PHP 5 RC1

Hi!

I think the PHP 5 tree has matured a lot and it's time to release an RC. I
think end of November or mid-December makes sense. I'd like to aim at the
former date.

thats too fast, or do you mean Nov/Dec. 2004? Why such a hurry? Shouldn't we
have a feature freeze first and try to release a stable and powerful relea?se
instead

Are there any major (as in "major") issues which in your
opinion still need to be dealt with?

• MySql License fix (should be done within the next days)

• Documentation
  most of new features aren't documented yet
  Zend2-API Documentation

• QA
  Only a few developers tested PHP5. We need a better qa

• DB Features
  pdo isn't ready yet
  lot of open points how to implement iterators, methods and properties.

More haste, less speed!!

Georg

Hey guys,

I think the PHP 5 tree has matured a lot and it's time to release an RC. I
think end of November or mid-December makes sense. I'd like to aim at the
former date.
Are there any major (as in "major") issues which in your
opinion still need to be dealt with?

RCs are meant for when you think the source tree is ready for release,
but want to make sure. Do you think it's ready?

• Stig

At 09:56 PM 11/11/2003 +0100, Georg Richter wrote:

I think the PHP 5 tree has matured a lot and it's time to release an RC. I
think end of November or mid-December makes sense. I'd like to aim at the
former date.

thats too fast, or do you mean Nov/Dec. 2004? Why such a hurry? Shouldn't we
have a feature freeze first and try to release a stable and powerful relea?se
instead

Do you want to see PHP 5 in Nov/Dec 2004?
I don't see a reason why to delay it for much longer. There are and always
have been backlogs of features/bugs in the past and that didn't stop us
from releasing. Except for a few minor issues the engine has been pretty
much in a feature freeze. As far as I know, most of ext/ too.
If we put an official feature freeze on the current tree, I think a few
weeks is enough to go for an RC. If we find it's not enough and there are
lots of very important bugs we can hold on the RC.

Are there any major (as in "major") issues which in your
opinion still need to be dealt with?

• MySql License fix (should be done within the next days)

This is a political problem and I don't think any PHP version, whether
minor nor major, should wait for this. It's not my fault MySQL have been
very slow to respond. And in any case, it's not as if we're stopping
support for MySQL. On the contrary, it's extremely important and people
will still be able to build it just like any other extension.

• Documentation
  most of new features aren't documented yet
  Zend2-API Documentation

Documentation has always lagged behind release. Stas wrote a short
OBJECTS2_HOWTO but it needs lots of updating. Maybe you can con Marcus into
adding his changes :)

• QA
  Only a few developers tested PHP5. We need a better qa

I don't think it's clear how many people have tested it. The fact is that I
know people who have been using PHP 5 in semi-production environments
already. Last time I wanted to check how many ppl downloaded beta 2 I
couldn't find the stats because either I'm looking in the wrong place or
someone broke the logger.

• DB Features
  pdo isn't ready yet
  lot of open points how to implement iterators, methods and properties.

Who said PDO is a PHP 5 feature? I think it's a nice initiative by Marcus
but the release shouldn't be tied to one ext/ or another (except what's in
ext/ historically).
PHP development doesn't end with PHP 5 and definitely not with the RC.
If/when PDO will be ready, it'll be ready, and possibly integrated into the
standard distribution.

Andi

I went over this list today. Some need to be addressed but the majority are
not engine problems.
I suggest everyone here takes a look at them and tries to solve what they can.
I have already planned time to go over the engine ones.

Andi

At 04:29 PM 11/11/2003 -0500, Ilia Alshanetsky wrote:

Andi,

According to:
http://bugs.php.net/search.php?cmd=display&bug_ntype[]=Feature%2FChange+Request&bug_ntype[]=Documentation+problem&bug_ntype[]=PECL+related&bug_ntype[]=PHP-GTK+related&bug_ntype[]=Website+problem&status=Open&search_for=&php_os=&boolean=0&author_email=&bug_age=0&by=&order_by=id&direction=ASC&phpver=5&limit=50&assign=&begin=50

There are currently over 60 bugs (most Zend Engine 2 related) that are
specific to PHP 5.X. Before moving to RC1 it may be a good idea to review the
situation. Quite a few of these bugs mention crashes and may expose even more
issues in production environment.

