slicc: isinstance bugfix

Information
Submitter:	Derek Hower
Repository:	gem5
Branch:	default
Bugs:
Depends On:
Reviewers
Groups:	Default
People:

Description

Changeset 9466:3253995768ef
---------------------------
slicc: isinstance bugfix

Testing Done

Don't you think this change will allow for creation of two symbols
with same identifier but different types? Note that even C and C++
do not allow "Address Address" kind of declarations.

Derek Hower Jan. 14, 2013, 4:14 a.m. (Jan. 14, 2013, 4:14 a.m.)

I don't think so. Code in SymbolTable will prevent two variables with the same id.

Nilay Vaish Jan. 17, 2013, 2:35 a.m. (Jan. 17, 2013, 2:35 a.m.)

Seems fine. But do we really want statements like "Address Address;" ?

Derek Hower Jan. 17, 2013, 3:05 a.m. (Jan. 17, 2013, 3:05 a.m.)

Historically, going back to GEMS, there are lots of protocols that do just that. SLICC uses the word 'address' pretty frequently (it is built in to actions, for example) to represent the block address it is working on. Because of that, it seems pretty natural that you would want to label the corresponding entity in messages, TBEs, etc. with the same name. Because the type is also called Address, 'Address Address' is unavoidable unless you force all the protocols to start using LineAddress or something else for the IDs.

You have a pending review.

Review Board 2.0.15

This change has been marked as submitted.

Screenshots

Files

Status: Closed (submitted)