Re: Verilog-AMS/2005 syntax and keywords

Graham,

You did a good job; it will definitely help to guide the discussions on 
the merger. I do have
a few comments, see below:

> Hello,
> 
> The Verilog-AMS/2005 syntax and keyword sections have been added to 
> the Mantis database (item 0000810). Please have a look at them. 

I've found a small discrepancy between your proposal and the 2.2 LRM.

In section A.2.6 of the merged_syntax document the
analog_function_item_declaration only allows inputs to be defined,
but the 2.2 LRM explicitly allows also outputs and inouts there.

> Below is a list of notes and question about these sections of the LRM.
> 
> 1) AMS was merged into the 2005 document (instead of 2005 being 
> merged into the AMS document). AMS extensions are identified by blue 
text.
> 2) Moved the analog_generate_loop_statement into the generate 
> construct section as it is extension of the 2005 generate construct.
> 3) Array and string parameter declaration syntax has merged into the
> 2005 parameter declaration syntax.
> 4) Keyword "abstol" in section B.2 need to be moved into B.1 as it 
> can be accessed from module declaration.
> 5) Verilog-AMS B.2 and B.3 sections have been collapsed into the B.1
> since 2005 did not have any corresponding 2005 sub-sections. Why 
> where these subsections created in the first place?
> 6) It appears from the document formatting that all of the keywords 
> are to be contained on a single page. To maintain this constraint 
> the font size was substantially reduced when AMS keywords were 
> added. Should the single page format be maintained or restore the 
> original font size and have a 2 page list of keywords?

This font-size is hardly readable for the visually impaired, so I think
the 3 column suggestion of Geoffrey is better.

> 7) Header and footer updated to reflect AMS merger.
> 8) Compiler directives and system task/functions has significant 
> syntax. Why is the syntax for constructs not documented in the syntax 
annex?

Given that there are syntax definitions of at least the compiler 
directives
in the LRM, I find it strange that they would be absent in the syntax
overview. The compiler directives are as much part of the language 
standard
as anything else. In a similar sense the system task/functions should also
be added, if just to be consistent with the 1394-2005 syntax in section
A.7.5.1.

> 
> Regards
> Graham

Marq