Re: Re: Revert session_serializer_name(), session_gc()

Hi, On Sat, Mar 15, 2014 at 6:36 AM, Yasuo Ohgaki <yohgaki@ohgaki.net> wrote: > Hi all, > > On Sat, Mar 15, 2014 at 1:15 PM, Yasuo Ohgaki <yohgaki@ohgaki.net> wrote: >>> >>> >>> Now back to the main topic: >>> >>> Please exclude session_serializer_name(), session_gc(), >>> session_reset(), session_abort() and the "session write short-circuit" >>> from the 5.6 branch. >> >> >> I removed session_serializer_name() and session_gc() >> (Although session_gc() is mandatory API, IMO) >> I like the idea removing INI modifying function in the future release. >> >> I don't understand reason why you insist removal of session_reset() >> and session_abort(). They are just missing API for session module, >> like session_gc(). >> >> There should be API (i.e. function/method or parameters) for distinct >> operations that user may use. >> >> I may agree if you could provide the reason why there should not be >> these APIs. session_abort(), session_reset() are just two functions that somebody asked about in 2002 via bugs.php.net and nobody responded (11 years without a single comment) right up until you just assumed it's a good idea and commited it. If it was "missing API", surely at least 1 user per year would've requested it. ;) I understand that you see them as small, trivial additions, but being a part of sessions automatically makes them very important and as such, they should be evaluated collectively, not by a single person. > I thought it might be better to explain what new functions do. > > session_gc() executes GC without tweaking INIs. > session_abort() aborts session without writing $_SESSION. There is no way > achieve w/o it. > session_reset() re-reads session and re-initializes $_SESSION. There is no > way achieve w/o it. (It could be done with session_abort(), though) > "write short circuit" omits "write" when $_SESSION hasn't change. > There is > no point calling write API and writing to storage for the same data. "write short circuit" as I understand it, is an exact copy of the 'lazy_write' option. This will be addressed in the previously mentioned RFC that I'll post later today. Cheers, Andrey.