P1076 January 16, 2014 Meeting Minutes

Attendees:

• CliffordWalinsky * PeterFlake * RyanHinton * JimLewis

Agenda:

Action Item Review

• Cliff: 276: '_' at the end of hex and octal reads
  • Current implementation does not read training '_'
  • Will write up proposal
• Cliff: 113: IR 2075/2092 impacts closely related array type conversions
  • Review with John Ries, was a change made in 2008?
• Ryan: 292 Configure Architecture of Direct Entity Instantiation.
  • Have initial proposal at ConfigureDirectInstantiation
• Need Owner
  • 207: Stop in environment package. Return values ambigous as to what can be done with them.
    • Reserve a range for the LRM uses?
    • Place requirement on implementation to make the return value available some how.
  • 139: PSL_Declaration should not be in Declaration Find Owner
• Steven Dovich: 218: Review and determine if it will be relevant given P1735 changes
• Steven Dovich: 149: IP protect
• Steven: 125: vhpiCbValueChange trigger event specification has redundancy (talk to Peter and see if it was done).
• Done Jim: 188: Explicitly declared terminals and quantities not defined in LRM AI Jim Forward to 1076.1 working group
• Done Jim: Annotated VHDL-2008. See Private Documents
• Peter: 104: String representation not correct for enums which are extended identifiers
  • Peter did write up in bugzilla will migrate to a proposal.
• David Bishop
  • 172 (lost) : Fixed point to_string error
  • 164 (lost) : Argument order: to_integer & to_real in float_generic_pkg
  • 161 (lost): Nit: Function declaration format
  • 160 (lost): ---
  • 158 (lost): logic operations and result sizing: fixed vs. float
  • 127: Read for Float

Review Bugzilla Issues

• Jim 100 : for 1076.1: DOMAIN_TYPE missing in 3.1.1.1
• 95- marked as invalid
  • Chucks concerns are that: A xnor B xnor C /= xnor(A,B,C).
  • And for xnor reduction operations, xnor A(2 downto 0) = not xor A(2 downto 0) and is not the same as A(2) xnor A(1) xnor A(0).
  • If one interprets xnor A(2 downto 0) as a 3 input xnor gate this exactly matches how hardware and Verilog works, and hence, the VHDL-2008 impementation is ok.
• 83 Marked as fixed: LRM was fixed as described in Peter's last paragraph.
• Next Item to review 66 or 297 or larger

Review and Approve Meeting Minutes:

• Motion: Ryan 2nd: Peter

Next Meeting: Thursday January 30, 2014, 8 am Pacific

Previous Meeting: Thursday January 2, 2014