Re: json_encode() bug

On Tue, 5 Feb 2008, Richard Lynch wrote: > On Sun, February 3, 2008 7:51 pm, Stanislav Malyshev wrote: > >> Like I mentioned before (I think), it should not return an empty > >> string of course because programmatically it's not possible to > >> check for this. As most of our functions return false in those > >> cases, so should this function. > > > > AFAIR false is not valid JSON, so it would break a lot of code. > > Also, I am not sure we should change json_encode to return false on > > whole structure if one of the fields contains invalid utf-8. [snip] > Put partially decoded value somewhere in a variable to be accessed, so > that the programmer can make an informed decision about what to do. We're talking about encoding here though... regards, Derick -- Derick Rethans http://derickrethans.nl | http://ezcomponents.org | http://xdebug.org