章 10. 基本语法

目录
从 HTML 中分离
指令分隔符
注释

从 HTML 中分离

当 PHP 解析一个文件时，会寻找开始和结束标记，标记告诉 PHP 开始和停止解释其中的代码。此种方式的解析可以使 PHP 嵌入到各种不同的文档中，凡是在一对开始和结束标记之外的内容都会被 PHP 解析器忽略。大多数情况下 PHP 都是嵌入在 HTML 文档中的，如下例所示。


<p>This is going to be ignored.</p>
<?php echo 'While this is going to be parsed.'; ?>
<p>This will also be ignored.</p>

还可以用更高级的结构：
例子 10-1. 高级分离术
<?php if ($expression) { ?> <strong>This is true.</strong> <?php } else { ?> <strong>This is false.</strong> <?php } ?>
上例可正常工作，因为当 PHP 碰到结束标记 ?> 时，就简单地将其后的内容原样输出直到碰到下一个开始标记为止。当然，上面的例子很做作，但是对输出大块的文本而言，脱离 PHP 解析模式通常比将所有内容用 echo() 或者 print() 输出更有效率。

可以在 PHP 中使用四对不同的开始和结束标记。其中两种，<?php ?> 和 <script language="php"> </script> 总是可用的。另两种是短标记和 ASP 风格标记，可以在 php.ini 配置文件中打开或关闭。尽管有些人觉得短标记和 ASP 风格标记很方便，但移植性较差，通常不推荐。

注: 此外注意如果将 PHP 嵌入到 XML 或 XHTML 中则需要使用 <?php ?> 以保持符合标准。

例子 10-2. PHP 开始和结束标记
1. <?php echo 'if you want to serve XHTML or XML documents, do like this'; ?> 2. <script language="php"> echo 'some editors (like FrontPage) don\'t like processing instructions'; </script> 3. <? echo 'this is the simplest, an SGML processing instruction'; ?> <?= expression ?> This is a shortcut for "<? echo expression ?>" 4. <% echo 'You may optionally use ASP-style tags'; %> <%= $variable; # This is a shortcut for "<% echo . . ." %>

上例中的 1 和 2 总是可用的，其中 1 是最常用，并建议使用的。

短标记（上例 3）仅在通过 php.ini 配置文件中的指令 short_open_tag 打开后才可用，或者在 PHP 编译时加入了 --enable-short-tags 选项。

注: 如果用 PHP 3 还可以通过 short_tags() 函数激活使用短标记。此方法只适用于 PHP 3！

ASP 风格标记（上例 4）仅在通过 php.ini 配置文件中的指令 asp_tags 打开后才可用。

注: 对 ASP 风格标记的支持是 3.0.4 版添加的。

注: 在以下情况应避免使用短标记：开发需要发行的程序或者库，或者在用户不能控制的服务器上开发。因为目标服务器可能不支持短标记。为了代码的移植及发行，确保不要使用短标记。

add a note User Contributed Notes

alfridus
24-Jul-2006 07:53


Only this work:



<?php 

$xml = '<?xml version="1.0" encoding="UTF-8" standalone="no"?>';

echo $xml;

 ?>



with space after '<?php' and before ' ?>', no spacing between

'<?xml' and a semicolon after '"no"?>';'.

php at pixo dot sk
05-Apr-2006 07:58


just for conclusion:



<?php echo('<?xml version="1.0" ?'.'>'); ?>



is the right choice :)

brettz9 at yahoo dot com
02-Apr-2006 11:04


I've essentially tried to synthesize this

discussion at 

http://en.wikibooks.org/wiki/Programming:

Complete_PHP/Escaping_from_HTML



One point not brought up yet...



The HEREDOC problem with some text editors

may be fixable in at least some text editors

by adding a comment with a quotation mark

as such afterwards (albeit necessarily on a

new line):



<?php



$version = "1.0";



print <<<HERE

<?xml version="

HERE;

//"



print $version."\"?>";



?>

Christoph
17-Jan-2006 09:08


Here's an inspiration on how to quickly fix all scripts relying on short_open_tag being enabled:



find -name '*.php' | xargs perl -pi -e 's/<\?= ?(.*?) ?\?>/<?php echo($1); ?>/g'

find -name '*.php' | xargs perl -pi -e 's/<\?/<?php/g'

find -name '*.php' | xargs perl -pi -e 's/<?phpphp/<?php/g'

Michael Newton (http://mike.eire.ca/)
13-Dec-2005 07:17


The XML declaration does not need to be handled specially.



You should output it via an echo statement, in case your code is ever used on a server that is (poorly) configured to use short open tags.



But there's no need to treat the ?> at the end of the string specially.  That's because it's in a string.  The only thing PHP ever looks for in a string is \ or $ (the latter only in double-quoted strings.)



