Re: Better key path creation
Re: Better key path creation
- Subject: Re: Better key path creation
- From: Jean-Francois Veillette <email@hidden>
- Date: Thu, 6 Jan 2011 08:52:25 -0500
I just read the referenced mail thread.
It seem very interesting !
Mike S. have real concern which are potential problem for specific cases.
Is it possible to have an option in the template to generate (or not) the extra classes and new keypath support ?
The option for extra keypath support could be enabled from the entity's user-info variable dictionary, or from a model.eogen variable (for all entities).
That way this could be enabled and used as needed.
jfv
Le 2011-01-05 à 04:25, Marc Guenther a écrit :
> How interesting. I almost missed this. I posted something similar a while ago (http://www.mail-archive.com/email@hidden/msg37440.html). Unfortunately it didn't generate much interest. I wonder if people just don't know what they are missing out on. Come on people, this stuff is way cool! :)
>
> Regarding naming collisions. Why don't you simply reuse the existing generated ERXKeys, instead of this "path" thingy? After all, you would replace them with subclasses of their own, and that wouldn't break anything for anybody. Then use uppercase everywhere. I never had any collisions that way.
>
> About the permgen issue, I have to say I simply don't care. John's proposal adds one additional class to the already existing two for each entity. That's certainly not dumping 3000 classes into the classpath and shouldn't be an issue at all.
> My proposal uses 4 additional classes (two of which are completely empty) cause I differentiate between keys and keypathes, and between to-one and to-many keys. I didn't have to increase permgen yet (150 entities) and even if I would have to, so what? It's a constant one time hit, which doesn't increase at runtime.
>
> Well, enough said. I use this stuff for a long time now, and it's so much easier to read and type. It's simply wonderful.
>
> Have fun,
> Marc
>
> --
> Sent from my iPhone
>
> On 21.12.2010, at 16:41, Ramsey Gurley <email@hidden> wrote:
>
>>
>> On Dec 20, 2010, at 10:15 AM, John Huss wrote:
>>
>>> The WOLips one is the real one. You can go vote for it if you'd like to have it.
>>>
>>> Thanks,
>>> John
>>>
>>> On Mon, Dec 20, 2010 at 5:30 AM, Timo Hoepfner <email@hidden> wrote:
>>> Already there...
>>>
>>> http://issues.objectstyle.org/jira/browse/WONDER-617
>>> http://issues.objectstyle.org/jira/browse/WOL-1189
>>
>> Does this still produce the collisions that Timo observed? If so, I know it isn't pretty, but I think the patch may stand a better chance if it has a prefix like
>>
>> MovieRole.$path.$talent().$photo().$photo().key()
>>
>> Otherwise, the next time someone wants to add a new method to ERXKey, the change is going to be subject to everyone getting upset because it brakes with their model. But then, I also remember Mike saying something about perm cache at some point in time...
>>
>> Ramsey
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden