[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-16 Thread Kris Wallsmith
Even if there were no BC issues I would be opposed to a sfWidgetFormInputBase class. It doesn't make semantic sense. The name implies it is the parent class of sfWidgetFormInput, which is true. However, one would likewise expect sfWidgetFormInputPassword to extend sfWidgetFormInput, which

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-16 Thread Jonathan Wage
In 1.3 we can break the BC of course but not 1.0, 1.1, 1.2, etc. - Jon On Thu, Apr 16, 2009 at 3:25 PM, Tom Boutell wrote: > > That would definitely be an appropriate response for 1.2. Does it make > sense in the 1.3 series? Isn't 1.3 supposed to be bleeding edge > developer-only stuff right no

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-16 Thread Tom Boutell
That would definitely be an appropriate response for 1.2. Does it make sense in the 1.3 series? Isn't 1.3 supposed to be bleeding edge developer-only stuff right now? We are talking about 1.3, right? (: On Thu, Apr 16, 2009 at 1:36 PM, Jonathan Wage wrote: > But that does not matter. Most people

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-16 Thread Jonathan Wage
But that does not matter. Most people don't pay attention to what changes are in a point release. They simply do pear upgrade/svn update and they expect that it isn't going to break their application. If it does, then that person lost the little bit of trust they had in symfony and probably won't e

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-16 Thread Dennis Benkert
I agree with Tom. In this special case I would say that the benefit of this changes is higher than the pain you'll get to change you're code. - Dennis 2009/4/16 Tom Boutell : > > I suppose it would be a BC break for anyone who is already using > instanceof() to check whether something is one of

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-15 Thread Jonathan Wage
True. I see now. Either way it is a BC break. - Jon On Wed, Apr 15, 2009 at 6:13 PM, Tom Boutell wrote: > > I suppose it would be a BC break for anyone who is already using > instanceof() to check whether something is one of sfWidgetFormInput's > subclasses. But isn't this the sort of very mino

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-15 Thread Tom Boutell
I suppose it would be a BC break for anyone who is already using instanceof() to check whether something is one of sfWidgetFormInput's subclasses. But isn't this the sort of very minor incompatibility that would almost have to be allowed for 1.3 to differ meaningfully from 1.2? On Wed, Apr 15, 20

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-15 Thread Kris Wallsmith
I considered this option but I see it as a BC break because sfWidgetFormInputPassword (for instance) would no longer extend sfWidgetFormInput. -- Kris Wallsmith | Community Manager kris.wallsm...@symfony-project.com Portland, Oregon USA http://twitter.com/kriswallsmith On Apr 15, 2009, at 3

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-15 Thread Jonathan Wage
So right now we can't identify a normal input field by doing instanceof sfWidgetFormInput because all the other widget types extend that? If so, what about this so that it maintains BC and you still use sfWidgetFormInput. 1.) Rename sfWidgetFormInput to sfWidgetFormInputBase(or something similar)

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-15 Thread Tom Boutell
I agree, this is a smart choice. On Wed, Apr 15, 2009 at 3:27 AM, Dennis Benkert wrote: > > Nice improvement. +1 for it. > > - Dennis > > 2009/4/14 Kris Wallsmith : >> I've just opened a ticket on Trac to add a sfWidgetFormInputText class to >> the 1.3 branch. The addition of this class will set

[symfony-devs] Re: RFC: sfWidgetFormInputText

2009-04-15 Thread Dennis Benkert
Nice improvement. +1 for it. - Dennis 2009/4/14 Kris Wallsmith : > I've just opened a ticket on Trac to add a sfWidgetFormInputText class to > the 1.3 branch. The addition of this class will set text fields on the same > level as password and file upload fields and allow developers to identify >