Maintain ID_C_expr_simplified

[tautschnig]

The aim is to avoid redundant invocations of the simplifier by storing is_simplified with each expression. This patch is the first part of this, then second half will be added once #25 is merged as there would be conflicting code changes.

[peterschrammel]

@tautschnig, can you explain why this code is commented out?
Otherwise, this looks mergeable.

[tautschnig]

I have cleaned up the code that really was for debugging purposes only.

[forejtv]

@tautschnig can you please rebase to fix the conflict and also see if all tests pass?

[tautschnig]

Rebased, all checks passing. As with #35, there is another question about performance, however.

[forejtv]

@kroening can you let me know if this is good to go or whom I should ask to evaluate the performance?

[tautschnig]

It would be nice to have some continuous performance evaluations set up. As a first step, the scripting outlined in https://github.com/diffblue/cprover-sv-comp/blob/master/README.md could be picked up. That is, one would likely want a two-fold evaluation: monitor SV-COMP overall results, but also build and run with profiling data as 1) not every change results in immediate overall verification improvements and 2) overall improvement on any specific set of benchmarks (such as SV-COMP) is not necessarily representative.

[hannes-steffenhagen-diffblue]

@tautschnig This has been stale for a while. I'm not sure if this still makes sense to do now 5 years later, if you want to revive this feel free to rebase and reopen, but closing for now.