Rafael Sá Menezes

@mikhailramalho and @fbrausse here is the output of `clang -c -v -m32 esbmc/src/c2goto/library/fenv.c`, I get the same error for every configuration. ``` Apple clang version 13.1.6 (clang-1316.0.21.2.5) Target: arm-apple-darwin21.5.0 Thread...

> Most importantly: hashing is not injective! There can be (unlikely, but possibly) hash collisions. Replacing a condition solely based on equality of hashes is dangerous. Can we include a...

> That's not what I meant :) do_crc() computes a hash of the type2t or expr2t. Equal do_crc() hashes do not imply equal type2t or expr2t. I see what you...

> ssa_algorithm::run() would seem to disagree by actually only iterating through the SSA-steps :) I could make every runOnX function be abstract, in this way you couldn't invoke ssa_algorithm directly....

> Disclaimer: personal opinion following, no technical reason, just a suggestion regarding API design. The types (goto_functionst and SSA_stepst) the 2 specific kinds of algorithms are operating on, respectively, are...

Ok, from @fbrausse suggestions, I think this should be put on hold. First I will create a PR for: 1. Convert `algorithms` into a template class (where `run` will receive...

I will still work on this PR but the initial results look good: 300s, Master ``` Statistics: 15648 Files correct: 8456 correct true: 5128 correct false: 3328 incorrect: 54 incorrect...

@fbrausse would you mind taking another look? I had to create a new hashing function in order to use `unordered_set`

> Do we really need more globals? I can think of some that may be needed. Especially for logging and invoking processes. However, I am proposing to actually put them...

There is also the dependency injection pattern. Which is probably better for this, and we may be able to remove every global.