Hopbot log for 2008-07-02 - Helma IRC channel: #helma on irc.freenode.net

2008-07-02:

[4:10] <midnightmonster> jkridner, I've got good clean sample code for your move-obj-creation-code-to-obj-instead-of-parent issue
[4:28] <midnightmonster> http://helma.pastebin.com/m47564fa0
[9:38] <lapusta> just found matsblog and photos from spring meeting, there was one interesting note with JS stroke out and Python written beside. can anyone tell me what this idea was about?
[9:39] <zumbrunn> one of the guys suggested it as a topic
[9:39] <zumbrunn> why JS? and not Jython
[9:39] <zumbrunn> but we never got around to really talk about that ;-)
[9:40] <zumbrunn> it was a bit tongue in cheek, I guess
[9:40] <zumbrunn> helma-ng in many ways does take some inspiration from phyton
[9:41] <zumbrunn> specially with the way it uses scopes with its module scope concept
[9:42] <zumbrunn> and some changes in ES4 as well are python inspired, I guess
[9:45] <zumbrunn> btw, the way helma 1 was written was intentionally architected in a way so it would be possible to add other languages
[9:46] <zumbrunn> it's one of the things that go in the way a bit in the helma 1 code
[9:46] <zumbrunn> that different parts of helma, like the orm, were to far separated from the javascript language itself
[9:47] <zumbrunn> helma-ng on the other hand does a lot more directly in javascript
[9:49] <zumbrunn> (which still doesn't mean other languages couldn't be used as well, of course, in addition to js)
[9:50] <zumbrunn> Example: Want ES4 in Helma today?
[9:50] <zumbrunn> Then take the just released Mascara Javascript 2 to Javascript 1 converter written in Python, run it using Jython on the Java Virtual Machine, and invoke it using Rhino inside of Helma, evaluating the results.
[9:50] <zumbrunn> ;-)
[9:56] <lapusta> BTW i was reading ng mail-lists and looking through sandbox projects and found out there are different views on some component implementation like orm(mainly java vs pure js). i think it would be nice to take pylons/merb way(small core with plugins) than rails/django(full-stack) to avoid such conflicts, though may be this way is already chosen
[9:56] <zumbrunn> nothing is chosen yet, I believe
[9:56] <zumbrunn> it's all experimental
[9:58] <zumbrunn> plus, for the next years, having diversity in the world of "javascript on the server-side" is a good thing, I think
[9:58] <zumbrunn> there will likely be an insane amount of different approaches
[9:59] <zumbrunn> and many will probably be built on top of the helma core
[10:00] <zumbrunn> in a sense, helma-ng *is* that small core you are talking about
[11:13] <jkridner> midnightmonster: thanks!
[11:18] <jkridner> it seems to me like this only allows one type of child determined by 'type'.
[12:22] <midnightmonster> jkridner, it allows only one type of child as specified in _children(MyChildType)
[12:23] <midnightmonster> I find for my convenience that often when I would do, e.g., projects = collection(Project) (Where there's already a _children = collection(SoemthingElse)) that I'm better served by doing projects = mountpoint(ProjectList), then ProjectList has _children = collection(Project)
[12:24] <midnightmonster> (the "type" variable is pulled dynamically from your type.properties)
[12:24] <jkridner> oh.
[12:25] <jkridner> would it be possible to do something like /parent/new?type=ChildType ?
[12:25] <midnightmonster> yes, trivial.
[12:25] <midnightmonster> I was thinking /parent/new_ChildType
[12:25] <jkridner> does it matter that the type could be different than _children(ChildTypeA)?
[12:26] <midnightmonster> are you using the object db or the sql?
[12:26] <jkridner> object db for simplicity.
[12:26] <midnightmonster> then your types will need to be able to figure out where they should attach themselves.
[12:26] <jkridner> I started with SQL, but the object db seemed more flexible, and I don't have a ton of objects.
[12:27] <midnightmonster> my code would probably need slight modification to work with the object db
[12:27] <jkridner> k
[12:27] <midnightmonster> er, not the code I posted
[12:27] <jkridner> k
[12:27] <midnightmonster> just the part where you actually save the new object
[12:28] <midnightmonster> if pseudoParent == future _parent, that's easy.
[12:28] <jkridner> this was quite helpful. I'm going to head into work now and give it some more thinking this afternoon.
[12:28] <midnightmonster> but if you have more than one type, and they don't all go in _children, you have to hard code more stuff
[12:29] <midnightmonster> also, if you support multiple types you need to whitelist which types are allowed
[12:29] <midnightmonster> alright. ttyl. gtg myself.
[12:29] <jkridner> that's going over my head just a bit right now, but I'll try to comprehend it later.
[12:29] <jkridner> thanks again!!!
[16:56] <decke> does anyone know what could be the cause of lots of "[2008/07/01 15:22:38] Stopping Thread for Request club3k/np/xXxsummergirlxXx/guestbook/new" error messages in the helma logfile?
[16:57] <decke> does this mean something like "the request took too long to respond and gets stopped now" ?
[20:39] <midnightmonster> decke, it means Helma decided xXxsummergirlxXx's name is too suggestive, and she must be stopped. it's for her own good, really.

 

 

In the channel now:

Logs by date: