Linux: coredump in try_compile might break build

While working on the packaging of dynamorio for debian #5153, I stumbled upon the following issue when building dynamorio in a minimal CI system. The reason is a unlucky combination of directory names of dynamorio (top level `core`) and the default coredump file name.

- DynamoRIO contains a directory called `core`
- The default coredump filename is `core`
- During the CMake feature probing (`try_compile`), a coredump happens which creates a file `core`, making it impossible to create the `core` directory later on

Citing `man 5 core`

```
Naming of core dump files
By default, a core dump file is named core, but the
/proc/sys/kernel/core_pattern file (since Linux 2.6 and 2.4.21)
can be set to define a template that is used to name core dump
files. The template can contain % specifiers which are
substituted by the following values when a core file is created:
```

To validate, please check the following files:

```bash
cat /proc/sys/kernel/core_uses_pid
# 0 is problematic
cat /proc/sys/kernel/core_pattern
# core is problematic
```
As a mitigation, enable `core_uses_pid` via `sysctl.d` or simply install `systemd-coredump`.

**/etc/sysctl.d/90-coredump.conf**

```
kernel.core_uses_pid = 1
```

Error pattern:

```
-- Configuring done
CMake Error: Cannot open file for write: /build/core/dynamorioloc.cmake.tmpcf7e2
CMake Error: : System Error: Not a directory
CMake Error in core/CMakeLists.txt:
Could not open file for write in copy operation
/build/core/dynamorioloc.cmake
[...]
```

This is especially problematic when building in containers, as there is no `coredump` namespace in the Linux kernel and by that the settings can only be configured system-wide. Also tests with `ulimit -c 0` inside the container did not solve the issue.

I hope, documenting this as an issue helps others to find the root cause (which is absolutely non trivial) and just apply the mitigations.