2016_MeetingApril7 < P1076

P1076 Web>Vhdl2019MeetingMinutes>2016_MeetingApril7 (2020-02-17, JimLewis) (raw view)EditAttach
<!--
Set ALLOWTOPICCHANGE = P1076AdminGroup
-->
---+!! P1076 April 07, 2016 Meeting Minutes
---++!! Attendees:

   * Rob Gaddi, Lieven Lemiengre, Jim Lewis, Brent Hayhoe, Ernst Christen, Patrick Lehmann

---++!! Agenda:

%TOC%

---++ Meeting Discussion
   * What's Next - see [[PrivateDocuments]] file: _summary_vhdl_requirements_priority.xlsx
   * Restart at line 60
   * 60: Array Type Generics
      * Currently cannot do array type operations on a generic type
      * No way to enforce that the actual is an array type
      * <span style="background-color: transparent;">May overlap with abstract packages</span>
      * May overlap with anonymous types discussion
      * Should we be able to restricted a type to a scalar type or an array type.
         * Current sense of the language is that the object is limited to operations on a scalar type
   * 63: Abstract Packages
      * Skipped to here due to overlap with 60
      * Currently can pass functions
      * Ernst: Alternative make checking generic package is to parse a generic package as a template and defer the checking until compiling a generic package instance
      * Lieven would prefer to use an abstract package
   * 61: New Predefined Attributes: 'actual and 'formal
      * For integers want to know the range of the actual that is passed to a subprogram vs having to pass 'low and 'high as parameters
      * Would not always want to do this as it would slow down all subprograms as the 'low and 'high would need to be implicitly passed by the simulator.
      * May need to be reconsidered with the integer extensions
      * Alternative: make range a first class object and have the range explicitly passed in the subprogram
      * Also see: [[RangeOperations]]
      * Could create an explicit range record. Make 'range available to return the record type.
      * solvable?
      * Nice and useful to have but not exactly must have for the level of complexity and/or time required for this proposal
      * Push to next revision
   * 62: Conditional Compilation
      * If limited to only predefined names is it enough?
         * Tool type (sim, synth), vendor name, tool name, tool revision, language revision
      * Could use tool directives similar to encryption
      * Lexing time vs. Parsing time?
      * Still contentious, need to talk about this again.
   * Next Meeting discuss interfaces/bundles
---++ Review and Approve Meeting Minutes:

Rob, Brent
---++ Next Meeting: Thursday [[2016_MeetingApril14][April 14, 2016]], 11 am Pacific
---+++ Previous Meeting: Thursday [[2016_MeetingMarch31][March 31, 2016]]
Topic revision: r3 - 2020-02-17 - 15:36:16 - JimLewis
P1076