Ed Leafe wrote: > On Oct 21, 2007, at 4:39 PM, Paul McNett wrote: > >> The tests used to work, so perhaps that expectation wasn't >> unfounded at >> the time? > > Dunno, but UserSQL has *always* been absolute, with no mods by the > framework.
Didn't we used to have some pattern matching that would extract the where clause if any from the existing SQL, and then put in the child filter? I know I used to be able to set UserSQL, FillLinkFromParent, and LinkField like I do in those tests, and it would work. If I have time this week I'll see about regression testing that to ascertain when it stopped working that way. >> We should write tests for all cases. So, if tests for the >> sqlbuilder are >> missing, those should be written. And if tests for UserSQL are broken, >> those should be fixed. But it isn't either/or. > > I meant that there are two ways to fix this: stay with UserSQL and > write the correct code, making it a test of how well the author of > the test wrote the SQL, or change it to SQLBuilder, and make it a > test of how well the framework manages parent-child requeries. I think that the child filter should be able to work whether the sqlbuilder or UserSQL was set by the developer. And I think that we should test both cases. For now though, I never actually use UserSQL in any of my code, so I'll rewrite those tests to use the sqlbuilder instead. -- pkm ~ http://paulmcnett.com _______________________________________________ Post Messages to: [email protected] Subscription Maintenance: http://leafe.com/mailman/listinfo/dabo-dev Searchable Archives: http://leafe.com/archives/search/dabo-dev This message: http://leafe.com/archives/byMID/dabo-dev/[EMAIL PROTECTED]
