Chris Hathhorn

It looks like our install/build instructions are out-of-date. Can you try this: ``` $ sudo apt-get install bison clang++-6.0 clang-6.0 cmake make coreutils \ diffutils flex libboost-test-dev libffi-dev libgmp-dev libjemalloc-dev...

Yeah, oops: ``` $ export PATH=$PATH:$(pwd)/.build/k/k-distribution/target/release/k/bin ``` Where `pwd` is the `c-semantics` repository root.

There's some discussion of this here: https://stackoverflow.com/questions/45350528/strncpyd-s-0-with-one-past-pointer I'd say it's probably well-defined -- a pointer to one past an object, at least, is usually considered perfectly valid. But there's clearly...

Hi, thanks a lot for the bug report! Could you include some output from the build (or other symptoms) so I know when I'm reproducing the issue you're seeing?

Can you try building with `LD=kcc`?

The build system is currently in a state of flux and the `INSTALL.md` has unfortunately gotten out of date again. The Jenkinsfile/Dockerfile are probably the best reference for how to...

We should move the fatpack stuff to Match instead of testing it (insofar as it is currently tested) twice.

`kcc` ran out of memory when trying to build `wrk` on your computer? What was the exact message? Have you tried maybe increasing the stack size limit (`ulimit -s`)?

Yes, we've recently been trying to add more support for gnuc builtins, so such a test suite would be very helpful. We just `#define __builtin_constant_p(x) 0` in a header, so...

Also, I think there might be a mistake in the list of tests `kcc` succeeds on -- I wouldn't expect the `*_overflow` or `__sync_bool_compare_and_swap` to work either, although maybe `kcc`...