<?xml version="1.0" encoding="iso-8859-1"?> <!-- $Revision: 1.31 $ --> <chapter id="faq.using"> <title>Using PHP</title> <titleabbrev>Using PHP</titleabbrev> <para> This section gathers many common errors that you may face while writing PHP scripts. </para> <qandaset> <qandaentry id="faq.using.anyform"> <question> <para> I would like to write a generic PHP script that can handle data coming from any form. How do I know which POST method variables are available? </para> </question> <answer> <para> PHP offers many <link linkend="language.variables.predefined"> predefined variables</link>, like the superglobal <varname> $_POST</varname>. You may loop through <varname>$_POST</varname> as it's an associate array of all POSTed values. For example, let's simply loop through it with <link linkend="control-structures.foreach"> foreach</link>, check for <function>empty</function> values, and print them out. <programlisting role="php"> <![CDATA[ <?php $empty = $post = array(); foreach ($_POST as $varname => $varvalue) { if (empty($varvalue)) { $empty[$varname] = $varvalue; } else { $post[$varname] = $varvalue; } } print "<pre>"; if (empty($empty)) { print "None of the POSTed values are empty, posted:\n"; var_dump($post); } else { print "We have " . count($empty) . " empty values\n"; print "Posted:\n"; var_dump($post); print "Empty:\n"; var_dump($empty); exit; } ?> ]]> </programlisting> </para> ¬e.superglobals; </answer> </qandaentry> <qandaentry id="faq.using.addslashes"> <question> <para> I need to convert all single-quotes (') to a backslash followed by a single-quote (\'). How can I do this with a regular expression? I'd also like to convert " to \" and \ to \\. </para> </question> <answer> <para> The function <function>addslashes</function> will do this. See also <function>mysql_escape_string</function>. You may also strip backslashes with <function>stripslashes</function>. </para> ¬e.magicquotes.gpc; </answer> </qandaentry> <qandaentry id="faq.using.stripslashes"> <question> <para> All my " turn into \" and my ' turn into \', how do I get rid of all these unwanted backslashes? How and why did they get there? </para> </question> <answer> <para> The PHP function <function>stripslashes</function> will strip those backslashes from your <type>string</type>. Most likely the backslashes magically exist because the PHP directive <link linkend="ini.magic-quotes-gpc">magic_quotes_gpc</link> is on. </para> ¬e.magicquotes.gpc; </answer> </qandaentry> <qandaentry id="faq.using.wrong-order"> <question> <para> When I do the following, the output is printed in the wrong order: <programlisting role="php"> <![CDATA[ <?php function myfunc($argument) { echo $argument + 10; } $variable = 10; echo "myfunc($variable) = " . myfunc($variable); ?> ]]> </programlisting> what's going on? </para> </question> <answer> <para> To be able to use the results of your function in an expression (such as concatenating it with other strings in the example above), you need to <function>return</function> the value, not <function>echo</function> it. </para> </answer> </qandaentry> <qandaentry id="faq.using.newlines"> <question> <para> Hey, what happened to my newlines? <programlisting role="php"> <![CDATA[ <pre> <?php echo "This should be the first line."; ?> <?php echo "This should show up after the new line above."; ?> </pre> ]]> </programlisting> </para> </question> <answer> <para> In PHP, the ending for a block of code is either "?>" or "?>\n" (where \n means a newline). So in the example above, the echoed sentences will be on one line, because PHP omits the newlines after the block ending. This means that you need to insert an extra newline after each block of PHP code to make it print out one newline. </para> <para> Why does PHP do this? Because when formatting normal HTML, this usually makes your life easier because you don't want that newline, but you'd have to create extremely long lines or otherwise make the raw page source unreadable to achieve that effect. </para> </answer> </qandaentry> <qandaentry id="faq.using.headers-sent"> <question> <para> I get the message 'Warning: Cannot send session cookie - headers already sent...' or 'Cannot add header information - headers already sent...'