ch02.htm

Web_Programming_with_Perl5,一个不错的Perl语言教程。

		<TD WIDTH="97" ALIGN="LEFT">PERLLOL</TD>

		<TD ALIGN="LEFT">Perl data structures: lists of lists</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLOBJ</TD>

		<TD ALIGN="LEFT">Perl objects</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLTIE</TD>

		<TD ALIGN="LEFT">Perl objects hidden behind simple variables</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLBOT</TD>

		<TD ALIGN="LEFT">Perl OO tricks and examples</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLIPC</TD>

		<TD ALIGN="LEFT">Perl interprocess communication</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLDEBUG</TD>

		<TD ALIGN="LEFT">Perl debugging</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLDIAG</TD>

		<TD ALIGN="LEFT">Perl diagnostic messages</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLSEC</TD>

		<TD ALIGN="LEFT">Perl security</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLTRAP</TD>

		<TD ALIGN="LEFT">Perl traps for the unwary</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLSTYLE</TD>

		<TD ALIGN="LEFT">Perl style guide</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLXS</TD>

		<TD ALIGN="LEFT">Perl XS application programming interface</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLXSTUT</TD>

		<TD ALIGN="LEFT">Perl XS tutorial</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLGUTS</TD>

		<TD ALIGN="LEFT">Perl internal functions for creating extensions</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLCALL</TD>

		<TD ALIGN="LEFT">Perl calling conventions from C</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLEMBED</TD>

		<TD ALIGN="LEFT">Perl: how to embed Perl in your C or C++ application</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLPOD</TD>

		<TD ALIGN="LEFT">Perl plain old documentation</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLAPIO</TD>

		<TD ALIGN="LEFT">Perl internal IO abstraction interface</TD>

	</TR>

	<TR ALIGN="LEFT" rowspan="1">

		<TD WIDTH="97" ALIGN="LEFT">PERLBOOK</TD>

		<TD ALIGN="LEFT">Perl book information</TD>

	</TR>

</TABLE>

<BR>

<BR>

The documentation also contains numerous additional sections corresponding to the

standard modules that ship with Perl. Most or all of these additional sections are

extracted from the embedded POD, which is to be found in the module file itself.

All Perl documentation is written first in POD and then translated to other formats.



<DL>

	<DT></DT>

</DL>



<P><B><TT>POD</TT></B>





<BLOCKQUOTE>

	<P>Plain Old Documentation. ASCII text documentation with markers corresponding to

	the various formatting elements. Can be embedded directly into Perl modules. See

	PERLPOD.



</BLOCKQUOTE>



<P>You can easily transform POD into standard UNIX *ROFF format, HTML, and a number

of other formats by using the pod2* (pod2man, pod2html, pod2text, and so on) converters.

There exist POD converters to many other types of formats, as well. The POD format

also implies that you can read the documentation directly, without any post-formatting

at all. Everything that I cover in this chapter is also documented in the Perl PODs,

and I give references to the specific sections as I go along, using the PERLBLAH

notation as previously mentioned. POD can be embedded directly into a Perl module,

or program, and nearly all of the modules have them already.</P>

<P>When Perl is installed on a typical UNIX site, the POD documentation, including

POD from the modules, is converted automatically into standard UNIX manpages. The

administrator usually installs it inside the primary @INC directory, usually in a

subdirectory called man. Macintosh Perl installations have the POD, converted to

HTML format, in a folder beneath the folder that contains the Perl application, named

pod. The Windows (ntperl) installation also has the documentation, converted to HTML,

in the directory called docs. You should find this directory, and be ready to refer

to the documentation within it.</P>

<P>As I mentioned previously, conversion tools are available for POD, including pod2html,

pod2text, pod2rtf, pod2tex, pod2inf, and now even pod2ps, and the pod2man program.

You can use any of them to convert from POD format to your preferred format, provided

that the tool has been written to work on your architecture. All these tools work

on UNIX, and some are configurable to work on other platforms.



<DL>

	<DT></DT>

</DL>



<P><B>A Note on Compatibility</B>





<BLOCKQUOTE>

	<P>There are, as you might guess, still a few incompatibilities when attempting to

	produce cross-platform Perl code, and not all scripts run on all architectures by

	default. I'll try to note these, as I progress through this chapter and through the

	book. Filepaths, for instance, shouldn't be hardcoded in Perl programs, but often

	are. This issue in general is being considered and worked on by the Perl Porters,

	a large group of brilliant people who help to bring you Perl.



</BLOCKQUOTE>



<P>Remember to check your favorite CPAN to get the latest versions of the pod2* programs.



<DL>

	<DT></DT>

</DL>



<P><B><TT>CPAN</TT></B>





<BLOCKQUOTE>

	<P>Comprehensive Perl Archive Network. A large group of well-connected Internet sites

	that maintains a copy of the master Perl archive. You can find more details later

	in this chapter and a complete history and description of the CPAN in Chapter 1.



