1. This project loads: ./vendor/laminas/laminas-coding-standard/ruleset.xml: https://github.com/laminas/laminas-mail/blob/2.13.0/phpcs.xml#L3
2. Which loads PSR2 <rule ref="PSR2"/>: https://github.com/laminas/laminas-coding-standard/blob/1.0.0/ruleset.xml#L10
3. Which defines <property name="lineLimit" value="120"/>, <property name="absoluteLineLimit" value="0"/>: https://github.com/squizlabs/PHP_CodeSniffer/blob/2.9.2/CodeSniffer/Standards/PSR2/ruleset.xml#L30-L36
4. so, according to PSR-2 soft limit should be 120 and no hardlimit: https://www.php-fig.org/psr/psr-2/#1-overview

and this is exactly how phpcs behaves:

[~/scm/php/zend-mail (override-cs-length)★] ➔ c cs-check
> phpcs
............................................................  60 / 152 (39%)
.............................................W.............. 120 / 152 (79%)
................................


FILE: /Users/glen/scm/php/zend-mail/test/Transport/SendmailTest.php
----------------------------------------------------------------------
FOUND 0 ERRORS AND 2 WARNINGS AFFECTING 2 LINES
----------------------------------------------------------------------
  97 | WARNING | Line exceeds 120 characters; contains 130 characters
 283 | WARNING | Line exceeds 120 characters; contains 122 characters
----------------------------------------------------------------------

Time: 3.61 secs; Memory: 24Mb

Script phpcs handling the cs-check event returned with error code 1

but the problem is that WARNING is treated as fatal, perhaps turn that off? I currently can't find how is that behaviour controlled.

This is the wrong direction because:

1. It opens the door for more errors of this kind in the future and for entire code instead of fixing them or to use existing features.
2. At the moment the old version of the coding standard (1.0) is still in use here, so modifications are not expedient.

@froschdesign such request came from @Ocramius:

• [2.14]: QA: Add specific type hints for $params #126 (comment)

perhaps you can jump in there and propose (show), how to set the exclude, as indicated there, previous attempts have failed (nothing works). the PR itself was extracted from larger PR as was unable to solve the cs exclude rule there.