Re: Re: cvs: php4(PHP_4_0_5) /sapi/fastcgi

From: Date: Wed, 21 Mar 2001 12:33:59 +0000
Subject: Re: Re: cvs: php4(PHP_4_0_5) /sapi/fastcgi
References: 1  Groups: php.dev php.qa 
Request: Send a blank email to php-dev+get-49608@lists.php.net to get a copy of this message
At 13:10 21/3/2001, Sascha Schumann wrote:
In my humble opinion (humility is a virtue), new modules are fine to add while in the release process, as long as there's at least one RC after them, to ensure they don't mess up the build or anything trivial like that.
    Oh, messing up the build is a trivial thing?
Yes, it is. Non trivial things are things that are difficult to find, which require long and thorough testing, such as changes to core API functions or very common modules. Figuring whether the build works or not is trivial.
It takes 2-3 months between each and every release, and it'd be a shame for the new module author to receive wide exposure and feedback for his new code.
    As the Midgard extension recently showed, a single line in an
    extension can break everything else.  And it can easily go
    unnoticed, especially if it is not reviewed by experienced
    people.
Easily unnoticed? I beg to differ. As soon as someone tried to build it, it failed. That's why it's trivial, and that's why it's relatively safe to do it even while RC's are going. The rule of 'dont break the CVS build' is crucial, and should be followed especially with RC's. But the worst case situation of adding a new module is a broken build, which is noticed immediately, and thus is not dangerous.
    All new modules can gain wide exposure by distributing them
    separately.  They don't need to be in our tree, and
    they especially don't need to be added during the release
    process.  It is not worth the risk.
The risk is pretty much non existent, as I demonstrated. Especially with a SAPI module, what you say isn't true.
    So few people have asked for the FastCGI module during the
    last two years that I doubt that it was necessary to push it
    into 4.0.5.
Zeus moved its recommendation from using the native ISAPI/Zeus module to using FastCGI, because of instability. So you can expect FastCGI to be more popular.
In my opinion, the CVS rules should be changed to reflect that, just to make Jani happy, or we'd stay outlaws in his mind forever :)
    The CVS rules are fine as they are.
Yes, so says you. But I beg to differ. Zeev -- Zeev Suraski <zeev@zend.com> CTO & co-founder, Zend Technologies Ltd. http://www.zend.com/

« previous php.dev (#49608) next »