On Oct 10, 2011, at 11:08 AM, Thomas Mortagne wrote:

> On Mon, Oct 10, 2011 at 10:18 AM, Denis Gervalle <[email protected]> wrote:
>> On Mon, Oct 10, 2011 at 09:49, Thomas Mortagne 
>> <[email protected]>wrote:
>> 
>>> On Mon, Oct 10, 2011 at 9:35 AM, Denis Gervalle <[email protected]> wrote:
>>>> On Fri, Oct 7, 2011 at 15:22, Thomas Mortagne <[email protected]
>>>> wrote:
>>>> 
>>>>> Just seen that we have a bunch of display related method based on
>>>>> Query plugin for which we don't have any alternative to propose right
>>>>> now.
>>>>> 
>>>>> Basically it's supposed to allow to easily display a search form for
>>>>> any field the same way you can display a edit form.
>>>>> 
>>>>> Note that from what I understood from a conversation  we had with
>>>>> Denis it's not working very well.
>>>>> 
>>>> 
>>>> I should confirm that, I would have said that it is not working at all if
>>>> you want to do anything more than very basic search with simple text
>>> field.
>>>> There is issue with list, multi-select, boolean, etc...
>>>> We have used it after some patches, part of them were commited to master,
>>>> but not all.
>>>> So I doubt there is any user of these functions that also intend to
>>> migrate
>>>> to the latest version.
>>>> 
>>>> 
>>>>> 
>>>>> So what do we do after all ?
>>>>> 
>>>>> 1) nothing
>>>>> 2) still deprecate and move to legacy, is app within minutes supposed
>>>>> to provide an alternative ?
>>>>> 3) move to retired
>>>>> 
>>>>> WDYT ?
>>>>> 
>>>>> I'm now -1 for 3) since it's impossible to move method to retired.
>>>>> 
>>>> 
>>>> From what I know of it, and +1 for 2) and even +1 for 3) if it is less
>>> work.
>>>> We use it on 2.4 currently, but if I have to upgrade any site using it to
>>>> 3.x, I will probably rewrite the search part anyway.
>>>> I know we do not have an alternative, but it is a function that is not
>>>> really used anyway due to its caveat. So this is not an issue to retire
>>> it
>>>> now. Of course, we should think about providing similar search feature in
>>>> the future.
>>> 
>>> I'm not really against the idea of retire it but I have no idea how to
>>> retire methods so it's either move them to legacy with aspects or
>>> delete them (which would be a lot easier than putting that in aspect
>>> since right now I have no idea how to put interface method in an
>>> aspect, need to look at AspectJ doc).
>>> 
>> 
>> This what 3) means to me, move the plugin to retire, and delete the method.
>> I am almost sure that building the aspect will be useless, so if this is
>> long and difficult leave it.
>> (I have never converted existing code in aspect, so I do not know well the
>> hard work it could be)
>> From what you say, +1 for 3).
> 
> Anyone else thinking like this ? I would certainly be happy to to
> having to find a way to put all that in aspects…

+1 for 2) ie move to legacy. The reason I say this is because I don't know if 
it's used or not in the wild so I prefer to take the careful route and 1) 
deprecate the methods, 2) move to legacy for a few releases and 3)  then 
completely remove in, say, 4.0M1.

If Ludovic tells us he doesn't remember it to be used anywhere then I'd be +1 
for 3).

Thanks
-Vincent

>> Denis
>> 
>> 
>>> 
>>>> 
>>>> Denis
>>>> 
>>>> 
>>>>> 
>>>>> I would says lets still deprecate and move it to legacy even if we
>>>>> don't provide right now an exact alternative, it's not like it was
>>>>> making impossible to search something since we don't use this at all
>>>>> in XE/XEM since a very long time AFAIK.
>>>>> 
>>>>> On Fri, Sep 30, 2011 at 11:10 AM, Vincent Massol <[email protected]>
>>>>> wrote:
>>>>>> 
>>>>>> On Sep 30, 2011, at 9:06 AM, Thomas Mortagne wrote:
>>>>>> 
>>>>>>> On Fri, Sep 23, 2011 at 1:25 PM, Vincent Massol <[email protected]>
>>>>> wrote:
>>>>>>>> 
>>>>>>>> On Sep 23, 2011, at 1:08 PM, Thomas Mortagne wrote:
>>>>>>>> 
>>>>>>>>> Hi dev,
>>>>>>>>> 
>>>>>>>>> The new query manager component work pretty well and starts to be
>>> used
>>>>>>>>> more and more so I think we should remove
>>> com.xpn.xwiki.plugin.query.
>>>>>>>>> 
>>>>>>>>> So here is the proposals:
>>>>>>>>> (1) move it to legacy
>>>>>>>>> (2) move it to retired is it's own maven project so that it's easy
>>> to
>>>>>>>>> build and produce a jar if someone wants to use it
>>>>>>>>> 
>>>>>>>>> Note planning to do it in 3.2, that's for 3.3M1.
>>>>>>>>> 
>>>>>>>>> WDYT ?
>>>>>>>>> 
>>>>>>>>> +1 for (2): it's the last thing that trigger dependency on
>>> jackrabbit
>>>>>>>>> so even if we get rid of jackrabbit in oldcore by moving it to
>>>>>>>>> legacy-oldcore it would still be in XE
>>>>>>>> 
>>>>>>>> I'd say (1) then (2) since it could be used by some external code.
>>>>>>> 
>>>>>>> It's the same for any other things we retired. The question is more
>>> do
>>>>>>> we still want to maintain it (legacy) or not (retired), in both case
>>>>>>> anyone can still use it if he really wants either by using the 3.2
>>>>>>> version or building the retired one.
>>>>>> 
>>>>>> Denis said he was still using it so I'd say yes to put it in legacy
>>> for
>>>>> now.
>>>>>> Denis, how long should it stay there in your opinion?
>>>>>> 
>>>>>> Thanks
>>>>>> -Vincent
_______________________________________________
devs mailing list
[email protected]
http://lists.xwiki.org/mailman/listinfo/devs

Reply via email to