This target marks each of its sources
with the <literal>.IGNORE</literal> attribute.
If you do not give it any sources, then it is
like giving the <option>-i</option> flag when
you invoke <application>PMake</application> &ndash;
errors are ignored for all commands.</entry>
</row>
<row valign="top">
<entry><maketarget>.INCLUDES</maketarget></entry>
<entry><para>The sources for this target are taken to be
suffixes that indicate a file that can be included in
a program source file. The suffix must have
already been declared with <literal>.SUFFIXES</literal>
(see below).
Any suffix so marked will have the directories on
its search path (see <maketarget>.PATH</maketarget>,
below) placed in the <makevar>.INCLUDES</makevar>
variable, each preceded by a <option>-I</option> flag.
This variable can then be used as an argument for
the compiler in the normal fashion. The
<filename>.h</filename> suffix is already marked in
this way in the system makefile. E.g. if you have
<programlisting>.SUFFIXES : .bitmap
.PATH.bitmap : /usr/local/X/lib/bitmaps
.INCLUDES : .bitmap</programlisting>
<application>PMake</application> will place
<literal>-I/usr/local/X/lib/bitmaps</literal>
in the <makevar>.INCLUDES</makevar> variable and you can
then say
<programlisting>cc $(.INCLUDES) -c xprogram.c</programlisting>
(Note: the <makevar>.INCLUDES</makevar> variable is
not actually filled in until the entire makefile has
been read.)</para></entry>
</row>
<row valign="top">
<entry><maketarget>.INTERRUPT</maketarget></entry>
<entry>When <application>PMake</application> is
interrupted, it will execute the commands in the
script for this target, if it exists.</entry>
</row>
<row valign="top">
<entry><maketarget>.LIBS</maketarget></entry>
<entry>This does for libraries what
<maketarget>.INCLUDES</maketarget> does for include
files, except the flag used is
<option>-L</option>, as required by those linkers
that allow you to tell them where to find libraries.
The variable used is <makevar>.LIBS</makevar>.
Be forewarned that <application>PMake</application>
may not have been compiled to do this if the linker
on your system does not accept the <option>-L</option>
flag, though the <makevar>.LIBS</makevar> variable
will always be defined once the makefile has been
read.</entry>
</row>
<row valign="top">
<entry><maketarget>.MAIN</maketarget></entry>
<entry>If you did not give a target (or targets) to
create when you invoked
<application>PMake</application>, it will take the
sources of this target as the targets to
create.</entry>
</row>
<row valign="top">
<entry><maketarget>.MAKEFLAGS</maketarget></entry>
<entry>This target provides a way for you to
always specify flags for <application>PMake</application>
when the makefile is used. The flags are just as they
would be typed to the shell (except you can not use shell
variables unless they are in the environment), though
the <option>-f</option> and <option>-r</option>
flags have no effect.</entry>
</row>
<row valign="top">
<entry><maketarget>.NULL</maketarget></entry>
<entry>This allows you to specify what
suffix <application>PMake</application> should pretend
a file has if, in fact, it has no known suffix. Only
one suffix may be so designated. The last source on the
dependency line is the suffix that is used (you
should, however, only give one suffix...).</entry>
</row>
<row valign="top">
<entry><maketarget>.PATH</maketarget></entry>
<entry>If you give sources for this target,
<application>PMake</application> will take them as
directories in which to search for files it cannot
find in the current directory. If you give no
sources, it will clear out any directories added to
the search path before. Since the effects of this
all get very complex, we will leave it till <xref
linkend="gods"> to give you a complete
explanation.</entry>
</row>
<row valign="top">
<entry><maketarget>.PATH<replaceable>suffix</replaceable></maketarget></entry>
<entry>This does a similar thing to
<maketarget>.PATH</maketarget>, but it does it only
for files with the given suffix. The suffix must
have been defined already. Look at Search Paths
(<xref linkend="searchpaths">) for more
information.</entry>
</row>
<row valign="top">
<entry><maketarget>.PRECIOUS</maketarget></entry>
<entry>Similar to <maketarget>.IGNORE</maketarget>,
this gives the <literal>.PRECIOUS</literal> attribute to
each source on the dependency line, unless there are
no sources, in which case the <literal>.PRECIOUS</literal>
attribute is given to every target in the file.</entry>
</row>
<row valign="top">
<entry><maketarget>.RECURSIVE</maketarget></entry>
<entry>This target applies the <literal>.MAKE</literal>
attribute to all its sources. It does nothing if you
do not give it any sources.</entry>
</row>
<row valign="top">
<entry><maketarget>.SHELL</maketarget></entry>
<entry><application>PMake</application> is not
constrained to only using the Bourne shell to
execute the commands you put in the makefile. You
can tell it some other shell to use with this
target. Check out <quote><xref linkend="ashell"
endterm="ashelltitle"></quote> (<xref
linkend="ashell">) for more
information.</entry>
</row>
<row valign="top">
<entry><maketarget>.SILENT</maketarget></entry>