I have never had need for the following, as some have suggested below:



<?php

$xml=rawurldecode('%3C%3Fxml%20version%3D%221.0%22%3F%3E');

echo($xml);

?>



<?php echo '<?xml version="1.0" ?'.'>' ?>



<?php echo "<?xml version=\"1.0\"\x3F>" ?>

php [AT] jsomers [DOT] be
23-Sep-2005 03:37


PEAR states:



Always use <?php ?> to delimit PHP code, not the <? ?> shorthand. This is required for PEAR compliance and is also the most portable way to include PHP code on differing operating systems and setups.



It are these small things that enhance readability in group projects, or libraries.

pablo [] littleQ.net
24-Jul-2005 05:06


Just another more "feature" of IE...



Content-Disposition: attachment; filename=\"__FILE__\";



__FILE__ can't have spaces or :



Regards

01karlo at gmail dot com
26-Jun-2005 11:44


Or, use the following:



<?php

$xml=rawurldecode('%3C%3Fxml%20version%3D%221.0%22%3F%3E');

echo($xml);

?>



What is does it the value of the variable $xml is the RAW Url Encoded version of the XML thing.

Then it decodes it and echo it to the visitor.

p o r g e s at the gmail dot com server
02-Apr-2005 12:02


mike at skew dot org, I believe the differentiation is that "x"-"m"-"l" as a PI target is explicitly excluded from the definition of processing instructions.

Lachlan Hunt
29-Mar-2005 01:06


The person that suggested the use of this meta element above is wrong:



<meta http-equiv="Content-Type" content="application/xml+xhtml; charset=UTF-8" />



That meta element and the XML declaration serve completely different purposes, and that meta element should not be used.  Such information should be set using the HTTP Content-Type header (see the header() function).



Any XHTML page that just uses that meta element without proper HTTP Content-Type header, will be processed as text/html by browsers regardless, and when the HTTP headers do serve as application/xhtml+xml (or other XML MIME type), that charset parameter in the meta element will be ignored.

mike at skew dot org
22-Oct-2004 07:53


mart3862 mentions "XML processing instructions" and quotes their syntax from the spec, but is mistaken in using



<?xml version="1.0" ...?>



as an example. This little bit of markup that appears at the beginning of an XML file is in fact not a processing instruction at all; it is an "XML declaration" -- or, if it appears in an entity other than the main document, a "text declaration". All three constructs are formatted slightly differently, although they all do begin and end with the same.



The difference between a processing instruction, an XML declaration, or a text declaration is more than just a matter of subtle differences in syntax, though. A processing instruction embodies exactly two opaque, author-defined pieces of information (a 'target' and an 'instruction') that are considered to be part of the document's logical structure and that are thus made available to an application by the XML parser. An XML or text declaration, on the other hand, contains one to three specific pieces of information (version, encoding, standalone status), each with a well-defined meaning. This info provides cues to the parser to help it know how to read the file; it is not considered part of the document's logical structure and is not made available to the application.

stooges_cubed at racerx dot net
21-Oct-2004 04:13


In the note above about escaping XML/PHP style <?xml tags, the following code was used:





<?php  // Html safe containers





   echo <<<EOD


<?xml version="1.0"?>


...all sorts of XML goes here...


Nothing will affect the output of this code until:


EOD;


?>





EOD is just an example stop/start name.





This works too:





<?php  // Html safe containers





  $myOutput = <<<MYHTMLSAFEOUTPUT


<?xml version="1.0"?>


<html>


  <title>PHP Example</title>


  <body>


   <p>...all sorts goes here...</p>


  </body>


</html>


MYHTMLSAFEOUTPUT;





echo $myOutput;





?>





Only disadvantage of using this is that all the code highlighting programs I've seen never get it right, making your code look eronous in the majority of viewers.





Another alternative is to keep the XML / HTML in a separate include file and read in when needed. I don't know how efficient/inefficient this is for (idiots like yourselves) small amounts of text.





xmlheader.txt:


<?xml version="1.0"?>





mypage.php:


<?php


  include("xmlheader.txt");


?>

crtrue at coastal dot edu
01-May-2004 02:02


Although you can use the above methods to pass a document off as a valid for the W3C parser, a simpler-and-perfectly-legal method of doing so is to simple declare the document type in a meta tag. Something along these lines (mind the values in 'content' - I haven't personally used the Content-Type method in awhile):



<meta http-equiv="Content-Type" content="application/xml+xhtml; charset=UTF-8" />



Of course if you're using just XML, and don't use such functions, then the above methods will work just as fine.

mart3862 at yahoo dot com dot au
19-Apr-2004 12:29


Now the ultimate truth on how you should output xml processing instructions:



There have been several posts suggesting ways to include the text <?xml version="1.0" encoding="utf-8"?> in your output when short_tags is turned on, but only the following should be used:



<?php echo '<?xml version="1.0" ?'.'>' ?>

or

<?php echo "<?xml version=\"1.0\"\x3F>" ?>



Using one of these methods, and not making use of short tags, means your source code will also be a valid XML document, which allows you to do many things with it such as validation, XSLT translations, etc, as well as allowing your text editor to parse your code for syntax colouring.  Every PHP tag will simply be interpreted as an XML processing instruction (commonly referred to as PI).



