On 4/11/06, Andrus Adamchik <[EMAIL PROTECTED]> wrote: > Yeah, if you could extract a reproducible subset, it would be great.
Yes, I have now reduced this down to a trivial model containing only WOMS_WORK, STATUSABLE_WORK, and WORK, each of which contains only an ID primary key field. > >> Also do you know if this happened with 1.1 or early 1.2 milestones (< > >> M10), i.e. before we've changed the commit code? > > Don't spend much time on that - I was just trying to narrow down the > possibilities by comparing old and new commit code, but this is > really unimportant. Well, it was worth trying anyway as I wanted to run my conversion, even if it meant using an older version of Cayenne. Unfortunately, the differences between HSQLDB 1.8 and 1.7 make it next to impossible to "downgrade" at this point.
