Re: [RFC] Revert/extend/postpone original RFC about read_only, lazy_write sessions

Hi Stas, Thank you for the comment on github. On Tue, Mar 18, 2014 at 2:45 PM, Stas Malyshev <smalyshev@sugarcrm.com>wrote: > > I agree. > > If nobody objects, I'll remove the option. Please send mail ASAP! > > I think defaults should not change in minor version, unless there's a > very good reason for it, and here it's not a very good reason. Session > module is very sensitive, and doing changes in it should be very > conservative to not break people's workflows. While new behavior serves > an important use case, it may not be everybody's use case, especially > with all handlers. Thus, I would rather have it an option - exactly as > voted. When we move into PHP 6, where certain major changes are > expected, we can change the defaults. But I think introducing such > change in a minor version is unnecessary risk. > And also it subverts RFC process as it's not what the vote was for, and > it's a substantial change - default behavior change (means, everybody > affected) instead of option (means, only people using the option affected No problem. How about making it a INI option? It makes session_start() option handling code a little simpler. It's not mandatory, though. > > When is the beta due? I made PR so that RMs can merge it. > > I'd recommend waiting for it to be thoroughly reviewed before merging. I > myself will try to take a look in coming days, as my schedule allows, > but it may take time to polish some things. Sessions are very sensitive > area, so we need to be very careful. I agree. Even for beta, we should be careful. I found issue with object based user save handler. Unlike procedural user save handlers, object based user save handler uses previously used save handler as it's base class (some what) I think we are better to have another SessionHandler object that support new APIs. We can handle create_sid() method rename with new object also. We may keep current implementation undocumented and may document it new one (createSid()) only. I will name it "SessionUpdateTimestampHandler". If anyone has suggestions, I would appreciate it. Regards, -- Yasuo Ohgaki yohgaki@ohgaki.net