Michal Čihař schrieb:

Hi

On Thu, 14 Jun 2007 07:39:32 -0400 Marc Delisle <Marc.Delisle@cegepsherbrooke.qc.ca> wrote:

...

phpMyAdmin has been invited by a group of well-known open source projects to participate in an initiative that is meant to promote PHP 5 and break the cycle where web hosts are not encouraged to upgrade to 5 because apps works on 4.

Specifically, if we agree to this, effective 5 February 2008, any new-feature release will have a minimum required PHP version of 5.2.0.

I like the idea of a group of projects doing this (so no one is penalized by being the first); I am in favor of this and would like your feedback.

I think it only helps us when we don't have to stay compatible with PHP 4.

??? yes, this is how i read it, rising PHP requirements to PHP 5.2.0 - breaking PHP 4 cmpatibility

On Thu, 14 Jun 2007 08:39:13 -0400 Marc Delisle <Marc.Delisle@cegepsherbrooke.qc.ca> wrote:

Sebastian Mendel a écrit :

...

Michal Čihař schrieb:

...

Hi

On Thu, 14 Jun 2007 07:39:32 -0400 Marc Delisle <Marc.Delisle@cegepsherbrooke.qc.ca> wrote:

...

phpMyAdmin has been invited by a group of well-known open source projects to participate in an initiative that is meant to promote PHP 5 and break the cycle where web hosts are not encouraged to upgrade to 5 because apps works on 4.

Specifically, if we agree to this, effective 5 February 2008, any new-feature release will have a minimum required PHP version of 5.2.0.

I like the idea of a group of projects doing this (so no one is penalized by being the first); I am in favor of this and would like your feedback. I think it only helps us when we don't have to stay compatible with PHP

???

yes, this is how i read it, rising PHP requirements to PHP 5.2.0 - breaking PHP 4 cmpatibility

I think what Michal meant is that rising the requirements is a positive thing for our project.

Exactly ;-) -- Michal Čihař | http://cihar.com | http://blog.cihar.com

Sebastian Mendel a écrit :
> Michal Čihař schrieb:
>> On Thu, 14 Jun 2007 08:39:13 -0400
>> Marc Delisle <Marc.Delisle@cegepsherbrooke.qc.ca> wrote:
>>
>>> Sebastian Mendel a écrit :
>>>> Michal Čihař schrieb:
>>>>> Hi
>>>>>
>>>>> On Thu, 14 Jun 2007 07:39:32 -0400
>>>>> Marc Delisle <Marc.Delisle@cegepsherbrooke.qc.ca> wrote:
>>>>>
>>>>>> phpMyAdmin has been invited by a group of well-known open source 
>>>>>> projects to participate in an initiative that is meant to promote PHP 5 
>>>>>> and break the cycle where web hosts are not encouraged to upgrade to 5 
>>>>>> because apps works on 4.
>>>>>>
>>>>>> Specifically, if we agree to this, effective 5 February 2008, any 
>>>>>> new-feature release will have a minimum required PHP version of 5.2.0.
>>>>>>
>>>>>> I like the idea of a group of projects doing this (so no one is 
>>>>>> penalized by being the first); I am in favor of this and would like your 
>>>>>> feedback.
>>>>> I think it only helps us when we don't have to stay compatible with PHP
>>>>> 4.
>>>> ???
>>>>
>>>> yes, this is how i read it, rising PHP requirements to PHP 5.2.0 - breaking
>>>> PHP 4 cmpatibility
>>> I think what Michal meant is that rising the requirements is a positive 
>>> thing for our project.
>> Exactly  ;-)
> 
> ... mhm, oh, when !== if ... i understand ... ;-)

I don't detect any objections in your replies, so we "sign" to be part 
of this initiative?

Marc

Hi Marc & list, I just noticed this interesting discussion and felt like joining it. I definitely support the idea of dropping php 4 support. phpMyAdmin has a manpower problem so getting rid of costly compatibility unmaintainable might be a step into the right direction. Marc Delisle schrieb:

...

and drop MySQL 3.x and 4.0 support to?

=20 And I don't know about the market penetration for MySQL 3.x and 4.x. I =

am in favor of dropping support for MySQL 3.x.

