Uri\Rfc3986\Uri::toString() doesn't drop an empty port

[kamil-tekiela]

Description

The following code:

<?php

require 'vendor/autoload.php';

$uri = new Uri\Rfc3986\Uri('https://example.com:');
var_dump($uri->getPort(), $uri->toRawString(), $uri->toString());

Resulted in this output:

NULL
string(20) "https://example.com:"
string(20) "https://example.com:"

But I expected this output instead:

NULL
string(20) "https://example.com:"
string(19) "https://example.com"

PHP Version

8.5.4

Operating System

No response

[iluuu1994]

/cc @kocsismate @TimWolla

[kamil-tekiela]

FYI Reposted from thephpleague/uri-src#215 (comment)

I have opened several bug reports for polyfill, but you may want to check whether some of them aren't PHP bugs. https://github.com/thephpleague/uri-src/issues?q=is%3Aissue%20author%3Akamil-tekiela

[kocsismate]

The normalization in question is mentioned in RFC 3986 (with a SHOULD though):

Likewise, an explicit ":port", for which the port is empty or the default for the scheme, is equivalent to one where the port and its ":" delimiter are elided and thus should be removed by scheme-based normalization.

Currently, uriparser doesn't implement this normalization yet:

php-src/ext/uri/uriparser/src/UriNormalize.c

Line 688 in bb73942

/* Is there a port even? */

I'll try to implement it sometime soon (although uriparser reviews take a lot of time nowadays).

[kocsismate]

P.S. this normalization option is mentioned wrt scheme-based normalization (section 6.2.1 in the RFC) which - by itself - is out of scope for a library implementing the generic URI syntax (uriparser/uriparser#173).

However, I'd argue that normalizing "https://example.com:" to "https://example.com" falls in the scheme-based normalization category, so maybe it's still worth a try to implement in uriparser, especially because the RFC references this case in another section:

URI producers and normalizers should omit the ":" delimiter that separates host from port if the port component is empty. Some schemes do not allow the userinfo and/or port subcomponents.

[TimWolla]

3.2.3 states:

URI producers AND normalizers should omit the port component and its ":" delimiter if port is empty or if its value would be the same as that of THE scheme's default.

This I would expect normalization to happen in uriparser.

[TimWolla]

I proposed uriparser/uriparser#342