Re: Re: com php-src: Fix bug #67064 => Countable interface inconsistency

On Fri, Jul 25, 2014 at 11:54 AM, Matteo Beccati <php@beccati.com> wrote: > On 23/07/2014 13:01, Dan Ackroyd wrote: > > Regarding the change in behaviour, I think the patch > > > > https://bugs.php.net/patch-display.php?bug=67064&patch=bug67064-BC&revision=1402667581 > > should definitely be applied. > > > > Currently the engine is doing magic stuff by modifying the parameters > > that are passed to the function. i.e. you call count() with no > > parameters and it arrives at the function with parameters. \o/ > > > > Even if that wasn't a BC break, having magic behaviour in a language > > is very, very bad. > > I agree, that's why I wrote the tiny patch in the first place ;) > > IMHO, the functionality as it is now breaks the principle of least > surprise. > > Bob, what's your opinion on this? You also had a conversation with Remi on irc, but I don't remember what was the decision if any. The PR from Remi for imagick to support 5.6 doesn't seem so bad (as in: doesn't require too big of a change from the extension's part), but we also have a couple of classes implementing the countable interface in core: http://lxr.php.net/search?q=spl_ce_Countable&defs=&refs=&path=&hist=&project=PECL Whats the current plan? Removing the internal BC break, or keeping it and accomodating the affected parts in php-src? -- Ferenc Kovács @Tyr43l - http://tyrael.hu