Re: Remider: conference call April 4th 1:30pm Pacific

In light of the agenda, I've uploaded two documents:
1) a revised pdf for Section 5
http://www.eda.org/verilog-ams/htmlpages/public-docs/signals_2.3_draft.pdf

2) a "comparison" pdf showing strikeouts and additions
http://www.eda.org/verilog-ams/htmlpages/public-docs/signals_2.3CMP.pdf

Note: in the "comparison" document, I have removed change-bars
where the only change is in whitespace, such as places where I
inserted tabs to fix alignment (per mantis 1404, etc.)  This
makes it easier to see the changes.


Relevant to these changes are the following Mantis items:
 819 Switch branch syntax clarification in the LRM
 - switch branch has not been added to the BNF, so it remains
   a semantic restriction.

 855 Array vs Port identifier declarations in Section 5.2, Syntax 5-1
 - Syntax 5-1 was completely replaced, so this issue is resolved

 887 Value retension example very confusing
 - value_ret added in 5.3.1.3; do we agree this fixes the issue?

 1404 missing inout declarations
 - resolved.


I have essentially swapped 5.2 and 5.2.1.

I believe I have captured most of the semantic restrictions
that were found in one draft of the updated syntax and repeated
in my e-mail of 28 March.

A.6.10 filters for contributions and conditionals
- see paragraph under the syntax box on page 105

A.6.10 indirect contributions and conditionals
- see new para at bottom of page 107

A.8.2 restrictions for branch_probe and port_probe
- see paragraph immediately above Syntax 5-1 on page 100
  (nature_attribute_identifier can't be "units", use of
  local ports)

A.8.9 nature attribute reference
- see para immediately above Syntax 5-2 on page 103


One restriction I don't know how to word:

NOTE: Semantically restrict nets used in analog_port_reference
  and analog_net_reference to be continuous only.

Probably need something near Syntax 5-1.  But how do I say
that the nets are continuous?  Something about the discipline
declared for the nets?


New issue: in "merged_syntax.pdf" (as found in Mantis 810) ddx()
is not listed as a possible item for indirect_branch_contribution,
whereas it was listed in a previous version (syntax_2.3_draft.pdf
as found in the "current discussion documents" area).


-Geoffrey 




Sri Chandra wrote:
> > Agenda:
> > * We will be focussing on the Section 5, updates that Geoffrey has made
> > to reflect LRM2.3