Meeting 10-04-04

Attendance
Steve Bailey
John Ries
David Bishop
Charlie Guy
Eric Marshner

FT-19

Reword to allow reading of global in subprogram declared within the architecture, process, and
blocks.
Discussion on if reading the global signals is an error, erroneous, and allowed.
Steve doesn't like it allowing globals in some subprograms and not other subprograms.
David states that synthesis tools don't support globals effectively.  Some discussion of
Xilinx and support for global resets. 
Steve proposes that we pull out the construction of the sensitivity list automatically.
The process_comb, ... should be advisory and tools can do the checks but it would not be
allowed.
Dave Bishop proposes that we include process_comb but not the other ones.
Eric says that we should do the complete checks on the inputs for sensitivity lists.
Charlie we need to check that latches and flip-flops are being created, should be error not erroneous.

Define Process (ALL) to automatically extract all signal read and insert them into a list.
process_comb implies process(ALL) plus the checks on no internal state and all outputs assigned
signals.  process_latch would have the process(ALL) sensitivity list extract and checks for
level sensitive outputs.  Process_ff has extraction of edge and asynch 

Strike the restrict on globals. 
Proposal will require analysis as elaboration time.
Make sure that combination process is "pure".

FT10B
  Looked at grammar changes. Some comments on 285 and simplifying the grammar.
Could be simplified but would have to hand ambiguity with simple assignments in other
parts of the LRM.


FT10A
 Needs to require last "else" condition.  Eric stated that it was still optional.
 (I looked at the changes  the changes for clause 7.1 at line 175 seem to require
  that the "else" clause be present)