Hopbot log for 2007-06-26 - Helma IRC channel: #helma on irc.freenode.net
2007-06-26:
[12:41] <fab_> hi, is there a way to generate a sql schema from a collection of type.properties files?[12:42] <zumbrunn> hi fab_
[12:42] <zumbrunn> maybe with the sqlshell
[12:42] <fab_> or is there a better way to create a O/R mapping than writing all type properties _and_ the sql schema
[12:44] <zumbrunn> So you know about Rabbit and Rocket?
[12:44] <fab_> zumbrunn, i tried using the sqlshell, but as far as i understand, it can't create type properties from sql schemas, but is more like an editor for type.properties
[12:44] <fab_> rabbit and rocket?
[12:44] <zumbrunn> http://dev.helma.org/wiki/Rabbit/
[12:45] <zumbrunn> And then there is Warp, as well...
[12:45] <zumbrunn> http://helma.org/pipermail/helma-dev/2007-May/003571.html
[12:45] <fab_> sounds interesting :), thank you
[16:14] <fab_> is there a way to define abstract base classes or interfaces in the O/R mapper?
[16:16] <fab_> err, interfaces don't make sense... but base classes would
[16:17] <zumbrunn> not sure what you mean
[16:19] <fab_> i have something like "b/type.properties: _extends = a" and want objects of type b, but no objects of type a
[16:20] <fab_> to specify that no objects of type a can exists, only objects of subclasses of a
[16:22] <zumbrunn> I don't think there is a way to prevent that
[16:23] <zumbrunn> why would you need to prevent it?
[16:23] <zumbrunn> I mean, if you don't create any objects of type "a", they won't exist
[16:24] <zumbrunn> well, at least one will exist, in a sense, since b.prototype would be an object of type "a", right?
[16:25] <fab_> if you can specify that no such objects exist, you dont have to handle them in your code
[16:25] <fab_> otherwise someone might accidentaly create such objects and you have unknown paths of your program
[16:26] <fab_> for that reason abstract classes exists in most OO programming languages
[16:28] <fab_> for example you could have some classes of objects for which a method exists, which makes sense for each subclass but not for the "abstract" base class
[16:29] <zumbrunn> yes, makes sense...
[16:29] <zumbrunn> but js doesn't offer that
[16:29] <fab_> yes, but i thought maybe helma does ;)
[16:30] <fab_> well, its no problem to get around that, but it would be "nice-to-have"
[16:32] <zumbrunn> you could get the same effect in js using closures
[16:33] <zumbrunn> but that would only work if you create the prototypes programatically
[16:34] <zumbrunn> which would get you into trouble when you want to use type.properties...
[16:34] <zumbrunn> since they can'^t be set programmatically right now as far as I know
[16:34] <zumbrunn> *that* would be nice as well
[16:36] <zumbrunn> (basically, prototype "a" would exist only inside a closure at the you create the "b" prototype)
[16:36] <zumbrunn> ...at the time you...
[16:39] <fab_> hhm
[16:43] <fab_> would it be hard to make the prototypes accessible from js?
[16:44] <zumbrunn> uhm, what do you mean by that?
[16:44] <zumbrunn> they are accessible from js
[16:44] <fab_> i mean the type definition
[16:45] <fab_> somehow i don't like the concept of defining them by type properties very much
[16:45] <zumbrunn> I'm not sure how much is possible through code right now
[16:46] <fab_> for example it would be nice to derive the object relations and fields from the database schema at startup time
[16:46] <zumbrunn> that's kind of what Warp does
[16:46] <zumbrunn> http://helma.org/pipermail/helma-dev/2007-May/003571.html
[16:47] <fab_> ah, then i misunderstood the post
[16:47] <zumbrunn> or maybe I did ;-)
[16:48] <zumbrunn> Warp does this in java, though
[16:49] <zumbrunn> it creates prototypes based on db schemas and exposes them to Helma as code repositories
[16:50] <fab_> will definitly try it out the...
[16:50] <fab_> have to leave now...
[16:51] <fab_> thanks for your help!
[16:51] <zumbrunn> ok, let us know how it goes!
[16:51] <fab_> will be back tomorrow