Compatibility drops like this should not be made often, otherwise you create many legacy versions the need to be maintained and security-patched. This is why I'd suggest that support for older MySQL branches should be dropped together with php 4. Current phpMyAdmin versions should support the full feature set of MySQL 3.23, 4.0 and 4.1. Users of those branches won't benefit much from future phpMyAdmin releases, imho. I know, that the idea of working with MySQL >=3D 5 only might sound a bit= exaggregated at first glace, but dropping support for older MySQL releases would make a lot of compatibility code obsolete. Dropping MySQL 3.23, supporting MySQL >=3D 4.0.12: * UNION and TRUNCATE may be used without checks. * Privileges introduces with 4.0 are supported by the server. According to MySQL, they do not offer any kind of support for 3.23 anymore, so why should we? Dropping MySQL 4.0, supporting MySQL >=3D 4.1.7: * The server supports recoding. All non-UTF-8 language files and all the recoding stuff may be dropped. * The server supports subselects. The 4.0 branch has reached its "end of product lifecycle" according to MySQL. Dropping MySQL 4.1, supporting MySQL >=3D 5.0.15: * information_schema can be used for queries on the structure. This is often much faster than the stuff we have to do in versions before 5.0 (for instance when calculation the statistics on the databases page). The old SHOW commands don't give us all the information we need in more recent MySQL versions - for instance we need to query information_schema if we want to distinguish between base tables and views. * We don't need a controluser anymore to query user privileges, information_schema provides the information we need. Like with 4.0, MySQL provides extended support for 4.1, but does not provide patches for it anymore. Furthermore, upgrading from 4.1 to 5.0 is really not that painful as it was when upgrading from 4.0 to 4.1. MySQL 5.0 is stable since October 2005 (v5.0.15), so I guess it should be safe to drop support for earlier versions in 2008. With MySQL, I see a similar problem as with php: webhosters for some reason stick with legacy versions (3.23 or 4.0). If you want to put pressure on them, you should not only do this with php, you should do it with MySQL as well. Regards, Alexander

On Thu, Jun 14, 2007 at 02:24:40PM +0200, Sebastian Mendel wrote:

and drop MySQL 3.x and 4.0 support to? Drop MySQL-3.x sure, but please don't drop 4.0 for another 6 months or so. I still have some consulting customers (with paid MySQL support contracts) that are only jumping from 4.0 to 5.1 when Christmas comes.

-- Robin Hugh Johnson E-Mail : robbat2@users.sourceforge.net GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85

Robin H. Johnson schrieb:

On Thu, Jun 14, 2007 at 02:24:40PM +0200, Sebastian Mendel wrote:

...

and drop MySQL 3.x and 4.0 support to? Drop MySQL-3.x sure, but please don't drop 4.0 for another 6 months or so. I still have some consulting customers (with paid MySQL support contracts) that are only jumping from 4.0 to 5.1 when Christmas comes.

yes, isn't Christmas before 2008? -- Sebastian

Marc Delisle schrieb:

P.S. We should also define what we mean by dropping support. Here are a few points:

- We announce the fact on phpmyadmin.net: yes

- We no longer test for non-supported platform: yes

- We block and/or emit warnings for non-supported platforms: ?

block with phpMyAdmin release of 2008-02? warn in next releases of upcoming change if PHP or MySQL is below 5

- We remove code that relates to those platforms: ?

of course

Sebastian Mendel a écrit :

Marc Delisle schrieb:

...

P.S. We should also define what we mean by dropping support. Here are a few points:

- We announce the fact on phpmyadmin.net: yes

- We no longer test for non-supported platform: yes

- We block and/or emit warnings for non-supported platforms: ?

- We remove code that relates to those platforms: ?

- use only UTF-8, including lang files?

Good point.

Thijs Kinkhorst a écrit :

On Thu, June 14, 2007 13:39, Marc Delisle wrote:

...

phpMyAdmin has been invited by a group of well-known open source projects to participate in an initiative that is meant to promote PHP 5 and break the cycle where web hosts are not encouraged to upgrade to 5 because apps works on 4.

Specifically, if we agree to this, effective 5 February 2008, any new-feature release will have a minimum required PHP version of 5.2.0.

I like the idea of a group of projects doing this (so no one is penalized by being the first); I am in favor of this and would like your feedback.

Thanks Thijs for your feedback.

I'm not a member of the development team, but I'm missing the most important thing here: what are the concrete gains??

- Coding without always having to stay compatible with PHP 4 (gains in terms of OOP, removing some dead code) - Because the PHP team will eventually stop to maintain PHP 4: avoiding to support our users who have PHP4-related problems

The most concrete drawback is that it will be hurting those users who have no control over their host's PHP config or have other reasons not to upgrade to PHP 5 yet.

