RFP: back out the incremental loader changes for 1.0

From: Dom Lachowicz (doml@appligent.com)
Date: Thu Feb 14 2002 - 18:53:34 GMT

  • Next message: Dom Lachowicz: "Re: Fwd: using the printing engine of abiword"

    Hi,

    I *really* don't want to step on anyone's toes here, especially
    Martin's, whom I told to check this in. This is some tremendous work and
    I certainly don't want to just throw away weeks' worth of hard work. I
    *really* love the incremental loader...

    when it is fully working, that is. My current problem is that after the
    new incremental loader, I and others have experienced a great deal of
    problems that we hadn't before then, including but not limited to:

    1) Unexplained crashes on document load
    2) Unexplained hanging on document load
    3) Unexplained crashes on zooming
    4) Unexplained exiting if closing 1 window while another doc is opening
    5) Inability to print from the command line any more, mostly due to a
    lack of an XAP_Frame to call nullUpdate() on while loading the doc.

    The worst part of #s 1-4 is that they are non-deterministic, though not
    hard to reproduce for us. "If at first you don't fail, try, try again."
    If these sorts of behavior aren't reproducible for you, then it's all
    the more frustrating.

    So I'm kindly asking that these changes be backed out for 1.0, if
    possible, and maintained in a separate patch or branch to be worked on
    and included at some later date. My reasoning is fairly simple - I'd
    rather have 1 known bug that is annoying yet tolerable (non-incremental
    loads) rather than a plethora of unknown bugs that basically render the
    product useless for me and others. I'd rather let this bug get pushed to
    post 1.0 than to have all of these new ones open so late in the
    development cycle, especially since these are potentially harder to
    identify and fix too.

    Alternately, a "magic bullet" bug-fix would also be appreciated. Martin,
    I would very much like to hear your opinion on this. We could delay 1.0
    until the incremental loader is working 100%, but I personally would
    rather not do that.

    Cheers,
    Dom





    This archive was generated by hypermail 2.1.4 : Thu Feb 14 2002 - 14:26:32 GMT