Shuhei Tanuma

yea, I'll plan to support 2.6 feature foreseeable feature. I'd like to use that. let me adjust my schedule :)

@dcelasun we uses this ext in production heavily. so I don't through away. (but my life style has been slightly changed. I'm still seek my life and OSS balance)

yea, but no one publish with namespaced class yet AFAIK. o.O

functional-php guy tried to put docs. but there was no manual on docs.php.net. http://old.nabble.com/DocGen%2C-PHPDOC%2C-PhD-and-namespaced-extensions.-tt31841428.html#a31841428 hm, I'll ask someone to create pecldocs later.

sorry for late response. > (all goes well with 10ms response times until about the 200th sample, at which point response times erratically grow into the thousands of ms. I'm...

you are right. I'd like to support that use case. for now I added `retry_socket` option to above branch. but it might change the option key. I'll test this and...

@EliFinkelshteyn btw, can you tell me your fluentd structure (send messages to local fluentd or remote fluentd directly) and distribution? I'd like to improve error handling regarding this issue so...

@kzk thanks pointing it. i'll test this today.

I merged some commits from retry_socket_opt branch and I'll check that behavior today (if I keep my guts. hehe)

AFAIK, fluentd doesn't have such interface. in_tail might solve your concerns. I recommend to use in_tail and FileLogger if you can switch fluentd configuration. http://docs.fluentd.org/articles/in_tail Basically, PHP is poor about...