Subject: Re: [sv-bc] Updated proposal SV-BC-53 Expand array querying functions
From: Greg Jaxon (Greg.Jaxon@synopsys.com)
Date: Fri Dec 05 2003 - 10:19:27 PST
Dave Rich wrote:
> This proposal incorporates the changes suggested at our last f2f meetings
>
> 1. Only introduce system functions, not methods
> 2. Incorporate Intel's proposal by Matt for $type. Now split into
> $typeof and $typename
> 3. Consolidate type equivalence definition into a single section.
>
> Dave
>
A few typos survived - in 5.8.1(4) the edits collided quite badly - where we say:
4. An user defined type (An typedef for an enum, unpacked struct, or unpacked
union, or a class) is equivalent to itself and variables using that type
within the scope of the type identifier.
Replace by:
4. A user-defined type (i.e., a typedef for either an enum, an unpacked struct,
or an unpacked union), or a class is equivalent to itself among variables
using that type within the scope of the typedef identifier.
I think points 3 & 4 should share most of their phrasing, all differences
between them should be significant. So here is one more stab at saying the thing
on which I think we all agree:
OR Replace 3 & 4 by:
3. A data type constructed anonymously by an *enum*, an unpacked *struct*, or an
unpacked *union* is only equivalent to itself among the entities declared
in the immediately following declaration list.
4. A data type constructed within a *typedef* declaration by an *enum*, an
unpacked *struct*, or an unpacked *union*, or which is declared as a *class*
is equivalent to itself among entities declared using that type
within the scope of the *typedef* or *class* identifier.
I think both constants and variables can be declared, and since "variables" seem
to include struct fields and array elements, I thought the term "entity" would
serve as shorthand for whatever we might call the things that can have a type.
Best regards,
Greg Jaxon
This archive was generated by hypermail 2b28 : Fri Dec 05 2003 - 10:23:01 PST