Re: [sv-ec] fork..join_none/join_any and automatic variables

From: Steven Sharp (sharp@cadence.com)
Date: Thu May 22 2003 - 12:17:39 PDT

  • Next message: drm@xilinx.com: "errata/140: PROPOSAL - Section 4.1.5: Definition of power operator result type"

    Precedence: bulk

    >I don't think 281 was ever formally resolved by the ETF,
    >just that nobody objected to closing it.

    I think that the reason everyone agreed to close it is that everyone
    agreed with our comments saying that the answer was already clear
    and what that answer was.

    >Regardless, if so many people are confused by the semantics of fork-join,
    >maybe we should add a short explanation in the 1364 LRM.

    Both times the issue has come from the same source, so I'm not convinced
    that this indicates widespread confusion.

    The difficulty with adding explanatory text about this specific situation
    is that it might have the reverse effect of convincing people that there
    is something special and unexpected about the situation. They might think
    that the text represents a specific decision about a special case that
    could have just as easily been made the other way. They might not realize
    that it is just an explanation of something that follows naturally and
    automatically from the existing scope rules.

    Mind you, trying to use the existing text to support an explanation of
    the rules shows flaws in the text. It could stand some reworking. But
    if so, it should be a better explanation of when allocation occurs, not
    a special description of one of the cases where it doesn't.

    Steven Sharp
    sharp@cadence.com

    This archive was generated by hypermail 2.1.4 : Thu May 22 2003 - 12:19:52 PDT and
    sponsored by Boyd Technology, Inc.