SV-EC Committee Meeting Monday August 14 2006 11:00am - 1:00pm PST [Minutes distributed for approval at next (sv-ec) committee meeting] (1) Day (4) (0) Month (8) (0) Year (6) --------- Attendees ---------- (-) Arturo Salz (-) Cliff Cummings (A) Dave Rich (A) Francoise Martinolle (-) Mehdi Mohtashemi (A) Neil Korpusik (A) Ray Ryan (-) Surrendra Dudani (A) Gordon Vreugdenhil (A) Steven Sharp (-) Phil Moorby (-) Doug Warmke (A) Stu Sutherland (-) Heath Chambers NOTE: Based on new p1800 PAR, meeting attendance/voting rights is reset from today's meeting. ** Minutes taken by Neil Korpusik ////////////////// Aug 14, 2006 ///////////////////////// Agenda: Mehdi wasn't able to make it to the meeting. Neil ran the meeting. 1. IEEE patent policy ref: http://standards.ieee.org/board/pat/pat-slideset.ppt Move: Dave - Assume that the patent policy was read Second: Steven Abstain: none Opposed: none passed 2. Review Meeting minutes of July 31st, 2006 http://www.eda.org/sv-ec/Minutes/SV-EC_Meeting_July_31_2006_Minutes.txt No discussion. Move: Dave - Approve the minutes of Aug 14, 2006 Second: Gord Abstain: none Opposed: none passed 3. Update on IEEE 1800 PAR The latest 1800 PAR, was approved on July 28th, 2006: http://standards.ieee.org/board/nes/projects/1800.pdf Summary of the last P1800 meeting. - VHDL - Currently looking for a user champion - Committee on interoperability - the svbc suggested this - Editor - will be going out for bid (Accellera will assist in this process). - The existing sub-groups will remain "as is" (svac, svbc, svcc, svec) - The champions group will start meeting again. - All of the errata that have been passed by the committees will go through the champions. - Nominations for officers have been solicited. - Voting rights: have been reset as of this Aug 14th meeting. We will resume the "last 2 out of 3" rule after the first 3 meetings. 4. Action items review Action items: July 31st, 2006 ----------------------------- [DONE] AI: 1490 Gord - send an example 2-dimension of array and how to do the reduction method to Neil. [Done] Neil - add an example to clarify how to deal with dimensions. how to write a reduction operation on an array of array of int. - reverse() - take out (unpacked or packed) from current LRM text. - merge 1490 into 978 - add text and example for specifying width of result (e.g. using with clause) AI: 1437 Dave - update 1437, also include a proposal From Previous meetings: AI: 890 Cliff and all - send email on the reflector to get the discussion started AI: 1330 Gord - new language proposal to handle the port connection issues. AI: 1330 Dave - examples via email on the issue with ports in interfaces new proposal was uploaded - not sure what type of examples needed. 5. Name Resolution sub-group update Gord is the owner from sv-bc. There was no discussion on this topic. 6. Review mantis item list. 978 5.15 array methods, minor errata (Shalom/Neil) [see 1270] Gord - unique() - the new text about removing elements could be confusing - Ordering doesn't apply to associative arrays FM - do these methods apply to multi-dimmensional arrays? Neil - array of arrays is how we described it at the last meeting FM - then just one dimension is affected Gord - you could define a with-expression for char, for example - the with-expression is optional if relational operators are defined for the element type. - not discussed in lrm how to use it but there is a good enum example on page 3 of the latest proposal for mantis 978 that makes use of it Steven - the original reason for having these methods was for constraints Ray - the problem here is that they are methods which seems to defeat the whole purpose (for constraints). Steve - casting is a cleaner way to do handle width extension of the result of calling a reduction method (as opposed to adding 0) Neil - This mantis proposal is not ready for a vote yet. AI/Neil - Copy info on with-expression to sorting section (see locator methods) - Update find, find_index, unique, unique_index -- update all 4 of these in the same way, as far as the ordering of the results are concerned. "The ordering of returned elements is unrelated to the ordering of the original array." - change - "syntax error" to "Shall be illegal" - change Gord's example to not require casting and then use a cast in last example instead. Mention "overflow" in the text describing the need for the cast. 1038 array ordering methods (Shalom) Move: Steven - mantis item 1038 is covered by mantis item 978 Second: Dave Abstain: none Opposed: none Passsed AI/Neil - close mantis item 1038. Add a bugnote explaining why. 1490 Array ordering methods on associative arrays (Dave) Dave - added a bugnote that this is now covered by mantis 978 - we agreed to leave it open until 978 passes 1437 04-26-06 Prototype syntax for array locator methods in 5.15.1 is not legal(Dave) (similar to 0082) No proposal yet. 1457 Problems with Associative arrays and wildcard index type(Dave) Dave - updated it with the previous feedback. Gord - Change text slightly. "computing the minimal length and using that representation for the value" Both of the following will match. 