Dave,
Recall that the default (and type) clauses have a recursive
definition over array axes etc.
So that part of the definition would be applicable, and might even
be needed for an unpacked array of scalar bits.
Being an author of one synthesis engine that's giving this an error,
I'm certainly willing to reconsider.
On a related note: we don't allow an assignment pattern to construct
an enum type.
As pointless as it seems to still think enums are strongly typed,
I'm still for requiring that this be done with cast operations,
unless the committee drastically rewrites section 6.19.3.
On 10/18/2010 10:55 AM, Bresticker, Shalom wrote:
Hi,
Dave.
Thanks
for
the reminder. I had forgotten that Mantis.
I
was surprised to find that all 3 simulators I tested allowed
an assignment
pattern assignment to a scalar target, though a synthesis
tool did not.
Shalom
You already reported this as
mantis 2708 last year.
I believe the LRM says that
each set of braces matches
each dimension of the array. In this case there are no
dimentions, so no braces
should be allowed.
However, intentionally or
unintentially, some implementations
have allowed references to a[0:0] to be the same
as a reference to a
scalar a. That needs to be resolved first.
always_comb
a
= '{default:'1};
Intel LAD
DA, Jerusalem, Israel
---------------------------------------------------------------------
Intel Israel (74) Limited
This e-mail and any attachments may contain confidential material
for
the sole use of the intended recipient(s). Any review or
distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
--
This message has been scanned for viruses and
dangerous content by
MailScanner,
and is
believed to be clean.
--
This message has been scanned for viruses and
dangerous content by
MailScanner, and is
believed to be clean.
Received on Wed Oct 20 08:49:31 2010