Minutes ISAC meeting
May 4 2011

Next meeting: June 8, 2011

Missing bugzillas.
	Somehow, a group of entered Bugzilla issues was lost from the data base. The most likely cause is that when the data base was moved to a new server an older version was restored. By the time the error was discovered all backups had been destroyed.
	As far as we can tell all missing issues dealt with intermediate language versions and all of them have been incorporated into VHDL-2008, so the loss is not as bad as it might have been.
	Action item: Chuck to 1. Check the Accelera reflector to see if any relevant data is available. 2. Prepare a list of non-recovered issues and send that list to the VHDL-2008 reflector. 3. Contact Dave Bishop to make sure that his issues
haven't been lost.

Preliminary Triage of VHDL-2008 open issues:

239 IR2130:Ability to overload the assignment operator: = would be useful.  email

	This is a language enhancement and will be forwarded to VHDL-201X. Several people requested more details use cases
and further clarification on this issue. It was also noted
that the new email is jonathan.bromley@verilab.com
	Action item: Chuck to contact bromley for more information.Chuck to create VHDL-201X Bugzilla product for use by the VASG working group.

240 IR2132: Method to allow functions that return arrays to have knowledge of the array bounds.

	It was pointed out that bugz 239 could be used to accomplish this.
	Action item: Chuck to forward to VHDL-201X.

241 IR2133: Not locally static expression is synthesized but not compiled for simulation.

	Important attachments are missing.
	Action item: Peter to recover attachments (done).

259 attribute declaration with same designator as predefined attribute.

	This was judged to be of medium importance/difficulty.

276 Read termination unclear when underlines are allowed .

	It was generally, but not  universally, agreed that this is a genuine issue.  The consensus is that once the data is exhausted then trailing underscores are not read. This was judged to be of medium importance.

277 Inconsistent rules for vector READS.

The consensus is that the intention of the LRM is for uniform
rules for all vector reads. This is high priority.

278 The logical reduction operators have the worng precedence.

The grammar rules give the correct precedence but the list in
Clause 9.2.1 is incorrect, since it does not distinguish unary
logical operators from binary logical operators. The LRM should explicity state that the precedence is determined by the grammar rules. It was also observed that NOTE 2 is wrong.This issue was given medium priority because the grammar is correct and it is believed that all implementations follow the grammar instead of the list.

281 Discrepancies between float_generic_pkg.vhdl and float_generic_pkg-body.vhdl

This is a simple typo with an obvious solution.

282 Inconsistent capitalization for ieee packages.

This is a low priority issue.

283 ?/= function in float_generic_pkg can give wrong answer

A specific example is needed to determine if there really is an issue. If there is a genuine problem then the issue is high priority.

Action: Chuck (the original submitter) to produce an example.

284 Sizing rules for fixed point reciprocal are wrong

The typo occurs in a non-normative part of the text, so the
priority is low although the solution is trivial.

285 textio write[line,boolean...] is not backward compatible across language versions

This issue is high priority. The question was asked if the 'image attribute is also changed. It appears that 'image always
produces ordinary enumeration names in lower case for all language versions.