<entry>When you use
<maketarget>.SILENT</maketarget> as a target, it
applies the <literal>.SILENT</literal> attribute to
each of its sources. If there are no sources on the
dependency line, then it is as if you gave
<application>PMake</application> the
<option>-s</option> flag and no commands will be
echoed.</entry>
</row>
<row valign="top">
<entry><maketarget>.SUFFIXES</maketarget></entry>
<entry>This is used to give new file suffixes
for <application>PMake</application> to handle.
Each source is a suffix
<application>PMake</application> should
recognize. If you give a
<maketarget>.SUFFIXES</maketarget> dependency line
with no sources, <application>PMake</application>
will forget about all the suffixes it knew
(this also nukes the null suffix). For those
targets that need to have suffixes defined, this
is how you do it.</entry>
</row>
</tbody>
</tgroup>
</informaltable>
<para>In addition to these targets, a line of the form:</para>
<programlisting>attribute : sources</programlisting>
<para>applies the attribute to all the targets listed as sources.</para>
</section>
<section id="modyvarex">
<title>Modifying Variable Expansion</title>
<para>Variables need not always be expanded verbatim.
<application>PMake</application> defines several modifiers
that may be applied to a variable's value before it is expanded.
You apply a modifier by placing it after the variable name with
a colon between the two, like so:</para>
<programlisting>${VARIABLE:modifier}</programlisting>
<para>Each modifier is a single character followed by something
specific to the modifier itself. You may apply as many
modifiers as you want &ndash; each one is applied to the
result of the previous and is separated from the
previous by another colon.</para>
<para>There are seven ways to modify a variable's expansion,
most of which come from the C shell variable modification
characters:</para>
<variablelist>
<varlistentry>
<term><literal>Mpattern</literal></term>
<listitem>
<para>This is used to select only those words (a word is a
series of characters that are neither spaces nor tabs)
that match the given pattern. The pattern is a
wildcard pattern like that used by the shell, where
<literal>*</literal> means <literal>0</literal> or more
characters of any sort; <literal>?</literal> is any
single character; <literal>[abcd]</literal> matches any
single character that is either <literal>a</literal>,
<literal>b</literal>, <literal>c</literal> or
<literal>d</literal> (there may be any number of
characters between the brackets);
<literal>[0-9]</literal> matches any single character
that is between <literal>0</literal> and
<literal>9</literal> (i.e. any digit. This form may be
freely mixed with the other bracket form), and
<literal>\</literal> is used to escape any of the
characters <literal>*</literal>, <literal>?</literal>,
<literal>[</literal> or <literal>:</literal>, leaving
them as regular characters to match themselves in a
word. For example, the system makefile
<filename>&lt;makedepend.mk&gt;</filename> uses
<literal>$(CFLAGS:M-[ID]*)</literal> to extract all the
<option>-I</option> and <option>-D</option> flags that
would be passed to the C compiler. This allows it to
properly locate include files and generate the correct
dependencies.</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>Npattern</literal></term>
<listitem>
<para>This is identical to <literal>:M</literal> except
it substitutes all words that do not match the given
pattern.</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>S/search-string/replacement-string/[g]</literal></term>
<listitem>
<para>Causes the first occurrence of search-string in
the variable to be replaced by replacement-string,
unless the <option>g</option> flag is given at the end,
in which case all occurrences of the string are
replaced. The substitution is performed on each word in
the variable in turn. If search-string begins with a
<literal>^</literal>, the string must match starting at
the beginning of the word. If search-string ends with a
<literal>$</literal>, the string must match to the end
of the word (these two may be combined to force an exact
match). If a backslash precedes these two characters,
however, they lose their special meaning. Variable
expansion also occurs in the normal fashion inside both
the search-string and the replacement-string, except
that a backslash is used to prevent the expansion of a
<literal>$</literal>, not another dollar sign, as is
usual. Note that search-string is just a string, not a
pattern, so none of the usual regularexpression/wildcard
characters have any special meaning save
<literal>^</literal> and <literal>$</literal>. In the
replacement string, the <literal>&</literal> character
is replaced by the search-string unless it is preceded
by a backslash. You are allowed to use any character
except colon or exclamation point to separate the two
strings. This so-called delimiter character may be
placed in either string by preceding it with a
backslash.</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>T</literal></term>
<listitem>
<para>Replaces each word in the variable expansion by
its last component (its <quote>tail</quote>).
