Yes, I did. Let me just follow up on this issue, which has been resolved about a week ago. In the end, it turned out that we had an issue with object cycles and resolved this through a small fix to the way we register objects loaded with the current transaction.
Let me just restate that I really appreciate any kind of feedback with regards to this new feature. Even if we'll be making available a first milestone build for the imminent 0.9.9 release some when next week, I'd still appreciate if as many of you as possible could run tests against the current CVS. Werner Bruce Snyder wrote: > On 7/19/05, Ralf Joachim <[EMAIL PROTECTED]> wrote: > > >>as far as I remember I got that failure at TC97 once after setting up >>the test database. After executing: >> >>update poly_owner set product=1; >> >>on the mysql commandline client this problem disapeared and never >>returned. At that time I thought that this may be a problem with the >>mysql.sql script but after checking this right now I think there is >>something going wrong with unidirectional relations to polymorph classes >>as we have from owner to product. >> >>The stacktrace of TC98 looks ok to me as we try to load the same object >>in 2 independent threads which should lead to a Locking problems. >>Loading the objects in a single thread works as you can see when >>debuging through the other test cases. > > > I'm using MySQL 5.0.7-beta-max. After running the update statement and > rerunning TC97, the test then passes. But TC98 is still failing. Did > Werner write these tests? > > Bruce ------------------------------------------------- If you wish to unsubscribe from this list, please send an empty message to the following address: [EMAIL PROTECTED] -------------------------------------------------
