> Could you include me in the joke? what is the issue?
This is totally an issue with Entity Modeler -- EOF is fine here .. I
just need my UI behavior to match EOF's. Basically what is going on
is that EM has worked fine, but iTMS would prefer (and it's a
reasonable request) that switching to EM should not cause all the
plist files to get munged around (even though they would end up "the
same" from a runtime perspective, they all look like they were
modified). So they submitted a patch to Entity Modeler that went
through and tweaked the entire plist writing process (and setting of
values that drive the plists) to 100% match EOModeler -- A pain-in-the-
ass job that was really cool of them to do. However, it turns out
that there is an issue with the way specifically allowsNull was being
handled. Essentially, almost everything in EOF I think defaults to
non-existence = "no", whereas it SEEMS to look like non-existence of
allowsNull = "yes" (I haven't confirmed this yet). So I rolled back
those particular changes (just to the allowsnull part) which I think
fixed things, but in exchange for no longer matching output formats
exactly.
I'm in the position at the moment where I just have to create some
sample models with each of the variations of:
* prototype allowsNull = yes, attribute allowsNull = yes
* prototype allowsNull = yes, attribute allowsNull = no
* prototype allowsNull = no, attribute allowsNull = yes
* prototype allowsNull = no, attribute allowsNull = no
* prototype not set, attribute allowsNull = yes
* prototype not set, attribute allowsNull = no
... and see what EOModeler writes into the plist for each scenario.
ms
This archive was generated by hypermail 2.0.0 : Wed Dec 19 2007 - 19:13:23 EST