UrlLinker is a PHP module for converting plain text snippets to HTML, and any web addresses in the text into HTML hyperlinks.
Usage:
print(htmlEscapeAndLinkUrls($text));
For a longer example, see UrlLinker-example.php.
UrlLinker assumes plain text input, and returns HTML. If your input is already HTML, but it contains URLs that have not been marked up, UrlLinker can handle that as well:
print(linkUrlsInTrustedHtml($html));
Warning: The latter function must only be used on trusted input, as rendering
HTML provided by a malicious user can lead to system compromise through
cross-site scripting. The htmlEscapeAndLinkUrls
function, on the other
hand, can safely be used on untrusted input. (You can remove existing tags from
untrusted input via PHP's strip_tags function.)
Note: This repository is no longer being actively maintained. If you want to use UrlLinker with Composer, consider this fork.
- Web addresses
- Recognized URL schemes: "http" and "https"
- The
http://
prefix is optional. - Support for additional schemes, e.g. "ftp", can easily be added by
tweaking
$rexScheme
. - The scheme must be written in lower case. This requirement can be lifted
by adding an
i
(thePCRE_CASELESS
modifier) to$rexUrlLinker
.
- The
- Hosts may be specified using domain names or IPv4 addresses.
- IPv6 addresses are not supported.
- Port numbers are allowed.
- Internationalized Resource Identifiers (IRIs) are allowed. Note that the job of converting IRIs to URIs is left to the user's browser.
- To reduce false positives, UrlLinker verifies that the top-level domain is
on the official IANA list of valid TLDs.
- UrlLinker is updated from time to time as the TLD list is expanded.
- In the future, this approach may collapse under ICANN's ill-advised new policy of selling arbitrary TLDs for large amounts of cash, but for now it is an effective method of rejecting invalid URLs.
- Internationalized top-level domain names must be written in Punycode in order to be recognized.
- If you need to support unqualified domain names, such as
localhost
, you may disable the TLD check by 1) replacing+
with*
in the$rexDomain
value and 2) replacing theif
statement line beneath the "Check that the TLD is valid" comment withif (true)
. This is obviously a quick-and-dirty hack, and may cause false positives.
- Recognized URL schemes: "http" and "https"
- Email addresses
- Supports the full range of commonly used address formats, including "plus addresses" (as popularized by Gmail).
- Does not recognized the more obscure address variants that are allowed by the RFCs but never seen in practice.
- Simplistic spam protection: The at-sign is converted to a HTML entity, foiling naive email address harvesters.
- Addresses are recognized correctly in normal sentence contexts. For instance, in "Visit stackoverflow.com.", the final period is not part of the URL.
- User input is properly sanitized to prevent cross-site scripting (XSS),
and ampersands in URLs are correctly escaped as
&
(this does not apply to thelinkUrlsInTrustedHtml
function, which assumes its input to be valid HTML).
A Stackoverflow.com question prompted me to consider the difficulty of this task. Initially, it seemed easy, but like an itch you just have to scratch, I kept coming back to it, to fix just one more little thing.
Feel free to upvote my answer if you find this code useful.
There's also a C# implementation by Antoine Sottiau.
To the extent possible under law, the author has waived all copyright and related or neighboring rights to UrlLinker.
For more information see: http://creativecommons.org/publicdomain/zero/1.0/