On 24/04/2019 21:32, Björn Larsson wrote:
With that experience in mind I wonder how different libraries will fare, given this > change? One should also have in mind that there has been a discussion on this list about extending the support cycle for PHP 7.4 like for 5.6, but to some degree it was rejected which doesn't help migration efforts.
I'm sure Nikita will pop up sometime in the next few days and put his recently downloaded 1000+ packages through his parser and give us some figures on how many use short open tags, if he hasn't already done so.
I would naturally expect it to be extremely small, after all it makes no sense to have a public package which requires a language feature to be enabled in the INI and which can't be relied upon.
If there's a problem, it will almost certainly be with internal code, which naturally there's a few million times more of, but that seems very much in-scope for individual developers to fix if they want to upgrade to PHP 8.
My main worry is, and remains, that it's being converted to just a "not parser significant" in PHP 8. The potential for unexpected data / code leaks is *significant*.
Until someone can convince me otherwise, I will continue to strongly believe that <? should lead to a compile-time fatal error for the foreseeable future (years).
I'm half tempted to RFC it so it can go to a vote. -- Mark Randall -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
