[dev] Fwd: Re: AJAX Framework changes

Michael J Rubinsky mrubinsk at horde.org
Fri Feb 3 19:59:51 UTC 2012


> Gunnar Wrobel <wrobel at pardus.de> schrieb:
>
> Michael M Slusarz <slusarz at horde.org> schrieb:
>
>> Quoting Michael J Rubinsky <mrubinsk at horde.org>:
>>
>>> Quoting Michael M Slusarz <slusarz at horde.org>:
>>>
>>>> Since the UI is going to be updated for Horde 4.1, it is a  
>>>> necessity to move common AJAX components from the applications  
>>>> (e.g. IMP, Kronolith) into Horde_Core. In fact, I've already been  
>>>> working on this for a few days.
>
> Sounds great.
>
>>>
>>> Wouldn't this be a BC break? Or do you mean that this is going to be
>>
>>> done for H5?
>>
>> Yes, of course this is a BC break. But to redo the UI for 4.1, this  
>> HAS to be done. Or else we are never going to be able to expand our  
>> AJAX and mobile views. How would the UI re-write work without this  
>> (i.e. if every application continues to have its own AJAX  
>> framework?) This is impossible. As a thought experiment: how would  
>> the portal work, for example?
>>
>> And I thought we are over the whole BC breaking thing for  
>> applications since this is the whole point of having PEAR packages.  
>> IMP 5.1/Kronolith 3.1 can require Horde 4.1 and a certain minimum  
>> version of Horde_Core. As long as existing methods don't completely  
>> disappear (so that older applications will still work), this is  
>> perfectly acceptable. We are over the whole "IMP 5.1 has to work  
>> with Horde 4.0" thing.
>
> I have no clue on the exact technical nature of the changes. But the  
> way you describe it I'm not certain it is a BC break. If this only  
> *adds* functionality we can require the newer package version in the  
> applications that have been adapted to the newer versions.
>
> But of course one should still be able to run Imp-5.0 on Horde-4.1.

I hate to bring this up again, but I think there *is* a break here  
that would prevent current x.0 applications from working correctly on  
4.1.  For instance, I just tried to use my local Hermes install, I am  
now completely unable to get it to load the dynamic view because it is  
still looking for the 'mode' session setting that was removed/changed  
for 4.1. This is exactly the type of issue I was trying to avoid by  
moving the view determination out of the application's scope into  
Core. Yes, Hermes isn't yet released for H4, but the method it used  
for determining the view to load is identical to the method Kronolith  
and Imp used.

There is also the loss of functionality I described in my other email  
regarding the ability to explicitly ask for the traditional vs dynamic  
view for each application, but that discussion I'll leave for the  
other thread.

These type of changes, IMO, should require a major version bump.
-- 
mike

The Horde Project (www.horde.org)
mrubinsk at horde.org



More information about the dev mailing list