Tidy
在线手册：中文 英文
PHP手册

Tidy 函数

Table of Contents

Tidy
在线手册：中文 英文
PHP手册
PHP手册 - N: Tidy 函数

用户评论:

badtzmari at gmail dot com (03-Dec-2007 03:52)

The manual doesn't really make it clear that the PECL ext is only for php4. For php5, you have to use the --with-tidy option. (At least, this was the case for me with php 5.2.5 on Mac OS 10.4 - Tiger.) Tiger ships with tidylibs already installed in /usr/include but there is broken header file. The easiest way to get php to compile with Tidy on 10.4 (and 10.5 Leopard also) is to download Macports and use that to install tidy (unless you want to build tidy from src.) After installing Macports via the .dmg, just su root and do:

port install tidy

(port is installed in /opt/local/bin)

Then configure --with-tidy=/opt/local (Macports installs stuff in /opt/local).

You can also use Macports to install all sorts of other libs such as png, libmcrypt, freetype and jpeg (although libpng and libjpeg are also available as package installs.)

The specific compile error is:

In file included from /usr/include/tidy/tidy.h:70,
                 from /Users/mari/Downloads/php-5.2.5/ext/tidy/
tidy.c:34:
/usr/include/tidy/platform.h:515: error: duplicate 'unsigned'
/usr/include/tidy/platform.h:515: warning: useless type name in empty declaration

Shuster (29-Oct-2007 07:45)

Valid XHTML STRICT

<?php

if (function_exists('tidy_repair_string'))
{
    $xhtml = tidy_repair_string($xhtml, array('output-xhtml' => true, 'show-body-only' => true, 'doctype' => 'strict', 'drop-font-tags' => true, 'drop-proprietary-attributes' => true, 'lower-literals' => true, 'quote-ampersand' => true, 'wrap' => 0), 'raw');
}

?>

Guillaume (15-May-2007 12:36)

To install correctly Tidy for PHP5 on Ubuntu, follow this link :
http://ubuntuforums.org/showthread.php?t=195636

In fact, you need to run a "make clean" before the commands "make" and "make install"

matteo dot contri at tecem dot it (12-Oct-2006 09:44)

i had many problem with a javascript that grab mouse event on image and tidy (obviously).
I found this solution:

'output-xhtml'  => false

and everything is working again!

info att tcknetwork doot com (15-Jul-2006 11:23)

I have been searching for an easy way to check an entire website against HTML/XHTML formatting (no error, compilant, etc.), tidy is very useful for that :
<?php

/** aready checked pages */
$e=array();
/** webpages to check */
$t=array("/web/test.com/");
/** forbidden extensions (typically linked ressources) */
$x=explode(",","jpg,gif,png,doc,xls,pdf");

echo "<pre>";
while ($t[0]) {
 // already checked or a ressource => skip
 if (in_array($t[0],$e) || in_array(substr($t[0],-3),$x)) array_shift($t);
 else { $c=array_shift($t); $e[]=$c; $t=array_merge($t,ck($c)); }
}
echo "</pre>";

/**
 check_vailidty($url,$server)
 return : list of the internal links of the page
*/
function ck($u,$s="http://127.0.0.1") {
 $c=array("indent"=>1,"output-xhtml"=>1,"accessibility-check"=>3);
 $t=tidy_parse_string(file_get_contents($s.$u),$c);
 tidy_clean_repair($t);
 if (tidy_error_count($t)) { // we have error, display them
  echo "FAIL ".htmlentities($u)." (".tidy_error_count($t)." errors)\n";
  echo htmlentities(tidy_get_error_buffer($t))."\n";
 } else { // all right
  echo "OK ".htmlentities($u)."\n";
 }
 // return all the links inside the page
 return gl(tidy_get_root($t),substr($u,-1)=="/"?$u:dirname($u)."/");
}
/**
 get_links($tinynode,$baseurl)
 return : list of the links
*/
function gl($t,$b) {
  $r=array();
  $c=count($t->child);
  for ($i=0;$i<$c;$i++) {
    $e=&$t->child[$i];
    if ($e->name=="a") { // a link
      $h=$e->attribute["href"]; // url
      if (substr($h,0,4)!="http") { // prevent external links
        $r[]=sp(substr($h,0,1)=="/"?$h:$b.$h);
      }
    } else { // not a link, search recursively inside
      $r=array_merge($r,gl($e,$b));
    }
  }
  return $r;
}
/**
 simplify_path($path)
 return : simplified path
*/
function sp($p) {
  while ($o!=$p) {
    $o=$p;
    $p=str_replace(array("//","/./"),"/",$p);
    $p=preg_replace("/\/[^\/]+\/..\//","/",$p);
  }
  return $p;
}

