From: Steven Sharp (sharp@cadence.com)
Date: Wed Feb 16 2000 - 11:41:16 PST
Someone from the PLI group has suggested that we remove the built-in rule
that an attribute on a module instantiation overrides the same attribute
attached to the module declaration. Their argument makes sense.
For everything else, we simply specify where the attributes are attached
and leave it up to the tool to interpret them. For example, someone can
attach a default value to a module and let the tool interpret that as
the default value for any object under that scope. In the same way, a
tool can interpret an attribute on a module instantiation as overriding
an attribute on the module declaration. However, if some tool needs a
different interpretation for this situation, it can be written to do so.
If we build this rule in, the tool no longer has the possibility of an
alternate rule for some purpose. I think that leaving the tool this
flexibility is more consistent with everything else about attributes.
Apparently the PLI access to attributes can distinguish between attributes
attached to instantiations and declarations. This means that the current
PLI standard already has the necessary flexibility.
Does anyone have any objections or other opinions on this?
This archive was generated by hypermail 2.1.4
: Mon Jul 08 2002 - 12:54:12 PDT
and
sponsored by Boyd Technology, Inc.