We are talking about not releasing for PHP 4 our new-feature releases after 5 Feb 2008. Users who want to run the current phpMyAdmin version on PHP4 hosts will be able to; meanwhile they can put pressure on their hosts. And remember, this is part of a bigger movement that should convince providers to switch.

I can see a move to PHP5 only have some benefits, but what are those concretely? Can we have a list of improvements that will be possible in phpMyAdmin with this decision? And do they offset the drawback for those users I mentioned?

Sorry I cannot produce such list of improvements for the moment.

And if you decide to move to PHP5 only, how will PHP4-bound users be security-supported?

You are right, we will have to deal correctly with this; the stable phpMyAdmin version on 5 Feb 2008 will have to be maintained for some time.

thanks, Thijs

Thijs Kinkhorst a écrit :

On Thursday 14 June 2007 17:20, Marc Delisle wrote:

...

...

The most concrete drawback is that it will be hurting those users who have no control over their host's PHP config or have other reasons not to upgrade to PHP 5 yet. We are talking about not releasing for PHP 4 our new-feature releases after 5 Feb 2008. Users who want to run the current phpMyAdmin version on PHP4 hosts will be able to; meanwhile they can put pressure on their hosts. And remember, this is part of a bigger movement that should convince providers to switch.

For me personally it's not about "hosting providers" but rather in-house custom developed systems. I've had the experience that the upgrade from PHP4 to PHP5 can be quite non-trivial, especially if you have a large custom application that needs to be ported.

I'm therefore wary of updating requirements of PHP 5, unless there's a demonstratable benefit to it. I'm also active in another project, SquirrelMail. We've noticed that there have only been very few items where a PHP 5 construct would have done something slightly better, but we were able to fix those just by using a couple of lines of code extra. If that's all it takes, it doesn't outweigh excluding those not ready to upgrade yet.

To me a logical point to stop supporting PHP4 would be when it either becomes a too big a burden for the code (which I'm seeing not much concrete examples of), or when PHP itself stops support for it.

...

...

And if you decide to move to PHP5 only, how will PHP4-bound users be security-supported? You are right, we will have to deal correctly with this; the stable phpMyAdmin version on 5 Feb 2008 will have to be maintained for some time.

I think this is the most important thing to take care of if you decide to go through with this.

...

- We announce the fact on phpmyadmin.net: yes

- We no longer test for non-supported platform: yes

- We block and/or emit warnings for non-supported platforms: ?

- We remove code that relates to those platforms: ?

If you go through with this, blocking users of PHP4 seems strange. A warning should be sufficient. If phpMyAdmin still works for a PHP4 user, why deny them that?

If you decide to go PHP5, then do it only when developing new code or where there's a measurable benefit in replacing the old code.

thanks, Thijs

Thijs, We have decided to go PHP 5. Thanks for your input. The exact steps we will do are still under discussion amongst the phpMyAdmin team and I'll discuss it also with other members of the "Go PHP5" coalition. We feel that the cycle where PHP dev team still has to support PHP 4 because providers still use it (because PHP dev team still supports it ...) needs to be broken. Marc

Hi Thijs & list, Thijs Kinkhorst schrieb:

...

I'm not a member of the development team, but I'm missing the most important thing here: what are the concrete gains??

As long as phpMyAdmin is expected to work on php 4, it has to be tested on php 4. Even if only a small patch is applied to the code, the developers have to ensure that it does not break the php 4 compatibility. Furthermore, old php versions may behave differently than the latest ones, so workarounds and compatibility code have to be written, which makes the code less maintainable. This takes a lot of time. Time that could be used for real improvements. On top of that, php 5 with its OO features allows to write code that is much more readable and maintainable. This would be a great advantage when reengineering parts of phpMyAdmin - which will definately be nessecary sooner or later.

...

The most concrete drawback is that it will be hurting those users who have no control over their host's PHP config

True, but honestly: if my webhoster did not offer php 5 by now, I'd be looking for a new one.

...

or have other reasons not to upgrade to PHP 5 yet.

It is possible to run php 4 and php 5 in paralell, so legacy applications that need php 4 may have php 4 while the rest is run under php 5. And if this is not an option, older versions of phpMyAdmin will still work - just like older versions of php. ;-)

...

And if you decide to move to PHP5 only, how will PHP4-bound users be security-supported?

This is something that needs to be worked out. As far as I can tell, this did not work well, when we dropped the php 3 support. Regards, Alexander