?>
Limitation : does not detect javascript-generated links. Check about set_time_limit(0) if you have a lot of webpages.

Paul Cook (11-Apr-2006 04:25)

To get libtidy and PHP 5.0.5 compiled on OS X Tiger this is what I needed to do:

1) download and upack the tidy source.
2) cd tidy-source-dir
3) >> /bin/sh build/gnuauto/setup.sh
4) then you can configure/make/make install as normal

PHP build generates errors because of tidy so I needed to edit the platform.h file like this (use your favorite command line editor):

5) >> sudo emacs /usr/local/include/platform.h
6) comment out line 508 which was causing the 'duplicate "unsigned" ' error in the PHP build.

7) configure/make/make install PHP as normal using --with-tidy=/usr/local

Restart apache and everything works now.  HTH someone.

patatraboum at nospam dot fr (25-Feb-2006 06:13)

<?php
//
//The tidy tree of your favorite !
//For PHP 5 (CGI)
//Thanks to john@php.net
//
$file="http://www.php.net";
//
$cns=get_defined_constants(true);
$tidyCns=array("tags"=>array(),"types"=>array());
foreach($cns["tidy"] as $cKey=>$cVal){
    if($cPos=strpos($cKey,$cStr="TAG")) $tidyCns["tags"][$cVal]="$cStr : ".substr($cKey,$cPos+strlen($cStr)+1);
    elseif($cPos=strpos($cKey,$cStr="TYPE")) $tidyCns["types"][$cVal]="$cStr : ".substr($cKey,$cPos+strlen($cStr)+1);
}
$tidyNext=array();
//
echo "<html><head><meta http-equiv='Content-Type' content='text/html; charset=windows-1252'><title>Tidy Tree :: $file</title></head>";
echo "<body><pre>";
//
tidyTree(tidy_get_root(tidy_parse_file($file)),0);
//
function tidyTree($tidy,$level){
    global $tidyCns,$tidyNext;
    $tidyTab=array();
    $tidyKeys=array("type","value","id","attribute");
    foreach($tidy as $pKey=>$pVal){
        if(in_array($pKey,$tidyKeys)) $tidyTab[array_search($pKey,$tidyKeys)]=$pVal;
    }
    ksort($tidyTab);
    foreach($tidyTab as $pKey=>$pVal){
        switch($pKey){
            case 0 :
                if($pVal==4) $value=true; else $value=false;
                echo indent(true,$level).$tidyCns["types"][$pVal]."\n"; break;
            case 1 :
                if($value){
                    echo indent(false,$level)."VALEUR : ".str_replace("\n","\n".indent(false,$level),$pVal)."\n";
                }
                break;
            case 2 :
                echo indent(false,$level).$tidyCns["tags"][$pVal]."\n"; break;
            case 3 :
                if($pVal!=NULL){
                    echo indent(false,$level)."ATTRIBUTS : ";
                    foreach ($pVal as $aKey=>$aVal) echo "$aKey=$aVal "; echo "\n";
                }
        }
    }
    if($tidy->hasChildren()){
        $level++; $i=0;
        $tidyNext[$level]=true;
        echo indent(false,$level)."\n";
        foreach($tidy->child as $child){
            $i++;
            if($i==count($tidy->child)) $tidyNext[$level]=false;
            tidyTree($child,$level);
        }
    }
    else echo indent(false,$level)."\n";
}
//
function indent($tidyType,$level){
    global $tidyNext;
    $indent="";
    for($i=1;$i<=$level;$i++){
        if($i<$level||!$tidyType){
            if($tidyNext[$i]) $str="|  "; else $str="   ";
        }
        else $str="+--";
        $indent=$indent.$str;
    }
    return $indent;
}
//
echo "</pre></body></html>";
//
?>

tonygambone at gNOSPAMmail dot com (06-Feb-2006 04:03)

