Re: Entity Modeler SQL generation trouble

From: Jerry W. Walker (jerrywwalke..mail.com)
Date: Wed Nov 14 2007 - 08:12:27 EST

  • Next message: Mike Schrag: "Re: WOLips Package explorer"

    On Nov 13, 2007, at 8:20 PM, Mike Schrag wrote:

    >> The "normal Java pattern" part was using static ivars as
    >> constants; the DB access is my own special sauce.
    >>
    >> So I guess it's me that drives me nuts, not WO. I really need to
    >> find a mid-level java overview book; not how-to-program, but how-
    >> things-work. I've never had occasion to learn much about the
    >> classloader, or how the JVM handle static initialization, but
    >> these are clearly things worth knowing about, since it was
    >> immediately obvious to you and Mike and Anjo exactly how I was
    >> abusing the language to shoot myself in the foot just so. Not that
    >> I expect a book to instantly enlighten me and keep me from posting
    >> stupid questions to every list I subscribe to, but it might help a
    >> little bit. Any recommendations?
    > To be perfectly honest, I consider these kind of things to be
    > fringe concepts ... It's obvious to us because we wallow in the mud
    > of EOF all the time, but there's really no book that will teach you
    > these things. There's a certain amount of learning any framework
    > that only comes from shooting yourself in the foot a few times. I
    > think this is true of any framework, but especially of a very large
    > application framework. If this isn't already an EOF commandment, it
    > might be worth adding it, because you're certainly not the first
    > person to have been screwed by this. However, I also don't think
    > it's necessarily EOF's _fault_ -- there are some problems that are
    > just nasty to solve, and the process of classloading is really a
    > complex problem.
    >
    > ms
    >

    So how would such an EOF Commandment be phrased? "Don't do EOF
    fetches in static initializers." or something more general, such as
    "Don't reference persistent data in static initializers."?

    Somehow, the former sounds too limited (i.e. doesn't cover the entire
    problem), but the latter seems overreaching. On the other hand,
    referencing the class loader or timings related to the class loader
    tend to be unusable (not understandable) by the very audience at
    which the commandments are aimed.

    Suggestions?

    Regards,
    Jerry

    --
    __ Jerry W. Walker,
        WebObjects Developer/Instructor for High Performance Industrial  
    Strength Internet Enabled Systems
    

    jerrywwalke..mail.com 203 278-4085 office



    This archive was generated by hypermail 2.0.0 : Wed Nov 14 2007 - 08:14:09 EST