Hopbot log for 2009-05-27 - Helma IRC channel: #helma on irc.freenode.net

2009-05-27:

[8:18] <hannesw> any suggestions for feature set or target date for Helma NG 0.4?
[8:18] <hannesw> just dump them here...
[8:59] <simono> persistence that works with mysql (or any other "usual" db)
[9:02] <simono> okay, just saw the hibernate ticket
[9:06] <hannesw> yep, hibernate is almost on par with other persistence backends now.
[10:17] <robi42> hi, ad upcoming helma ng 0.4 release:
[10:18] <robi42> just a short note/thought
[10:19] <robi42> what about providing some kind of code template generation script(s), probably as part of modules/helma/admin?
[10:20] <robi42> e.g., for conveniently generating skeletons for *.hbm.xml ORM files.
[10:21] <robi42> feedback's welcome. :-)
[10:44] <simono> hi robi42 can you show me such a hbm.xml file.
[10:46] <robi42> here's an example taken from the adapted storage demo app: http://github.com/hns/helma-ng/blob/7fe6bf0e38d04efc1c791c3153f9a398a2b8229a/apps/storage/config/Author.hbm.xml
[10:47] <robi42> resp.: http://github.com/hns/helma-ng/blob/7fe6bf0e38d04efc1c791c3153f9a398a2b8229a/apps/storage/config/Book.hbm.xml
[10:48] <simono> thx, never done any hibernate-xml-mapping. very verbose indeed
[10:50] <robi42> yep, it is. especially also all that DOCTYPE etc. boilerplate stuff.
[10:51] <simono> could those xmls be generated on the fly (on startup or so) and fed into hibernate?
[10:52] <robi42> this would be one possibility, yeah.
[10:52] <robi42> or at least generating the skeletons of the files by some kind of code generation script.
[10:53] <robi42> personally, i currently tend more towards the latter.
[10:54] <robi42> this option gives more power to the user.
[10:55] <simono> ah yes, probably nicer to see what gets fed to hibernate
[10:55] <robi42> yep.
[10:55] <robi42> and the xml mapping files are well-documented.
[11:05] <simono> hannesw, are those the same bug http://helma.zumbrunn.com/hopbot/2009-05-25 and http://dev.helma.org/bugs/show_bug.cgi?id=666
[11:06] <simono> else i will report another one :) size() not working is really annoying, though a couple of invalidates() help me work around it
[11:06] <hannesw> robi42: generator scripts == good idea
[11:07] <hannesw> simono: don't know, have to look, but i just fixed #666 locally
[11:07] <simono> awesome
[11:08] <robi42> hannesw: ok, then i'll put that in a new trac ticket.
[11:20] <hannesw> robi42: I also thought about on-the-fly generation of .hbm.xml files.
[11:21] <hannesw> i guess we should use/support whatever works best
[11:23] <robi42> hannesw: possible's both, yet more error-prone probably the on-the-fly generation option.
[11:23] <hannesw> yep. i think an admin script is best choice as first step
[11:24] <robi42> i think so, too.
[11:35] <hannesw> simono: the bugs are definitely not related.
[11:39] <simono> ok, i can file it later
[11:58] <hannesw> simono: i see, select_count is basically broken for oracle collections with limit/maxsize/offset.
[11:58] <simono> yes
[12:00] <simono> just commited it http://dev.helma.org/bugs/show_bug.cgi?id=667
[12:01] <hannesw> ok, thanks
[12:21] <hannesw> ouch, major refactoring due to that bug...
[12:25] <simono> oops... hm. offset is a new feature, right? smth we won't use much
[12:27] <hannesw> not a problem. refactoring is painful but necessary :-)
[13:14] <simono> should this groupedCollection.add(objectA) be equivalent to groupedCollection.get("objectAGroup").add(newObject)
[13:15] <simono> ^^ groupedCollection.get("objectAGroup").add(objectA)
[13:15] <simono> the problem i have is: how do i add / persist a hopobject which belongs to a grouped collection whichs does not yet exist
[13:16] <hannesw> simono: just add it to the main collection.
[13:17] <simono> okay, doesn't work for me.. maybe my fault
[13:17] <hannesw> maybe a bug.
[13:19] <simono> okay it works. but other grouped collections containing this objectA will only get updated after the request is finished or i do res.commit()... hm.
[13:19] <simono> no aggressive caching of course
[13:38] <simono> *facepalm* because this is how helma works. been working w/ agg-caching too much

 

 

In the channel now:

Logs by date: