Daniel Schemmel

I implemented the basic idea @davidhewitt proposed for a `tracing` based project (interop with `log` should just work (tm), although it is basically untested). While [my implementation](https://github.com/danielschemmel/tracing-for-pyo3-logging) is quite quick...

@ccadar The holdup is due to us deriving from `llvm::cl::parser` a.k.a. `llvm::cl::parser`. Before LLVM 9, this class was final ([LLVM 8](https://github.com/llvm/llvm-project/blob/19a71f6bdf2dddb10764939e7f0ec2b98dba76c9/llvm/include/llvm/Support/CommandLine.h#L952)), but starting with LLVM 9 [it is not](https://github.com/llvm/llvm-project/blob/0399d5a9682b3cef71c653373e38890c63c4c365/llvm/include/llvm/Support/CommandLine.h#L957). @MartinNowack...

As the CI aborts all pending runs once any part of it fails, it might be interesting to know that I can reproduce the LLVM 8 failure locally, while LLVM...

Thanks, @jbuening that was exactly the hint that I was missing. All the CI failures are now gone, @ccadar!

@ccadar done. I checked the four failing tests from Linux (Z3 only) and all of them are handled correctly. For three cases, KDAlloc *correctly* prints a use-after-free error, while the...

The commit is gone again. I think that splitting up kdalloc into multiple commits is not very useful, as it would just mirror the file structure in this commit, except...

@ccadar I have broken up KDAlloc into 6 commits

@251 split off #1537 and #1538. At least #1538 is required for the CI of this PR to be successful, so it should fail for now.

@ccadar I have created a common file for the nontrivial tests. The trivial tests (empty main) I have left as is, as there is really nothing to gain there.

I have tracked down the CI issue to the specific versions of GCC and Boost when using C++17. [Changing any of these parameters will resolve the issue](https://godbolt.org/z/xsnfcK6qx). Any of the...