Generic.Formatting.MultipleStatementAlignment.NotSame doesn't parse short list properly after upgrading to 3.2.3

[lcobucci]

After updating to v3.2.3 this sniff is complaining about incorrect spacing regarding this code:

<?php
declare(strict_types=1);

$payload = $message->getPayload();

[$userInput, $errors] = $this->normalizeUserInput($payload);

Output:

$ phpcs aa.php 
E 1 / 1 (100%)



FILE: aa.php
------------------------------------------------------------------------------------------------------------
FOUND 1 ERROR AFFECTING 1 LINE
------------------------------------------------------------------------------------------------------------
 4 | ERROR | [x] Equals sign not aligned with surrounding assignments; expected 14 spaces but found 1 space
------------------------------------------------------------------------------------------------------------
PHPCBF CAN FIX THE 1 MARKED SNIFF VIOLATIONS AUTOMATICALLY
------------------------------------------------------------------------------------------------------------

And it doesn't complain anymore if we use list() instead of [], like:

<?php
declare(strict_types=1);

$payload = $message->getPayload();

($userInput, $errors) = $this->normalizeUserInput($payload);

Output:

$ phpcs aa.php 
. 1 / 1 (100%)


Time: 131ms; Memory: 14Mb

[lcobucci]

Maybe these are related: #1911 #1984

[jrfnl]

@lcobucci Hi Luís, #1870, #1848, #1924 are all related to this, but I'm surprised you are seeing the message as your example code looks to use blank lines between the lines and that would "break" a multi-statement block.
Could it be that the code you are testing on does not have the blank lines ?

So, this should trigger the notice:

<?php
declare(strict_types=1);

$payload = $message->getPayload();
[$userInput, $errors] = $this->normalizeUserInput($payload);

This should not:

<?php
declare(strict_types=1);

$payload = $message->getPayload();

[$userInput, $errors] = $this->normalizeUserInput($payload);

[jrfnl]

Just ran some tests which confirm what I said above.
Also: whether you use the short list syntax or the list() construct - correctly - has no influence on the results according to my test.

[lcobucci]

@jrfnl the code is exactly the way I've posted (with the line in between), which indeed surprised me. It can sure be an influence of another sniff but I'm afraid it wouldn't have that output while only analysing the code.

[jrfnl]

@lcobucci Hmm... the blank line in between should ensure that this sniff isn't triggered.
Hang on... ok.. so I've tested this now on 3.2.3 and can confirm your findings. Looks like this has already been fixed in master via #1924 which fixed #1922.

[lcobucci]

@jrfnl let me try it with master

[lcobucci]

@jrfnl it's indeed fixed in master, we can close it as a duplicate then. Thanks for your help 👍

[jrfnl]

You're welcome.