</BLOCKQUOTE>



<P><B><TT>Readability Improvements</TT></B> The ability to provide easily readable

and reusable code has become more important as the level of formal training and skills

required to start a Web site has decreased. The responsibility is largely up to the

script author to implement this readability. Perl5 also provides some new functionality

that enhances the capability of the script author to do so.</P>

<P>The English module increases the readability and understanding of Perl code, and

it is a big step toward alleviating the boggling effect that raw Perl code sometimes

has on new programmers. The English module provides a mapping between Perl's eclectic

punctuation (special) variables with an English name corresponding to each one. The

regular-expression variables that correspond to the three components of a matched

string, for example, are often difficult to remember, even for the experienced Perl

programmer. The English module maps these variables as follows:</P>

<PRE><FONT COLOR="#0066FF">*MATCH = *&amp; ;

*PREMATCH = *';

*POSTMATCH = *';

</FONT></PRE>

<P>Thus, when you use the English module in your program, you can use $MATCH, $PREMATCH,

and $POSTMATCH instead of using $&amp;, $', or $' and chasing through the manual

to verify whether you need an ampersand, backtick, or single quotation mark following

the $, each time you want to access these built-in variables. See the complete English.pm

module in @INC, and its embedded POD documentation, or English.3, the POD converted

to a manpage, for more details. <B><TT>New Logical Operators</TT></B> The new logical

operators and, or, and not enable you to avoid using the &amp;&amp;, ||, and unary

! operators, respectively. The new operators are definitely more readable, to the

casual observer at least. The former also have lower precedence than the comma, and

certain other low-precedence operators. Consider the following:</P>

<PRE><FONT COLOR="#0066FF">$foo/=0 || print &quot;aak&quot;;



# prints: aak

</FONT></PRE>

<P>Using ||, it looks like you can divide by 0. (Actually you're dividing by the

||'d value of zero and one, the return value from print.) Now consider the same example,

using the or operator:</P>

<PRE><FONT COLOR="#0066FF">$foo/=0 or print &quot;aak&quot;;



# prints: Illegal division by zero

</FONT></PRE>

<P>Using or (and Perl5) gives the expected result. See PERLOP for more details on

operator precedence. <B><TT>Warnings and Stricture</TT></B> Other usability enhancements

in Perl5 include improvements to the -w command-line switch, which now gives more

useful and informative output. You should use it with all your programs. See PERLRUN

for a complete description of all command-line switches.</P>

<P>Also new is a set of pragmatic modules, which impose certain restrictions and

perform extended type and syntax checking on your code at compile time, to potentially

help you find bugs before they bite. Use of these pragmas within your code is also

highly recommended. See PERLDSC and PERLMOD for more details on the motivation for

and impact of using the strict modules. <B><TT>The New </TT>=&gt;<TT> Operator</TT></B>

The =&gt; operator is syntactic sugar for a comma. It makes certain declarations

look prettier and appear more sensible, as in the following example:</P>

<PRE><FONT COLOR="#0066FF">%hash = (

    `Name' =&gt; `Joe',

    `Address' =&gt; `123 Foo Street',

    `City' =&gt; `San Francisco',

    `State' =&gt; `CA'

);

</FONT></PRE>

<P>Note how we used the comma at the end of each hash key/value pair, but the =&gt;

operator between them. This makes such declarations easier to read, especially when

declaring more complex data structures. <B><TT>Function Prototypes</TT></B> Function

prototypes are one of the very latest new features in Perl5. They were finally included

as of Perl5.002, after a great deal of consideration and discussion. Essentially,

they provide you with a means to assure that the correct arguments are passed to

your subroutines, to emulate the behavior of built-in commands, for instance. See

PERLSUB for more details.

<CENTER>

<H4><A NAME="Heading6"></A><FONT COLOR="#000077">Lexical Scoping</FONT></H4>

</CENTER>

<P>The new my() operator enables you to declare variables that are truly lexically

scoped. Previously, the closest you could get to a lexical variable was with local().

The my() variables are visible only within the current block (within a set of curly

braces) and go out of scope immediately after exiting the block. The following example

illustrates their use:</P>

<PRE><FONT COLOR="#0066FF">$foo = &quot;I\'m foo in a global context\n&quot;;

print $foo;



{  # the curlies open a new lexical context (block)

    my $foo = &quot;But I\'m bar in this short lexical context\n&quot;;

    print $foo;

} # lexically scoped $foo goes out of scope here



print $foo;



# prints:

I'm foo in a global context

But I'm bar in this short lexical context

I'm foo in a global context

</FONT></PRE>

<P>The my() variable is used extensively in the modules that you'll explore in this

book. See PERLVAR for complete details on all types of Perl variables, including

lexical variables.

<CENTER>

<H4><A NAME="Heading7"></A><FONT COLOR="#000077">References</FONT></H4>

</CENTER>

<P>Standard Perl scalars can now be used to refer to other data types. References

act much like pointers in C. In the general sense, they simply refer to other data

types. When a reference is blessed into a package, it becomes a Perl Object, and

can be used to invoke the methods of its package (or class), as well as to access

the instance variables of the class, something akin to C++ references.</P>

<P>Because we'll explore the references in depth later in this chapter, I'll defer

most of the examples, and a complete description of them until that time. I'll also

explain how the bless() operator works at that time. The reference datatype is also

used extensively in the examples in this book.

<CENTER>

<H4><A NAME="Heading8"></A><FONT COLOR="#000077">Data Structures</FONT></H4>

</CENTER>

<P>Nested data types and other data structures are now possible, using references.