. </para> </question> <answer> <para> The functions <function>header</function>, <function>setcookie</function>, and the <link linkend="ref.session">session functions</link> need to add headers to the output stream but headers can only be sent before all other content. There can be no output before using these functions, output such as HTML. The function <function> headers_sent</function> will check if your script has already sent headers and see also the <link linkend="ref.outcontrol">Output Control functions</link>. </para> </answer> </qandaentry> <qandaentry id="faq.using.header"> <question> <para> I need to access information in the request header directly. How can I do this? </para> </question> <answer> <para> The <function>getallheaders</function> function will do this if you are running PHP as an Apache module. So, the following bit of code will show you all the request headers: <programlisting role="php"> <![CDATA[ <?php $headers = getallheaders(); foreach ($headers as $name => $content) { echo "headers[$name] = $content<br />\n"; } ?> ]]> </programlisting> </para> <para> See also <function>apache_lookup_uri</function>, <function>apache_response_headers</function>, and <function>fsockopen</function> </para> </answer> </qandaentry> <qandaentry id="faq.using.authentication"> <question> <para> When I try to use authentication with IIS I get 'No Input file specified'. </para> </question> <answer> <para> The security model of IIS is at fault here. This is a problem common to all CGI programs running under IIS. A workaround is to create a plain HTML file (not parsed by PHP) as the entry page into an authenticated directory. Then use a META tag to redirect to the PHP page, or have a link to the PHP page. PHP will then recognize the authentication correctly. With the ISAPI module, this is not a problem. This should not effect other NT web servers. For more information, see: <ulink url="&url.iis;">&url.iis;</ulink> and the manual section on <link linkend="features.http-auth">HTTP Authentication </link>. </para> </answer> </qandaentry> <qandaentry id="faq.using.netscape"> <question> <para> My PHP script works on IE and Lynx, but on Netscape some of my output is missing. When I do a "View Source" I see the content in IE but not in Netscape. </para> </question> <answer> <para> Netscape is more strict regarding HTML tags (such as tables) then IE. Running your HTML output through a HTML validator, such as <ulink url="&url.w3.validator;">validator.w3.org</ulink>, might be helpful. For example, a missing </table> might cause this. </para> <para> Also, both IE and Lynx ignore any NULs (<literal>\0</literal>) in the HTML stream, Netscape does not. The best way to check for this is to compile the <link linkend="features.commandline">command line</link> version of PHP (also known as the CGI version) and run your script from the command line. In *nix, pipe it through <literal>od -c</literal> and look for any <literal>\0</literal> characters. If you are on Windows you need to find an editor or some other program that lets you look at binary files. When Netscape sees a NUL in a file it will typically not output anything else on that line whereas both IE and Lynx will. </para> </answer> </qandaentry> <qandaentry id="faq.using.mixml"> <question> <para> How am I supposed to mix XML and PHP? It complains about my <?xml tags! </para> </question> <answer> <para> In order to embed <?xml straight into your PHP code, you'll have to turn off short tags by having the PHP directive <link linkend="ini.short-open-tag">short_open_tags</link> set to <literal>0</literal>. You cannot set this directive with <function> ini_set</function>. Regardless of <link linkend="ini.short-open-tag"> short_open_tags</link> being on or off, you can do something like: <literal><?php echo '<?xml'; ?