release 0.7

* fix #148

* add example for dynamic communicator

* changed generic functions for group parameters in CommunicatorHandle to use trait-objects to make them object-safe

* changed compare function in communicator traits to work with trait-objects

* sealed CommunicatorHandle enum to hide it from public API

* added sealed trait AsHandle to obtain comm handles from communicators

* use CommunicatorHandle in process instead of generic communicator.
 - changed constructors for Process type
 - changed AnyProcess analogous
 - implemented Communicator for Process, such that AsCommunicator can be implemented without violating borrow checker
 - changed functions in Communicator trait that return processes to be object-safe

* extracted sealed module into separate file

* remove obsolete syntax

Attributes now hide the keys from most user code, and are easy to
implement for new types. We place set_attr and get_attr in a new trait
with blanket implementation to support object safety with &dyn
Communicator.

Refactor attributes and add support for setting

* MSRV: add Cargo.lock and cargo update in CI

We frequently have CI failures because transitive dependencies have
releases that require new MSRV. The Rust project now suggests committing
Cargo.lock for reproducibility and to enable testing MSRV.

https://blog.rust-lang.org/2023/08/29/committing-lockfiles.html
https://doc.rust-lang.org/cargo/faq.html#why-have-cargolock-in-version-control
https://doc.rust-lang.org/cargo/guide/continuous-integration.html#verifying-latest-dependencies

Perhaps this can be removed if this Pre-RFC is implemented.

https://internals.rust-lang.org/t/pre-rfc-msrv-aware-resolver/19871

* doc: scrape examples on docs.rs

* ci: allow oversubscription for Open-MPI 5.x

* docs: fix homepage example indentation

* docs: add ci linting of code comments in nightly

* cargo fmt code in doc comments

---------

Co-authored-by: Jed Brown <jed@jedbrown.org>

* Bug/confusion introduced here: https://github.com/rsmpi/rsmpi/pull/157/files#r1462677040
* Process now holds AnyProcess and returns reference to it in as_communicator()
* Fix denied warning in examples/structured.rs and make more strict
* AnyProcess is now Copy and Clone, like Process
* Add Process::is_self() for common pattern

Fix process.as_communicator().rank() and examples/structured

Bumps [shlex](https://github.com/comex/rust-shlex) from 1.2.0 to 1.3.0.
- [Changelog](https://github.com/comex/rust-shlex/blob/master/CHANGELOG.md)
- [Commits](https://github.com/comex/rust-shlex/commits)

---
updated-dependencies:
- dependency-name: shlex
  dependency-type: indirect
...

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

Linting for edition=2021 and cargo upgrade

* Quote from the standard about erroneous use of ready send.

* Add loud warning in `examples/immediate.rs` where ready send is
used (sending to self) since it appears people have copied this example
into uses in which ready send is invalid.

Arguably, ready send should become unsafe because it's so risky to use
and unlikely to be caught by the MPI implementation's run-time checking.
I.e., detection of misuse will be non-deterministic and can easily be
very unlikely in practice.

Ready-mode sends requirements are difficult to ensure without heavy
synchronization (that would outweight the possible performance gains).
The ordering semantics are subject to race conditions and usually not
detectable by the implementation, thus erroneous use may not produce an
MPI error as would be possible for a quality implementation of most
other MPI routines.

Ready mode is unsafe