Using PHP 5.1.2 on Win32/IIS, I noticed that even with "output-xhtml: yes," tidy was adding the deprecated name attribute to form tags (using the value of the id attribute).  Grabbing the latest dll from the snaps link at the top of the page fixed this.

tom at expresshosting dot net (23-Aug-2005 06:50)

It should be noted that the examples on this page apply ONLY to PHP5. None of the functions in the manual apply to PHP4. The names are the same but arguments are different on some of them (tidy_parse_string).

If you wish to use tidy in PHP 4.3.x you can use the following example instead:

<?php

$tidyhtml = ob_get_contents();
if( function_exists( 'tidy_parse_string' ) ) {
        tidy_set_encoding('iso-8859-1');
        tidy_parse_string($tidyhtml);
        tidy_setopt('output-xhtml', TRUE);
        tidy_setopt('indent', TRUE);
        tidy_setopt('indent-spaces', 2);
        tidy_setopt('wrap', 200);
        tidy_clean_repair();
        $tidyhtml = tidy_get_output();
}
ob_end_clean();

echo $tidyhtml;

?>

Hope that helps somebody.

mohan at asix dot com dot my (11-Feb-2005 08:23)

To those who need to install libtidy on mac os x , here is a guide that worked for me :

If you're on Mac OS X, you'll need to tell the Makefile that you use
ranlib:

    $ export set RANLIB=ranlib

Change to the directory with the Makefile in it, and run make.
This example uses the GNU make Makefile.

    $ cd tidy/build/gmake/

    $ make
    if [ ! -d ./obj ]; then mkdir ./obj; fi
    gcc -o obj/access.o ...
    ... etc etc etc ...

Install the libs, headers and the tidy executable:

    $ sudo make install

If you're on Mac OS X, you'll have to run ranlib again on the installed
lib:

    $ sudo ranlib /usr/local/lib/libtidy.a

Jon Dowland (bugs at alcopop dot org) (01-Feb-2005 11:40)

Rough installation instructions for debian/testing:

Use debian's apt package manager to install the required development packages

$ apt-get install php4-dev php4-pear libtidy-dev

Then use pear to install tidy

$ pear install tidy

Note: I did /not/ have success installing the tarball locally. Only using this method was the .so put in the correct place.

I also had to add an entry to the php.ini

$ echo extension=tidy.so >> /etc/php4/apache/php.ini
$ apachectl restart

...and you're done.

doodleelephant (16-Nov-2004 01:34)

I'm installing PHP 5.0.2 on Redhat Linux (I forget the version. Enterprise WS 3 I think) I had troubles installing the libtidy. It consistently complained that it could not find 'libtidy'. I finally got a clue into how to install it (in build/gnuauto/readme.txt). This is how I finally got it to install (after lots of trial and error):

First, don't get the binary distribution of of tidy.sf.net. It's not what you want. You need the source distribution.

Command by command this is what I did:
=======

wget http://tidy.sourceforge.net/src/tidy_src.tgz
tar -xzf tidy_src.tgz
cd tidy
/bin/sh build/gnuauto/setup.sh
./configure --prefix=/usr
make
make install
cd [php source directory]
./configure --with-tidy=/usr --[other extensions]
make
make install

=======
Tada. Finally it doesn't complain when I configure PHP about the installation. The info I needed was stuck in that build/gnuauto/readme.txt file in the tidy directory.

Took me a while. Hope my trials can help others save time.

Doodleelephant

bill dot mccuistion at qbopen dot com (30-Oct-2004 06:53)

Installing tidy on Fedora Core 2 required three libraries:

tidy...
tidy-devel...
libtidy...

All of which I found at http://rpm.pbone.net

Then, finally, could "./configure --with-tidy"

Hope this helps someone out.  This was "REALLY" hard (for me) to figure out as no where else was clearly documented.