LCS2016_082 < P1076

P1076 Web>VHDL2017>LCS2016_082 (2017-04-02, PatrickLehmann) (raw view)EditAttach
---+!! Language Change Specification for Empty records syntax regularization

<br />%TOC%

| <sticky><b>LCS Number:</b></sticky> | LCS-2016-082 |
| <sticky><b>Version:</b></sticky>  | 2 |
| <sticky><b>Date:</b></sticky>  | 22-Dec-2016 (ver.1), 10-Feb-2017 (ver.2) |
| <sticky><b>Status:</b></sticky>  | |
| <sticky><b>Author:</b></sticky>  | [[Main.RyanHinton][Ryan Hinton]] |
| <sticky><b>Email:</b></sticky>  | %USERINFO{"RyanHinton" format="$emails"}% |
| <sticky><b>Source Doc:</b></sticky> | [[EmptyRecord][Syntax regularization: empty records]] |
| <sticky><b>Summary:</b></sticky>  | Allow record type definitions to have zero elements |

---++ Voting Results: Cast your votes here

Yes:
   1 %USERSIG{RyanHinton - 2016-12-22}% - Version 2
   1 %USERSIG{ThomasPreusser - 2016-12-22}% - Version 2
   1 %USERSIG{FarrellOstler - 2016-12-22}% - Version 1
   1 %USERSIG{BrentHahoe - 2016-12-23}% - Version 1
   1 %USERSIG{JakkoVerhallen - 2016-12-29}% - Version 1
   1 %USERSIG{MartinZabel - 2017-02-17}% - Version 2
   1 %USERSIG{RobGaddi - 2017-01-26}% - Version 1
   1 %USERSIG{JimLewis- 2017-02-06}% - Version 1
   1 %USERSIG{PatrickLehmann - 2017-02-16}% - Version 2

No:
   1 

Abstain:
   1 %USERSIG{KevinJennings - 2016-12-23}% - Version 1...don't think I'll make use of this, but not really opposed if it helps someone else
   1 %USERSIG{LievenLemiengre - 2017-01-27}%
   1 %USERSIG{BrentHahoe - 2017-02-16}% Version 2 - Abstain due to lack of personal time for review.
   1 %USERSIG{MartinThompson- 2017-02-17}% Version 2 - what KevinJ said

---++ Details of Language Change

Changes are shown in %RED%red font%ENDCOLOR%. Deleted text is %RED%<del>crossed out</del>%ENDCOLOR%. Editing comments are in %GREEN%green%ENDCOLOR%.

---+++ LRM 5.3.3 Record types

%GREEN%Modify BNF after first full paragraph, middle of page 51 (pdf 65).%ENDCOLOR%

<sticky><pre>%BLACK%
record_type_definition ::=
    <b>record</b>
        %RED%<del>element_declaration</del>%BLACK%
        { element_declaration }
    <b>end</b> <b>record</b> [ %GRAY%<i>record_type</i>%BLACK%_simple_name ]
</pre></sticky>

%GREEN%(The identical change should be made in the syntax summary in Annex C.)%ENDCOLOR%

---+++ LRM 9.3.5 Qualified expressions

A qualified expression is a basic operation (see 5.1) that is used to explicitly state the type, and possibly the subtype, of an operand that is an expression or an aggregate.

<pre>qualified_expression ::=
     type_mark ' ( expression )
   | type_mark ' aggregate
   %RED%| type_mark ' ( )%ENDCOLOR%
</pre>

%RED%For qualified expressions with an operand, <del>T</del>%ENDCOLOR%the operand shall have the same type as the base type of the type mark. The value of %RED%such%ENDCOLOR% a qualified expression is the value of the operand. The evaluation of %RED%such%ENDCOLOR% a qualified expression evaluates the operand and converts it to the subtype denoted by the type mark.  

%RED%A qualified expression without an operand is used to express a literal for a composite type or subtype with no elements.  The value of such a qualified expression is an empty value of the type or subtype denoted by the type mark.%ENDCOLOR%

-- %USERSIG{RyanHinton - 2016-12-22}%

---+ Comments

Suave uses "type foo is null record" to define empty records. How does a map to an empty record look like? Are empty record compatible?

-- %BUBBLESIG{PatrickLehmann - 2016-12-31}%

*PL:* Suave uses type foo is null record to define empty records.

*MZ:* Requiring null record does not work well with conditional compilation. The record might get empty based on conditions.

-- %BUBBLESIG{MartinZabel - 2017-01-09}%

An alternative for null record is:

<pre>type MyRecord is record
  null;
end record;</pre>

I changed my vote to NO, because I think an empty initialization should be addressed by this proposal too.

<pre>type MyRecord is record
`if DEBUG = "TRUE" then
  debug1 : std_logic;
`end if
end record;

entity E1 is
  port (P : MyRecord);
end entity;

entity E2 is
end entity;

architecture A2 of E2 is
begin
  inst : entity work.E1
    port map (P =&gt; (
`if DEBUG = "TRUE" then
        debug1 =&gt; '1'
`end if
      )
    );
end architecture;
</pre>

If DEBUG is FALSE, then the port map collapses to port map (P !=&gt; () ). Empty aggregation lists are not supported.

-- %BUBBLESIG{PatrickLehmann - 2017-01-09}%

Could we force a record aggregate that may be empty to use a qualified expression? In Patrick's example:

<pre>architecture A2 of E2 is
begin
  inst : entity work.E1
    port map (P =&gt; MyRecord'(
`if DEBUG = "TRUE" then
        debug1 =&gt; '1'
`end if
      )
    );
end architecture;</pre>

-- %BUBBLESIG{RobGaddi - 2017-01-27}%

Version 2 created to use Rob's great idea.

-- %BUBBLESIG{RyanHinton - 2017-02-10}%



What happens for string&#39;(&#39;A&#39;) is then the (&#39;A&#39;) understood as a element of an aggregate?
string&#39;() = &#34;&#34;

I am not saying that there are any issues, but there may be some rough edges

-- %BUBBLESIG{JimLewis - 2017-03-21}%
%COMMENT%
Topic revision: r22 - 2017-04-02 - 15:42:45 - PatrickLehmann
P1076