[EDIF] EDIFNetlist.collapseMacroUnisims() to not clobber cell

[eddieh-xlnx]

.. plus optimize expandMacroUnisims().

Currently, the following scenario is not considered:

• Design has OBUFDS cell instantiations
• Only some of those OBUFDS cells are expanded into OBUFDS_DUAL_BUF
• Collapsing causes the OBUFDS_DUAL_BUF cell to be renamed to OBUFDS, without awareness that OBUFDS already exists.
• Now there are two OBUFDS cells being instantiated in the netlist, but only one in the library.

[clavin-xlnx]

I recall you found an issue regarding the attribute being on a port/net would have to be artificially migrated to the instance for the exception to be triggered. Will that case need consideration here?

[eddieh-xlnx]

I recall you found an issue regarding the attribute being on a port/net would have to be artificially migrated to the instance for the exception to be triggered. Will that case need consideration here?

I think you mean #646, along with its derivative PRs? If so, I think this PR is orthogonal: #646 and the like deal with IOStandard properties on the EDIFCellInst (or top-level EDIFNet it is connected to). This PR deals with the EDIFCell definition.

[clavin-xlnx]

I think you mean #646, along with its derivative PRs? If so, I think this PR is orthogonal: #646 and the like deal with IOStandard properties on the EDIFCellInst (or top-level EDIFNet it is connected to). This PR deals with the EDIFCell definition.

Good, thanks for clarifying.