Re: Pecl request for new extension
| From: | Joshua Thijssen | Date: | Sat, 31 Mar 2012 07:19:01 +0000 |
| Subject: | Re: Pecl request for new extension | ||
| References: | 1 2 3 4 | Groups: | php.pecl.dev |
| Request: | Send a blank email to pecl-dev+get-9276@lists.php.net to get a copy of this message | ||
Patrick Allaert wrote:
> Discard all what I've said Joshua!
I will, but I see your point :-) I think i've tried to be as thread-safe
as possible, but coming from OS-level programming I know this stuff is
hard, even with the TSRML_* macro's :)
Michael Wallner wrote:
> Looks good at a first glance, except... there's one show-stopper:
> According to coding-standards the method names should me 'studlyCaps':
I've updated this on github. All PHP_METHOD files should be studlyCaps.
There is only one problem that I decided not to fix but maybe you guys
know of a better way. It's about cloning the object. The problem there
lies with the actual shout object that is stored in the object store.
This is a datastore so strings can be duplicated and numerical values
can be copied, but it also hosts some socket and queue information. Now
I COULD duplicate the sockets with dup() and such, but I'm not 100% sure
if this will actually function and if shoutcast is able to understand
this. There is another option: these sockets and queues are only used as
soon as you actually have opened the connection. In that case, cloning
an object would depend on wether or not the connection is open, and to
me that doesn't sound like a very good requirement (you can clone, but
not always).
Therefor I decided to disable cloning of these objects alltogether, but
I'm sure there are other ways to go around this (serialization maybe?)