From: Paul Graham (pgraham@cadence.com)
Date: Fri Oct 19 2001 - 14:07:06 PDT
Precedence: bulk
> When in doubt...
> Accepting more will bring less complaints like,
I can't agree with this. Accepting illegal constructs that no other tools
accept will only tick off your customers. What's the point of coding up
some great design if you have no simulator support for it? On the other
hand, rejecting illegal constructs which all your competitors accept will
also tick off your customers. To an extent, it all comes down to what the
market wants. But then again, this is a standards group.
> Would you accept an attribute like this? This was a question back
> during discussions, but fell out of the final standard.
>
>
> (* my_data = "This is an extremenly long attribute.
> In fact it spans multiple lines and may require something
> to show that this is really legal.
>
> Can I just do this, or must I include escapes "\" to
> prefix newline characters?
>
> " *)
The answer to the question should not depend on the rules for attributes,
but on the rules for verilog strings. The syntax for a verilog string:
string ::= " { Any_ASCII_Characters_except_new_line } "
disallows embedded newlines, so I would reject the above attribute instance.
Paul
This archive was generated by hypermail 2.1.4
: Mon Jul 08 2002 - 12:54:47 PDT
and
sponsored by Boyd Technology, Inc.