Re: Should I document session_get_cookie_params => setcookie?

On Tue, 2 Jun 2020 at 09:12, Christoph M. Becker <cmbecker69@gmx.de> wrote: > On 02.06.2020 at 09:35, Peter Cowburn wrote: > > > On Tue, 2 Jun 2020 at 02:58, Karoly Negyesi <karoly@negyesi.net> wrote: > > > >> Hello, > >> > >> Today I wanted to check from JavaScript whether a session was started > but > >> the system sets the session cookie to httponly so I can't do that so I > >> wanted to do setcookie('has_session', 1, session_get_cookie_params()). > This > >> of course (?) doesn't work. Should I add a code snippet on how to make > it > >> work to session_get_cookie_params and / or setcookie or should I just > >> submit to Stackoverflow as a self answer? > >> > > > > Probably the latter. > > > > > >> I already added > >> https://marc.info/?l=php-doc-cvs&m=159105012229429&w=2 > >> because that's generic enough. Speaking of, should I amend the return > >> values of setcookie to include this null or just this commit is enough? > >> > > > > I'm not sure where the null is coming from. The function should be > > returning false for the case you've described, right? > > Actually, setcookie() is not supposed to fail if there are spurious > elements in the options array; see e.g. > <https://3v4l.org/h4dWr>. > E_NOT_ENOUGH_COFFEE -- yes, it should emit the warning(s) but not fail *due to the extra options*. > > -- > Christoph M. Becker >