The reason why all the other suggested methods are not advisable is because they contain the characters ?> inside the PHP tag, which the XML parser will interpret as the end of the processing instruction.



A processing instruction is defined in XML as:



PI ::= '<?' PITarget (S (Char* - (Char* '?>' Char*)))? '?>' 



In other words, it explicitly forbids the characters ?> to occur together within a processing instruction, unless they are delimiting the end of the tag.  It also requires a PITarget (an identifier starting with a letter) immediately after the initial start delimiter, which means that all short tag formats are also invalid XML.



Following these guidelines will result in code that is portable to servers with any configuration and allow you perform many useful tasks on your XML or XHTML source documents.  Even if you do not intend to validate or translate your source documents, and you can ignore some incorrect syntax colouring in your text editor, it is still best to get into good habits early.

Anon
22-Feb-2004 10:05


Yet another way of adding the XML processing instruction is to use:



<?php echo '<?xml version="1.0" ?'.'>' ?>



Because the ? and > are separated, the parser will not terminate before it is supposed to.



As a side note, the W3C's parser seems to recognise this method (assuming it even checks for the PI).

TarquinWJ
06-Feb-2004 10:54


Not spotted any messages like this one - delete it if there was one.



My hosting service allows <? and ?>, but I like to use valid XHTML, so I came up with this simple solution:



It is possible to use the short tags <? ?> with XHTML or XML documents. The only problem is that X(HT)ML requires a declaration using <? and ?>



<?xml version="1.0" encoding="UTF-8"?>



To avoid the problem, simply replace <? with <<? ?>?

and ?> with ?<? ?>>



<<? ?>?xml version="1.0" encoding="UTF-8"?<? ?>>



This inserts a blank piece of PHP in between the < and ?, and when parsed will output the regular tag

<?xml version="1.0" encoding="UTF-8"?>

mwild at iee dot NO_SP_AM dot org
19-Dec-2003 09:12


The text between <script> and </script> in XHTML is PCDATA, so <  and & characters in it should be interpreted as markup. This is a bit limiting for PHP, which is often used to output tags, though you can of course use &lt; and &amp; instead. To avoid that, which makes your code look peculiar and is easy to forget to do, you can mark the PHP as CDATA, eg :



<script language="PHP">

//<![CDATA[ 

echo('Today is <b>'.date('l F jS').'</b>');

//]]>

</script>



If you don't do this, and your code contains < or &, it should be rejected by an XHTML validator.

johnbeech at (not saying) mkv25 dot net
08-Dec-2003 08:42


In the note above about escaping XML/PHP style <?xml tags, the following code was used:



<?php  // Html safe containers



   echo <<<EOD

<?xml version="1.0"?>

...all sorts of XML goes here...

Nothing will affect the output of this code until:

   EOD;

?>



EOD is just an example stop/start name.



This works too:



<?php  // Html safe containers



  $myOutput = <<<MYHTMLSAFEOUTPUT

<?xml version="1.0"?>

<html>

  <title>PHP Example</title>

  <body>

    <p>...all sorts goes here...</p>

  </body>

</html>

MYHTMLSAFEOUTPUT;



echo $myOutput;



?>



Only disadvantage of using this is that all the code highlighting programs I've seen never get it right, making your code look eronous in the majority of viewers.



Another alternative is to keep the XML / HTML in a separate include file and read in when needed. I don't know how efficient/inefficient this is for small amounts of text.



xmlheader.txt:

<?xml version="1.0"?>



mypage.php:

<?php

  include("xmlheader.txt");

?>

dave at [nospam] dot netready dot biz
18-Mar-2002 08:21


A little "feature" of PHP I've discovered is that the <?PHP token requires a space after it whereas after the <? and <% tokens a space is optional.



The error message you get if you miss the space is not too helpful so be warned!



(These examples only give a warning with error_reporting(E_ALL) )



<?PHP/*<Some HTML>*/?> fails...

<?/*<Some HTML>*/?> works...

mrtidy at mail dot com
12-Dec-2001 04:36


[Ed Note:


This is because of short_tags, <?xml turns php parsing on, because of the <?.


--irc-html@php.net]





I am moving my site to XHTML and I ran into trouble with the <?xml ?> interfering with the <?php ?> method of escaping for HTML.  A quick check of the mailing list confirmed that the current preferred method to cleanly output the <?xml ?> line is to echo it:<br>


<?php echo("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n"); ?>

add a note