mirror/php-htmlpurifier
1
0
Fork 0
mirror of https://github.com/ezyang/htmlpurifier.git synced 2025-08-05 05:37:49 +02:00
Code Issues Releases Wiki Activity

Dramatically rewrite null host URI handling.

Browse Source 
Basically, browsers don't parse what should be valid URIs correctly, so
we have to go through some backbends to accomodate them.  Specifically,
for browseable URIs, the following URIs have unintended behavior:

    - ///example.com
    - http:/example.com
    - http:///example.com

Furthermore, if the path begins with //, modifying these URLs must
be done with care, as if you remove the host-name component, the
parse tree changes.

I've modified the engine to follow correct URI semantics as much
as possible while outputting browser compatible code, and invalidate
the URI in cases where we can't deal.  There has been a refactoring
of URIScheme so that this important check is always performed,
introducing a new member variable allow_empty_host which is true
on data, file, mailto and news schemes.

This also fixes bypass bugs on URI.Munge.

Signed-off-by: Edward Z. Yang <ezyang@mit.edu>
This commit is contained in:
Edward Z. Yang
2011-01-25 18:56:46 +00:00
parent a32d5b52e1
commit e76f4b45d0
15 changed files with 216 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files

93
library/HTMLPurifier/URI.php
View File

@@ -67,14 +67,6 @@ class HTMLPurifier_URI
$chars_gen_delims = ':/?#[]@';
$chars_pchar = $chars_sub_delims . ':@';
// validate scheme (MUST BE FIRST!)
if (!is_null($this->scheme) && is_null($this->host)) {
$def = $config->getDefinition('URI');
if ($def->defaultScheme === $this->scheme) {
$this->scheme = null;
}
}
// validate host
if (!is_null($this->host)) {
$host_def = new HTMLPurifier_AttrDef_URI_Host();
@@ -82,6 +74,21 @@ class HTMLPurifier_URI
if ($this->host === false) $this->host = null;
}
// validate scheme
// NOTE: It's not appropriate to check whether or not this
// scheme is in our registry, since a URIFilter may convert a
// URI that we don't allow into one we do. So instead, we just
// check if the scheme can be dropped because there is no host
// and it is our default scheme.
if (!is_null($this->scheme) && is_null($this->host) || $this->host === '') {
// support for relative paths is pretty abysmal when the
// scheme is present, so axe it when possible
$def = $config->getDefinition('URI');
if ($def->defaultScheme === $this->scheme) {
$this->scheme = null;
}
}
// validate username
if (!is_null($this->userinfo)) {
$encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . ':');
@@ -96,32 +103,48 @@ class HTMLPurifier_URI
// validate path
$path_parts = array();
$segments_encoder = new HTMLPurifier_PercentEncoder($chars_pchar . '/');
if (!is_null($this->host)) {
if (!is_null($this->host)) { // this catches $this->host === ''
// path-abempty (hier and relative)
// http://www.example.com/my/path
// //www.example.com/my/path (looks odd, but works, and
// recognized by most browsers)
// (this set is valid or invalid on a scheme by scheme
// basis, so we'll deal with it later)
// file:///my/path
// ///my/path
$this->path = $segments_encoder->encode($this->path);
} elseif ($this->path !== '' && $this->path[0] === '/') {
// path-absolute (hier and relative)
if (strlen($this->path) >= 2 && $this->path[1] === '/') {
// This shouldn't ever happen!
$this->path = '';
} else {
} elseif ($this->path !== '') {
if ($this->path[0] === '/') {
// path-absolute (hier and relative)
// http:/my/path
// /my/path
if (strlen($this->path) >= 2 && $this->path[1] === '/') {
// This could happen if both the host gets stripped
// out
// http://my/path
// //my/path
$this->path = '';
} else {
$this->path = $segments_encoder->encode($this->path);
}
} elseif (!is_null($this->scheme)) {
// path-rootless (hier)
// http:my/path
// Short circuit evaluation means we don't need to check nz
$this->path = $segments_encoder->encode($this->path);
}
} elseif (!is_null($this->scheme) && $this->path !== '') {
// path-rootless (hier)
// Short circuit evaluation means we don't need to check nz
$this->path = $segments_encoder->encode($this->path);
} elseif (is_null($this->scheme) && $this->path !== '') {
// path-noscheme (relative)
// (once again, not checking nz)
$segment_nc_encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . '@');
$c = strpos($this->path, '/');
if ($c !== false) {
$this->path =
$segment_nc_encoder->encode(substr($this->path, 0, $c)) .
$segments_encoder->encode(substr($this->path, $c));
} else {
$this->path = $segment_nc_encoder->encode($this->path);
// path-noscheme (relative)
// my/path
// (once again, not checking nz)
$segment_nc_encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . '@');
$c = strpos($this->path, '/');
if ($c !== false) {
$this->path =
$segment_nc_encoder->encode(substr($this->path, 0, $c)) .
$segments_encoder->encode(substr($this->path, $c));
} else {
$this->path = $segment_nc_encoder->encode($this->path);
}
}
} else {
// path-empty (hier and relative)
@@ -150,6 +173,9 @@ class HTMLPurifier_URI
public function toString() {
// reconstruct authority
$authority = null;
// there is a rendering difference between a null authority
// (http:foo-bar) and an empty string authority
// (http:///foo-bar).
if (!is_null($this->host)) {
$authority = '';
if(!is_null($this->userinfo)) $authority .= $this->userinfo . '@';
@@ -157,7 +183,12 @@ class HTMLPurifier_URI
if(!is_null($this->port)) $authority .= ':' . $this->port;
}
// reconstruct the result
// Reconstruct the result
// One might wonder about parsing quirks from browsers after
// this reconstruction. Unfortunately, parsing behaviro depends
// on what *scheme* was employed (file:///foo is handled *very*
// differently than http:///foo), so unfortunately we have to
// defer to the schemes to do the right thing.
$result = '';
if (!is_null($this->scheme)) $result .= $this->scheme . ':';
if (!is_null($authority)) $result .= '//' . $authority;