[sv-bc] genblk counting -- known during analysis, or does it depend on elaboration?

12.4.3 of IEEE Std 1364-2005, discusses "External names for unnamed generate blocks", or informally "genblk names".

I believe the counts used in these genblk names are fully known by the end of analysis.  But it's been argued to me that because hierarchical generate block names cannot be determined until elaboration time, then neither can the genblk count.  For example, it could be argued that the following nested generate blocks (conditioned on 0) don't exist, have no implicit name, and don't get counted when naming genblks.

   for (i = P; i < Q; i = i + 1)
     if (0)
        assign out[P] = 0;

I don't agree, but can't find an airtight argument from the LRM.  The phrase "... construct that appears textually ..." is suggestive, but isn't proof.

Is the genblk count known by the end of analysis, or not until elaboration?

-- Brad





-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.