></literal>. The default for this directive is on. </para> </answer> </qandaentry> <qandaentry id="faq.using.editor"> <question> <para> How can I use PHP with FrontPage or some other HTML editor that insists on moving my code around? </para> </question> <answer> <para> One of the easiest things to do is to enable using ASP tags in your PHP code. This allows you to use the ASP-style <% and %> code delimiters. Some of the popular HTML editors handle those more intelligently (for now). To enable the ASP-style tags, you need to set the <link linkend="ini.asp-tags">asp_tags</link> &php.ini; variable, or use the appropriate Apache directive. </para> </answer> </qandaentry> <qandaentry id="faq.using.variables"> <question> <para> Where can I find a complete list of variables are available to me in PHP? </para> </question> <answer> <para> Read the manual page on <link linkend="language.variables.predefined"> predefined variables</link> as it includes a partial list of predefined variables available to your script. A complete list of available variables (and much more information) can be seen by calling the <function>phpinfo</function> function. Be sure to read the manual section on <link linkend="language.variables.external">variables from outside of PHP</link> as it describes common scenarios for external variables, like from a HTML form, a Cookie, and the URL. </para> ¬e.registerglobals; </answer> </qandaentry> <qandaentry id="faq.using.freepdf"> <question> <para> How can I generate PDF files without using the non-free and commercial libraries <link linkend="ref.cpdf">ClibPDF</link> and <link linkend="ref.pdf">PDFLib</link>? I'd like something that's free and doesn't require external PDF libraries. </para> </question> <answer> <para> There are a few alternatives written in PHP such as <ulink url="&url.pdf.ros;">&url.pdf.ros;</ulink>, <ulink url="&url.pdf.fpdf;">&url.pdf.fpdf;</ulink>, <ulink url="&url.pdf.pdf4php;">&url.pdf.pdf4php;</ulink>, and <ulink url="&url.pdf.phppdflib;">&url.pdf.phppdflib;</ulink>. There is also the <ulink url="&url.pdf.panda;">Panda</ulink> module. </para> </answer> </qandaentry> <qandaentry id="faq.using.cgi-vars"> <question> <para> I'm trying to access one of the standard CGI variables (such as $DOCUMENT_ROOT or $HTTP_REFERER) in a user-defined function, and it can't seem to find it. What's wrong? </para> </question> <answer> <para> It's important to realize that the PHP directive <link linkend="ini.register-globals">register_globals</link> also affects server and environment variables. When register_globals = off (the default is off since PHP 4.2.0), <varname>$DOCUMENT_ROOT</varname> will not exist. Instead, use <varname>$_SERVER['DOCUMENT_ROOT'] </varname>. If register_globals = on then the variables <varname>$DOCUMENT_ROOT</varname> and <varname>$GLOBALS['DOCUMENT_ROOT']</varname> will also exist. </para> <para> If you're sure register_globals = on and wonder why <varname>$DOCUMENT_ROOT</varname> isn't available inside functions, it's because these are like any other variables and would require <literal>global $DOCUMENT_ROOT</literal> inside the function. See also the manual page on <link linkend="language.variables.scope">variable scope</link>. It's preferred to code with register_globals = off. </para> ¬e.superglobals; </answer> </qandaentry> <qandaentry id="faq.using.shorthandbytes"> <question> <para> A few PHP directives may also take on shorthand byte values, as opposed to only <type>integer</type> byte values. What are all the available shorthand byte options? And can I use these outside of &php.ini;? </para> </question> <answer> <para> The available options are K (for Kilobytes) and M (for Megabytes), these are case insensitive. Anything else assumes bytes. <literal>1M</literal> equals one Megabyte or <literal>1048576</literal> bytes. <literal>1K</literal> equals one Kilobyte or <literal>1024</literal> bytes. You may not use these shorthand notations outside of &php.ini;, instead use an <type>integer</type> value of bytes. See the <function>ini_get</function> documentation for an example on how to convert these values. </para> </answer> </qandaentry> </qandaset> </chapter> <!-- Keep this comment at the end of the file Local variables: mode: sgml sgml-omittag:t sgml-shorttag:t sgml-minimize-attributes:nil sgml-always-quote-attributes:t sgml-indent-step:1 sgml-indent-data:t indent-tabs-mode:nil sgml-parent-document:nil sgml-default-dtd-file:"../../manual.ced" sgml-exposed-tags:nil sgml-local-catalogs:nil sgml-local-ecat-files:nil End: vim600: syn=xml fen fdm=syntax fdl=2 si vim: et tw=78 syn=sgml vi: ts=1 sw=1 -->