Re: [PHP4BETA] PHP 4.0.0 Release packaged

At 21:08 21/05/2000 , Manuel Lemos wrote:
>Hello Stanislav,
>
>On 21-May-00 06:40:03, you wrote:
>
>AP>>> This is total garbage, there are many people putting thier jobs
>AP>>> and businesses on the line by using open source.  This would be
>AP>>> the smack in the face that would get someone to never want to be
>AP>>> associated with open source ever again nevermind loosing thier job
>AP>>> or having thier business ruined.
>
> >Face the reality. Commercial software companies *always* release
> >known-buggy software, there's no way to meet deadline otherwise. Also, I
> >do not see why current release of PHP 4 would ruin someone's business. I
>
>Not that it would ruin the business but some outstanding bugs in PHP 4 and
>incompatibilities with PHP 3 prevent people from upgrading to PHP 4 until
>some developer takes the time to address them.
>
>I can only talk of bugs that affect me, but I'm afraid there are other
>outstanding bugs that affect others.  There is this particular PostgreSQL
>related bug that makes it trash environment variables. That simple prevents
>me from using PostgreSQL in a production site that I have.  I need
>PostgreSQL because it is free and unlike MySQL it supports transactions and
>sub-queries that I need in that site.
>
>I spent an afternoon isolating a minimal sequence of PHP code that is
>enough to reproduce the behaviour so that it will be easier for developers
>to find and fix the bug.  I posted the bug report to the bug database but
>it is still waiting for a developer to give some attention to it.
>
>While this bug is not fixed, I can't recommend the use of PHP 4 with
>PostgreSQL.
>
>
>AP>>> Software development isn't a popularity contest, you do it right,
>AP>>> or you shouldn't do it at all.
>
> >Well, not quite so. If trying to do it right means never finishing doing
> >it, you'd better do it 99% right and then release and then continue
> >working on remaning 1%. Or you never release at all.
>
>At least reproducible bugs should be fixed and documentation should be
>completed with an official release.  That's a feasable goal, unless when
>developers keep on adding features with new bugs every day instead of
>agreeing on a code freeze period that is meant to complete what was already
>started, no new features until the release is officially done.

I'm sorry that the PostgreSQL bug hasn't been fixed;  It's my code, and I 
haven't had the time to look into this yet.  I disagree with you about the 
two points you make:

- The readiness of the documentation shouldn't have anything to do with the 
availability of the software itself.  PHP began with no documentation *at 
all*, and people hacked their way around it.  When documentation was added 
it could be used by a lot more people, but the lack of documentation does 
not mean it's not usable/useful.

- In my opinion, software should be released as soon as it's stable for the 
vast majority of people.  PHP 4.0 satisfies that.  No, it doesn't mean it's 
stable for you, and I won't try to convince you that it's really working 
fine for you when you say there's a PostgreSQL bug - but that PostgreSQL 
bug, whatever it may be, will not affect most people.  It can be fixed in 
between 4.0.0 and 4.0.1.

I'm not sure why people feel 4.0.0 should mark a 'golden version', 100% 
bug-free version of PHP.  Not a single version of PHP in the past (or any 
other piece of software longer than "Hello, world!"), definitely not 3.0.0, 
has been 100% bug free.  4.0.0 today is at a MUCH better situation than PHP 
3.0.0 was at the time.

Zeev
--
Zeev Suraski   <[email protected]>  http://www.zend.com/