On Sun, Jul 20, 2008 at 2:02 PM, Lukas Kahwe Smith <[EMAIL PROTECTED]> wrote:
> > On 20.07.2008, at 07:45, Nathan Nobbe wrote: > > On Sat, Jul 19, 2008 at 4:55 AM, Lars Strojny <[EMAIL PROTECTED]> wrote: >> >> Hi everbody, >>> >>> regarding my mail from yesterday, I've also created an RFC for the new >>> error level. >>> >>> http://wiki.php.net/rfc/e-user-deprecated-warning >>> >> >> >> i definitely like the E_USER_DEPRECATED :D im curious though, about >> E_DEPRECATED. is this for deprecated functions at the C level, or just >> the >> php api? the reason i ask is because deprecation notices are already >> issued >> via the E_STRICT level, for example, when using is_a() today, w/ E_STRICT >> enabled, the following is generated >> >> Strict standards: is_a(): Deprecated. Please use the instanceof operator >> in >> ... >> >> would, perhaps, the deprecation level for is_a() move to E_DEPRECATED (i >> noticed its still there in 5.3)? >> > > > deprecation of is_a() was a mistake that has been reverted. thats interesting, may i ask when that would be a better choice that instanceof ? and anyway, there are other examples as well, for example mktime(), PHP Strict standards: mktime(): You should be using the time() function instead in ... would this be another candidate to port to E_DEPRECATED ? so i imagine there will be some work to convert those E_STRICT errors regarding deprecation to the new E_DEPRECATED level. maybe that should go along w/ the patch. as a matter of consistency, i think it would be best to move all E_STRICT errors regarding deprecation to E_DEPRECATED in a single patch, rather than a piecemeal fashion. -nathan
