When notified of process exit, we can't catch the ptrace exit event if
we're not still attached. Trying to do so will instead reap the final
exit status, which will cause the Process to fail to do so.

Disable forkproxy by default

* Disable forkproxy / O(n)-waitpid-workaround by default
* Fix a shutdown bug in thread_ptrace that this exposed
    
Closes #1261

Not clear yet how best to handle these; avoid using them until we do.
Some options

* Avoid ever wrapping them; ideal but maybe difficult.
* Just go ahead and wrap them. bindgen's static assertions *should*
  catch if any structs are a different size than expected. They won't
  catch if fields are reordered, but that's probably unlikely.
* Wrap them as opaque types. This would also protect us from reordering,
  though we'd then need to create C accessors for those fields.
* Dynamically generate bindings instead of checking them in. This would
  guarantee that the types match the types on the machine where it's
  built. It wouldn't help with distributed binaries, though.

Don't ignore unknown topology attributes

Bumps [serde](https://github.com/serde-rs/serde) from 1.0.125 to 1.0.126.
- [Release notes](https://github.com/serde-rs/serde/releases)
- [Commits](serde-rs/serde@v1.0.125...v1.0.126)

Signed-off-by: dependabot[bot] <support@github.com>

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Add Rust bindings for Host

* Create a Rust type `SimulationTime` to help avoid mixing with other time types.
* Add a Host struct in Rust to wrap host_*, and expose name and IP for logging
* Add methods to Worker to get current host, sim time, and thread-id

Progress on #1301

Renamed topology 'ip' to 'ip_address', and removed the 'asn' and 'packet_loss' node attributes

* Copy file names instead of holding a reference when we also need to
  get a mutable reference. Holding an immutable + mutable reference
  simultaneously would break Rust's soundness contract, and currently
  results in a run-time panic in the C wrappers.
* Fix missing NULL checks for return values of process_getReadablePtr
  and process_getWriteablePtr.
* Remove unnecessary NULL checks for user pointers. The MemoryManager
  functions will themselves return NULL if the user pointer isn't
  accessible.

Fix fstatat

* TypedPluginPtr: make slice method generic over RangeBounds
* MemoryManager.read_ptr: let caller slice instead of taking offset
* MemoryManager: generalize read_ptr -> readv_ptr
* Split MemoryReader.copy to read_some and read_exact
* MemoryReader: split as_ref into ref_some and ref_exact
* Reimplement process_getReadableString using new Reader API
* Add process_readString
* fileat syscalls: fix memory access bugs
* Add test for fstatat

Fixes #1332

- Rename vertex to node, since gml calls them nodes
- Use host instead of node where appropriate
- Use network node instead of node for clarity
- Replace some use of topology with graph or network graph

Update docs for 2.x

Includes:

- more details about what shadow is and why it's important on the top-level readme
- re-ordered the ToC in docs/readme to better organize contents into sections
- added line breaks to long lines across all docs
- more consistent use of terms like vertices->nodes, topology->graph
- separate tgen example from shadow log parsing

and more...

Bumps [vsprintf](https://github.com/sporksmith/vsprintf) from `ad76219` to `1269b19`.
- [Release notes](https://github.com/sporksmith/vsprintf/releases)
- [Commits](sporksmith/vsprintf@ad76219...1269b19)

---
updated-dependencies:
- dependency-name: vsprintf
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>

…printf-1269b19

Bump vsprintf from `ad76219` to `1269b19` in /src/main

Formatting changes to docs

Minor updates to top-level readme

Final merge main