Fix warnings from Wall Wextra Wpedantic

[hainest]

No description provided.

[bbiiggppiigg]

If on_breakpoint can be called multiple times, then I assume the memset should be kept?

[hainest]

std::pair isn't trivially constructible, so using memset is UB. Values in indexes are only used immediately after writing them, so the memset was always unneeded.

indexes[cur_index].first = j;
indexes[cur_index].second = k;
bps[j][k]->setData(indexes+cur_index);

The mixture of bracket operator and C-style array-plus-index is...

[bbiiggppiigg]

What about the memset for bp_addrs?
It is an array of Dyninst::Address.
It is defined before use in executeTest.
But it is also checked in on_breakpoint L:107.

As for the mixture of C-style array-plus-index, it seems like setData is expecting a void *, so that seems valid to me.
Unless it should be something like setData(&indexes[cur_index])?

[hainest]

What about the memset for bp_addrs? It is an array of Dyninst::Address. It is defined before use in executeTest. But it is also checked in on_breakpoint L:107.

Dyninst::Address is typedef'd to a primitive type, so that one is fine. I think it's also ok to convert T[][] to void*. I'd have to double-check.

As for the mixture of C-style array-plus-index, it seems like setData is expecting a void *, so that seems valid to me. Unless it should be something like setData(&indexes[cur_index])?

It's valid, but there's no need to write assembly in C++. It would be more clearly written &indexes[cur_index], as you pointed out.