Meeting called to order 8:30 February 9, 2004 Attendance: 00111000000000000111110000 \ 21210987654433211221109998 / Month 01000011112021121101022002 \ 92136814691740073628413996 / Day aaaaaaaaaa-aaaaaaaaaaaa*aa Steven Sharp -aaaapaaaaaaaaaaaaaaa-a*aa Karen Pieper aaaaa-a----a--aa-aa--a-$aa Cliff Cummings aaaaaaaaaaaaaaaaaaaaaaa=-a Shalom Bresticker aaaa--a-aaaaaaaaaaaaaaa*aa Stefen Boyd -a------a--aaaaaaaaaaaa*aa Dennis Marsa a-aaaaaaa-aaaaa-aaa-aaa$aa James Markevitch -----aaaa-aa-a-aaaaaaa-=-a Gordon Vreugdenhil ------aaa--aaa-aa-aaaaa$aa Anders Nordstrom -------------------aaa-$a- Ted Elkind aaaaaaaaaaaaaaaaaaaaa-a*a- Brad Pierce aaaaaaaaaa-aaaaa-aaaaaa*a- Charles Dawson a-aaaaa-aa-a---aa-a-aa-$a- Mike McNamara aaaaa--aa-aaaaaaaaaa---*aa Stu Sutherland -aaaaa--a--------------*a- Tom Fitzpatrick -----------------------*aa Elliot Mednick -a--a-a-------aa---------- Don Mills -----aaaa-----a----------- Jay Lawrence --a-aa-------a------------ Mehdi Mohtashemi a--a--aaaa--aa------------ Kurt Baty --------a----------------- David Smith a--aaaa-a----------------- Dennis Brophy aaaaaaaa------------------ Francoise Martinolle --aaaa-------------------- Drew Lynch a---a--------------------- David Roberts ---a---------------------- Ennis Hawk a-a----------------------- Ronald Goodstein --a----------------------- Keith Gover aa------------------------ Jason Woolf a attended p proxy - absent The co-chair directed everone's attention to http://www.verilog.com/IEEEVerilog.html where a link to the IEEE policy is prominent. Review of the minutes of the last ETF meeting (Jan 12, 2004). Shalom moves to approve the minutes. Steven Sharp seconds. No opposed. No abstain. Passes. ------------------------------ Open Action Items: (7/14/03) 16: Shalom (11/18/02) generate proposal for 17, 113, 255. We now have a proposal in an advanced stage. (12/1/03) 22: There is no concensus to a fix, so we will leave this issue alone until there is one. (12/1/03) 33: Mike MacNamera to make a proposal. (11/18/02) Steven and committee to return with an @* proposal for issues 22, 82, 84 Group has agreement on some of the more contentious issues. (11/18/02) Evaluating TBD Errata. The tasks are: Shalom 170 formatting of bnf non-terminals (11/4/02) Steven will proposing a wording to fix 172. It will be a significant rewrite. Partially rewritten through the generate proposal. More work will remain after the generate proposal is ready. (11/18/02) Evaluating TBD Errata. The tasks are: Charles 197 sscanf/"string" incompatibility Should add nulls in the white space description in 17.2.4.3a (11/18/02) 198, 254, 58, 318: Shalom (7/14/03) 204: James volunteered to drive He needs input on what the simulators do. There are some complex issues here dealing with port collapsing and inout. That is covered by issue 54. As to the exact issue in 204, how about adding examples indicating that examples 2, 3, 4 are illegal? James will do that. He has not written this up yet. Shalom suggested that this and 7 other errata be evalutated together. All of the issues are Brad's (filed by him?) He does not want to drive. James will develop a proposal for all of them including 204. 204, 227, 233, 234, 292, 332, 345, 452, 486(?). (7/14/03) 204, 227, 233, 234: James (2/10/03) Issue 237: SV-BC19-41, SV-BC19-42 Dave Roberts and Shalom to fix the proposal expanding the BNF and creating language to support it. Dave Roberts said he'd send something out later today. (7/14/03) 245: Shalom (7/14/03) 247: May be partially covered by generate proposal (10/06/03) 428: Brad and Shalom (11/3/03) Shalom has the action to add an issue for the situation where you need a hierarchical reference that allows an array of scopes. This is a follow on from issue 209. Done. This is issue 500 (11/3/03) Shalom will open an issue on parameters and signing reflecting issues raised in 441. Done. This is issue 499 ------------------------------ James leads a discussion of a group of issues that he is driving, 204, 227, 233, 234, 292, 332, 345, in order to get the sense of the committee, so that he could write formal proposals. Issue 345 closed as a duplicate. Issue 237. Discussion led by David Roberts and Shalom, describing various apparently irrational differences in the semantics of timing checks vs. the semantics of the rest of Verilog. Because timing checks are involved with signoff and occur in much legacy code, there is a consensus that simply changing the semantics is not an option. It is further decided that at least the existing semantics, especially its most problematic aspects, should be more thoroughly documented, and that ideally, so should any extensions that are already supported by the major tools. There is also much support for the idea of introducing a second syntax for timing checks that could be given a more rational semantics without impacting legacy code. Generate. Discussion by Jason and Steven Sharp about some issues raised in e-mail by Brad and Shalom. Some key points are that 1) for each genvar, there may be many implicit localparams of the same name, and it is the latter which are used in constant_expressions 2) hence genvar_expressions indicate those places in the syntax where the name really denotes the genvar itself and not one of the implicit localparams with which it is associated 3) a major design constraint required by the user community was that generate scopes be anonymous wherever possible. It is also agreed that Shalom's defparam example was so unlikely as to be considered 'pathological', so can be excluded in the LRM with special case language. The meeting adjourned a little after 10:30 Pacific time.