sim: support checkpointing std::set<std::string>'s

Information
Submitter:	Curtis Dunham
Repository:	gem5
Branch:	default
Bugs:
Depends On:
Reviewers
Groups:	Default
People:

Description

The next patch, RB2931 "tag-based checkpoint versioning" uses sets of strings to represent checkpoint versions. This patch adds support for sets containing strings in the checkpointing infrastructure.

Testing Done

Issue Summary

Description	From	Last Updated	Status
I can we are doing this separate handling of the first element for vector and list as well. Do we ...	Nilay Vaish	July 13, 2015, 10:29 a.m.	Open
Please explain the Pixel disagreement more.	Nilay Vaish	July 13, 2015, 10:29 a.m.	Open

src/sim/serialize.cc (Diff revision 1)

I can we are doing this separate handling of the first element for vector and list as well. Do we really need it? Would we some problem in creating tokens?

Show all issues

Curtis Dunham July 13, 2015, 1:18 p.m. (July 13, 2015, 1:18 p.m.)

The checkpoint upgrader adds a bunch of superfluous whitespace as well and everything seems reasonably robust to that. So I think the answer is no, it's not really needed, but that's the way all the other arrayParamOut methods are written.

src/sim/serialize.cc (Diff revision 1)

Please explain the Pixel disagreement more.

Show all issues

Curtis Dunham July 13, 2015, 1:18 p.m. (July 13, 2015, 1:18 p.m.)

Pixel doesn't have an operator<. If we added one, we could do a full template instantiation with set<> using the existing infra by adding to the INSTANTIATE_PARAM_TEMPLATES macro. It didn't seem that important when we only need those two templates to be instantiated - arrayParam{In,Out} for string sets.

You have a pending review.

Review Board 2.0.15

This change has been marked as submitted.

Screenshots

Files

Issue Summary

Status: Closed (submitted)