8 bit 32 bit Dave - there are existing rules for invalid index values for arrays - another part of this section talks about what to return if the index value is invalid (no-op for write, return default for read) Gord - 4-state integral index value... is what it refers to - bullet 2 - pre-elaboration error bullet 3 - runtime error Dave - there is no such thing as a dynamically typed expression Steven - handles are not integral types (can't add something to it). - compilation - before execution in 1364 Move: Dave - Approve the proposal for mantis 1457 Second: Gord Abstain: none Opposed: none passed Gord - Errors to be caught before simulation versus runtime. We need better terminology in the LRM for these types of errors. Need to look at all of these in the LRM. The same terminology should be used Dave - this should be done after the LRM merge. 976 5.9-5.10 clarifications (Shalom, DaveR) Gord - doesn't first item in 5.9.5 conflict with 1457? Dave - 1457 was just for wildcard indices Steven - 5.10.8 "If the index type is an integral value that is smaller than the size of the corresponding index..." - what is the "corresponding index"? Gord - ix in example is an output need a term to say what ix is (the actual argument) "If the argument has an integral type that is smaller than the size of the corresponding array index type..." Move: Dave - Approve the proposal for mantis 976 Second: Steven Abstain: none Opposed: none passed 1517 07-03-06 Array reduction methods used in constraints (Dave) Dave - wants a proposal from Arturo - prefers to change the rules for randomize() instead of making the reduction "operators" an operator. 1330 Missing BNF for parameterized virtual interfaces (Dave) [1324 is duplicate -- can be closed if above approved] Neil - new text should be blue. FM - "... shall include any actual ..." Dave - same as module instances - can declare a different virtual interface type Assignments are legal or illegal. The declarations are always legal. Steven - first one in example no interface instantiation in this design neither p16 nor p32 can be assigned to this declaration. Gord - typo: "and is also pert" --> "which is also part" Fm - shouldn't these rules be part of the section on assignment compatibility rules? - parameterized classes - compatibility is in that section. Dave - there would be some circular references if we did that. Assignment compatibilty rules are not all in that section. real vs. integer is in the section on real (not in that same section). - this section has implicit rules. - implicit compatibility rules are scattered around the LRM Gord - parameterized classes - universe of classes different from what is being discussed here. - virtual interfaces - a weird use of type prefers to keep it all together - easier for people to follow Fm - we could create a typedef of a virtual interface - we have forward references already in the LRM. - concern - won't have a complete spec in one section AI/Dave - will upload a new proposal. 6. Next meeting: Proposal: August 28th, 2006 The SV-BC cancelled their meeting for August 28th due to conflicts for several people. Stu - can't make the 28th ========== ADDITION TO the minutes, not part of regular ========= ========== meeting discussions, informational purposes ========= ========== Action Items updated based on August 14th, 2006 ==== Action items: August 14, 2006 ------------------------- AI: 978 Neil - Copy info on with-expression to sorting section (see locator methods) - Update find, find_index, unique, unique_index -- update all 4 of these in the same way, as far as the ordering of the results are concerned. "The ordering of returned elements is unrelated to the ordering of the original array." - change - "syntax error" to "Shall be illegal" - change Gord's example to not require casting and then use a cast in last example instead. Mention "overflow" in the text describing the need for the cast. AI: 1038 Neil - Close mantis item 1038. Add a bugnote explaining why. AI: 1330 Dave - will upload a new proposal. ==================================================================================