Ilia

Andi,

According to:
http://bugs.php.net/search.php?cmd=display&bug_ntype[]=Feature%2FChange+Request&bug_ntype[]=Documentation+problem&bug_ntype[]=PECL+related&bug_ntype[]=PHP-GTK+related&bug_ntype[]=Website+problem&status=Open&search_for=&php_os=&boolean=0&author_email=&bug_age=0&by=&order_by=id&direction=ASC&phpver=5&limit=50&assign=&begin=50

There are currently over 60 bugs (most Zend Engine 2 related) that are
specific to PHP 5.X. Before moving to RC1 it may be a good idea to review the
situation. Quite a few of these bugs mention crashes and may expose even more
issues in production environment.

Ilia

This message was transferred with a trial version of CommuniGate(tm) Pro
Hi,

I think the PHP 5 tree has matured a lot and it's time to release
an RC. I
think end of November or mid-December makes sense. I'd like to aim at the
former date.
Are there any major (as in "major") issues which in your
opinion still need to be dealt with?

i would suggest to wait with RCs and go on with beta-releases. The big
problem with
php5 is, that most of the php-"user" do not know exactly what will change,
what php5
will do better/different and what they will have to change within their
applications.
In my opinion there are two reasons for it:

1.) far to less communication
2.) far to less documentation

these points make it really difficult for the normal developer to migrate to
php5.
Well, you can say now there are lots of articles around, explaining this and
that,
but this is only one point. The most important thing is imho putting these
infos
into the official docs. And - maybe Zend can help here - to enforce the
communication
of php5 and it's features (i.e. pay some good people to write an
extra-column on
zend-website, php-website and wellknow php-portals around the world,
introducing
STEP BY STEP and CONTINUOUSLY the new php5-features, ...). I would call this
a
campaign.
I understand you at zend wish to enforce speed in release of php5, but think
about
"preparing the users for it" and "make it stable"

well, just my two cents

-Wolfgang

--
PHP-Knotenpunkt Dynamic Web Pages:
http://www.dynamicwebpages.de/
PHP 4.3. Professionelle Losungen fur
dynamisches Webpublishing: http://www.php-buch.de/

At 10:39 PM 11/11/2003 +0100, Wolfgang Drews wrote:

This message was transferred with a trial version of CommuniGate(tm) Pro
Hi,

I think the PHP 5 tree has matured a lot and it's time to release
an RC. I
think end of November or mid-December makes sense. I'd like to aim at the
former date.
Are there any major (as in "major") issues which in your
opinion still need to be dealt with?

i would suggest to wait with RCs and go on with beta-releases. The big
problem with
php5 is, that most of the php-"user" do not know exactly what will change,
what php5
will do better/different and what they will have to change within their
applications.
In my opinion there are two reasons for it:

1.) far to less communication
2.) far to less documentation

these points make it really difficult for the normal developer to migrate to
php5.
Well, you can say now there are lots of articles around, explaining this and
that,
but this is only one point. The most important thing is imho putting these
infos
into the official docs. And - maybe Zend can help here - to enforce the
communication
of php5 and it's features (i.e. pay some good people to write an
extra-column on
zend-website, php-website and wellknow php-portals around the world,
introducing
STEP BY STEP and CONTINUOUSLY the new php5-features, ...). I would call this
a
campaign.

I agree. I think we need to take care of this. I wouldn't count on any one
company for this (although this is also important).
How about the people from internals@ who made changes volunteer to write
about their changes?

I understand you at zend wish to enforce speed in release of php5, but think
about
"preparing the users for it" and "make it stable"

Actually it's more of a personal ambition.

Andi

Sorry to interject, but shouldn't the feature set be frozen before
moving to a release candidate? This stage should mean to people that
they can start building software on the new system because all the
features are fixed and are not going to change, even though the system
may still have bugs and is not yet production ready (in fact, that's
what beta should mean, but there was no alpha, so...).

Cheers,

Marco

Andi Gutmans wrote:

Hey guys,

I think the PHP 5 tree has matured a lot and it's time to release an RC.
I think end of November or mid-December makes sense. I'd like to aim at
the former date.
Are there any major (as in "major") issues which in your opinion still
need to be dealt with?

Thanks,
Andi

On Tuesday, Nov 11, 2003, at 19:05 Europe/Copenhagen, Andi Gutmans
wrote:

Hey guys,

I think the PHP 5 tree has matured a lot and it's time to release an
RC. I think end of November or mid-December makes sense. I'd like to
aim at the former date.
Are there any major (as in "major") issues which in your opinion still
need to be dealt with?

Some major win32 build issues that need to be dealt with before RC1:

1. php4 -> php5 rename
2. bundling of libxml2 and dealing with linking issues (static vs.
   dynamic)
3. getting rid of external dll dependancies (iconv.dll atm) which is
   related to issue (2)

Edin