Re: Quality control of PECL modules
| From: | Andi Gutmans | Date: | Thu, 23 Feb 2006 04:06:59 +0000 |
| Subject: | Re: Quality control of PECL modules | ||
| References: | 1 2 3 4 5 6 7 | Groups: | php.pecl.dev |
| Request: | Send a blank email to pecl-dev+get-3531@lists.php.net to get a copy of this message | ||
Steph,
I actually had no idea what there was a discussion going on. I just sent this email after I was asked by someone to recommend which extensions are stable and which aren't. It was purely an out-of-context email and it seems that the timing made you interpret it differently.
I still believe that forget Windows issues or whatever, we need to find a much better way to make sure that we can recommend extensions and warn of others. This is more than just a "stable" mark, but has to do with maturity, maintenance, docs, etc... This is the more interesting discussion.I don't really care what tactical band-aids are used for Windows builds.
Andi
At 08:00 PM 2/22/2006, Steph Fox wrote:
Andi, This entire issue came up because Pierre was re-developing a PECL extension from scratch in CVS HEAD at the point of a PHP release and there was no way for Edin's current system to know it. It was a purely win32 issue, and that's why a win32 issue (only) was discussed on IRC. Whatever he's coming up with next is probably also important, but please don't dismiss the actual problem we were attempting to deal with in the first place. - Steph ----- Original Message ----- From: "Andi Gutmans" <andi@zend.com> To: "Pierre Joye" <pierre.php@gmail.com>; "Steph Fox" <steph@zend.com> Cc: <pecl-dev@lists.php.net> Sent: Thursday, February 23, 2006 5:28 AM Subject: Re: [PECL-DEV] Quality control of PECL modulesAt 06:52 PM 2/22/2006, Pierre Joye wrote:As far as stability goes - yeah, we've already been talkingOn 2/23/06, Steph Fox <steph@zend.com> wrote:irc in the context of win32 .dll distributions. PECL alreadyabout this > inmarking code stability for src releases.has a way > ofJust about marking code stability. The problem is not to mark it stable, the problem is to define what is stable. For example, I was quietly loughing when I saw the first or second filter release with a beta state. Not that the code was bad (it actually did not work that well for beta), but the fact is that codes need users to get mature, and tests.Exactly. I think we need another level. Also, please remember, not all of us have the ability to hang out on IRC, so I think giving us the opportunity to also give our $.02 on email is good/fair practice. I'd definitely like this to be also discussed on this mailing list. It's not rocket science but being able to provide feedback like the current "stable" not working is important IMO. Andi __________ NOD32 1.1380 (20060125) Information __________ This message was checked by NOD32 antivirus system. http://www.eset.com