For example, given:</para>
<programlisting>OBJS = ../lib/a.o b /usr/lib/libm.a
TAILS = $(OBJS:T)</programlisting>
<para>the variable <makevar>TAILS</makevar> would expand
to <literal>a.o b libm.a.</literal></para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>H</literal></term>
<listitem>
<para>This is similar to <literal>:T</literal>, except
that every word is replaced by everything but the tail
(the <quote>head</quote>). Using the same definition of
<makevar>OBJS</makevar>, the string
<literal>$(OBJS:H)</literal> would expand to
<literal>../lib /usr/lib.</literal> Note that the final
slash on the heads is removed and anything without
a head is replaced by the empty string.</para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>E</literal></term>
<listitem>
<para><literal>:E</literal> replaces each word by its
suffix (<quote>extension</quote>). So
<literal>$(OBJS:E)</literal> would give you
<literal>.o .a.</literal></para>
</listitem>
</varlistentry>
<varlistentry>
<term><literal>R</literal></term>
<listitem>
<para>This replaces each word by everything but the
suffix (the <quote>root</quote> of the word).
<literal>$(OBJS:R)</literal> expands to
<literal>../lib/a b /usr/lib/libm</literal>.</para>
</listitem>
</varlistentry>
</variablelist>
<para>In addition, the System V style of substitution is also
supported. This looks like:</para>
<programlisting>$(VARIABLE:search-string=replacement)</programlisting>
<para>It must be the last modifier in the chain. The search is
anchored at the end of each word, so only suffixes or whole
words may be replaced.</para>
</section>
<section id="moreexercises">
<title>More Exercises</title>
<bridgehead>Exercise 3.1</bridgehead>
<para>You have got a set programs, each of which is created from
its own assembly-language source file (suffix
<filename>.asm</filename>). Each program can be assembled into
two versions, one with error-checking code assembled in and one
without. You could assemble them into files with different
suffixes (<filename>.eobj</filename> and
<filename>.obj</filename>, for instance), but your linker only
understands files that end in <filename>.obj</filename>. To top
it all off, the final executables must have the suffix
<filename>.exe</filename>. How can you still use
transformation rules to make your life easier (Hint: assume the
errorchecking versions have ec tacked onto their prefix)?</para>
<bridgehead>Exercise 3.2</bridgehead>
<para>Assume, for a moment or two, you want to perform
a sort of <quote>indirection</quote> by placing the name of
a variable into another one, then you want to get the value
of the first by expanding the second somehow. Unfortunately,
<application>PMake</application> does not allow constructs like:</para>
<programlisting>$($(FOO))</programlisting>
<para>What do you do? Hint: no further variable expansion is
performed after modifiers are applied, thus if you
cause a <literal>$</literal> to occur in the expansion,
that is what will be in the result.</para>
</section>
</chapter>
<!--
Local Variables:
mode: sgml
sgml-indent-data: t
sgml-omittag: nil
sgml-always-quote-attributes: t
sgml-parent-document: ("../book.sgml" "chapter")
End:
-->