|
From: | David Ayers |
Subject: | Re: GDL2, postgres adaptor issues. |
Date: | Fri, 07 May 2004 15:23:18 +0200 |
User-agent: | Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6) Gecko/20040113 |
Simon Stapleton wrote:
Hi again.OK, I've been playing with some fairly heavy duty postgres usage under GDL2, and I've hit a problem or two. Fixable, though.First off, EOKeyValueCoding seems to be broken, around the performXXXforKey: handling. The problem shows itself when you try to use eogenerator to build source files.
This turns out to be an undocumented feature of EOF which allows you to omit the the key used for computational functions. ([anArray valueForKey:@"@myFunc.someKey"] vs. [anArray valueForKey:@"@myFunc"]) My original test showed that with the exception of @count (which is special in other ways) all standard functions need the key. This was actually documented in GDL2 and asserted. But after some private feedback from Simon (and a little support from Alex wrt some efficient string handling :-) ) we "fixed" GDL2 to comply with EOF's behavior and updated the documentation.
Thanks! David
[Prev in Thread] | Current Thread | [Next in Thread] |