Re: Return type declaration and &return.falseforfailure;

We should be adding <void> if there is no return value. If a function can return null, then that's a valid return type. The issue here is that many of the PHP functions return way too many types. The inability of the function/method signature correctly express itself is a pain. You read the signature. It says "I return stuff". I then have to read the rest of the documentation to find that it returns something useful or false. <type>int</type><type>false</type> (as true and false are now types), is very explicit, understandable, parseable and can be used by other systems in a concise way. And don't get me started on the fact the structure of the signature is not PHP compliant! mixed function(type param) is not correct. function(type param): mixed is what it should be. For a newbie reading the PHP dox for the first time and seeing the wrong structure, trying to use it, failing, ... On 11 February 2018 at 22:19, Christoph M. Becker <cmbecker69@gmx.de> wrote: > On 11.02.2018 at 22:31, Richard Quadling wrote: > > > DocBook 5.2 now supports multiple types, so, really, the form > SHOULD/COULD > > be > > > > <type>int</type><type>bool</type> > > > > In my opinion, <type>mixed</type> is lazy. We now have scalar type > hinting > > (though not union type hints which is a shame as that matches the > language > > itself quite a lot). > > > > If the documentation can be accurate, then it should be. > > Well, then we'd have to add NULL as return type to almost every > function. IMHO, not really helpful. > > > On 5 December 2017 at 10:15, Andrew Gromov <andrewgrom@rambler.ru> > wrote: > > > >> I see many changes from anonymous users like this: > >> > >> - <type>int</type><methodname>preg_match</methodname> > >> + <type>int|bool</type><methodname>preg_match</methodname> > >> > >> > >> or this: > >> > >> - <type>int</type><methodname>preg_match_all</methodname> > >> + <type>mixed</type><methodname>preg_match_all</methodname> > >> > >> > >> Formally them correct, but by spirit them is bad (imho). > >> > >> Accept or reject them? > > -- > Christoph M. Becker > -- Richard Quadling