Thank you for taking the time to write to us, but this is not
a bug. Please double-check the documentation available at
http://www.php.net/manual/ and the instructions on how to report
a bug at http://bugs.php.net/how-to-report.php

This is not a bug, for an extensive explanation see Bug #80114

https://bugs.php.net/bug.php?id=80114

@sjon: But apparently the Windows builds aren't showing this new behavior.

Since this was marked as "Not a bug", i'd like to point out the inconsistency across the several platforms, as well as the current behaviour in userland. The used test stems from current PSR-7 implementations where it is expected to be an invalid URI. See also: https://github.com/guzzle/psr7/blob/7858757f390bbe4b3d81762a97d6e6e786bb70ad/tests/UriTest.php#L137

$ x64\Debug\php -v
PHP 8.0.0-dev (cli) (built: Oct 16 2020 18:46:05) ( NTS Visual C++ 2019 x64 DEBUG )
Copyright (c) The PHP Group
Zend Engine v4.0.0-dev, Copyright (c) Zend Technologies

C:\php-sdk\phpdev\vs16\x64\php-src-8.0
$ x64\Debug\php -r "var_dump(parse_url('//portcannotbezero.com:0'));"
array(1) {
  ["host"]=>
  string(20) "portcannotbezero.com"
}

IOW, I cannot reproduce.

@cmb: in fact, you could reproduce what the Windows build of the GH actions runner didn't. Aside, if this is now expected behaviour, why is the "port" key with a value 0 missing in the array - according to the previously linked bug, the port number 0 should be treated as valid?

Firstly, this is not particularly related to PHP 8.0,
but rather affects PHP-7.3+.

> […] in fact, you could reproduce what the Windows build of the
> GH actions runner didn't.

The GH actions runner uses PHP 7.4.11 (which does not have that
bugfix) and some PHP 8.0.0-dev version (which presumably does not
have the bugfix either).

> Aside, if this is now expected behaviour, why is the "port" key
> with a value 0 missing in the array

From the commit message:

| URIs with a 0 port are generally valid, so `parse_url()` should
| recognize such URIs, but still report the port as missing.

Apparently, we should document that behavior.

> Apparently, we should document that behavior.

Yes, that would be greatly appreciated, also a mention in UPGRADING would be helpful.

Thanks for clearing this up and thank you for your time! I think this can be closed then. Cheers!

| URIs with a 0 port are generally valid, so `parse_url()` should
| recognize such URIs, but still report the port as missing.

I do wonder whether that's the right behavior. It would make more sense to include the zero port in the output to me.

> It would make more sense to include the zero port in the output
> to me.

After further consideration, I have to agree.