On Fri, 30 Aug 2019 at 11:45, Zeev Suraski <z...@zend.com> wrote: > > > On 30 Aug 2019, at 12:33, Nikita Popov <nikita....@gmail.com> wrote: > > > > Hi internals, > > > > Relating to the recent discussions on undefined variables & co. One thing > > that is particularly annoying about the undefined variable case is that > our > > default error_reporting level (without a php.ini) does not include > E_NOTICE. > > > > Thankfully distros do tend to have more reasonable defaults, but if you > > spend as much time with custom PHP builds as I do, not seeing *anything* > > for undefined variables is a pretty big annoyance. > > > > Does anyone see an issue with making error_reporting=E_ALL the default in > > PHP 8? It can of course still be manually downgraded via php.ini and > > php.ini-production will retain the existing recommendation that excludes > > E_DEPRECATED/E_STRICT. > > > > PR: https://github.com/php/php-src/pull/4659 > > By the way - I'd consider going a bit further than that, and re-sync our > internal defaults with the values of php.ini-development. In the past - > when we had -dist and -recommended - the intended behavior was that having > php.ini-dist as your php.ini would be identical to not having a php.ini > file at all. I think it can be useful if we do the same for > php.ini-development (by changing the internal defaults to correspond to it > of course, not the other way around) instead of having to agree > individually on each change. For the future, agreeing on a change in > php.ini-development would directly imply changing the internal defaults. > > Zeev > -- > PHP Internals - PHP Runtime Development Mailing List > To unsubscribe, visit: http://www.php.net/unsub.php
I'm not sure we can totally execute on that otherwise we would need to change the default of the short_tag directive, which from my understanding was one of the pain points of the Short Tag RFC. Other than that I wholeheartedly agree that syncing the dev one with the defaults would be a good idea but then I'm not totally sure what it's point *would* be. Best regards George P. Banyard
