Re: errata/486: 12.3.3: Implicit nets are not necessarily unsigned

From: Steven Sharp (sharp@cadence.com)
Date: Mon Sep 22 2003 - 11:40:00 PDT

  • Next message: Steven Sharp: "Re: errata/485: Multi-line string literals?"

    Precedence: bulk

    The following reply was made to PR errata/486; it has been noted by GNATS.

    From: Steven Sharp <sharp@cadence.com>
    To: etf-bugs@boyd.com, Brad.Pierce@synopsys.com
    Cc:
    Subject: Re: errata/486: 12.3.3: Implicit nets are not necessarily unsigned
    Date: Mon, 22 Sep 2003 15:27:52 -0400 (EDT)

    >According to 12.3.3, "Implicit nets shall be considered unsigned.",
    >but the comments in the example below it (and common sense) contradict
    >that.
     
    >And I don't know what this sentence means --
    >
    > "Nets connected to ports without and explicit net declaration shall
    > be considered unsigned, unless the port is declared as signed."
     
     Presumably it is talking about a net that was only declared as a port.
     Of course, that is an implicit net, which it just said was always unsigned.
     
     My guess is that it is distinguishing the case of an implicit net that
     is created by a port declaration without an explicit net declaration
     (which gets the signedness of the port declaration), from an implicit
     net that is created by use in an instance port expression or LHS of a
     continuous assignment (which is always a scalar, and presumably unsigned).
     
     Note that the first case of implicit nets was not actually mentioned in
     the 1995 standard, but was documented in the 2001 standard. Perhaps this
     text in 12.3.3 was written before this was changed, so that the term
     "implicit net" didn't include ports without explicit net declarations yet.
     
     Assuming this was the intent, I would suggest reversing these two sentences
     and adding the word "other", as follows:
     
     "Nets connected to ports without an explicit net declaration shall be
     considered unsigned, unless the port is declared as signed. Other implicit
     nets shall be considered unsigned."
     
     Steven Sharp
     sharp@cadence.com
     

    This archive was generated by hypermail 2.1.4 : Mon Sep 22 2003 - 11:42:41 PDT and
    sponsored by Boyd Technology, Inc.