iox2-51 shm long name support

[elfenpiff]

Notes for Reviewer

Do not review before #52 is merged, since it is based on it.

MacOS has the restrictions that only 31 characters are allowed for the shared memory name but iceoryx2 requires much more. The trick is that a file is created to:

• be able to list all shared memorys in mac os (already present)
• store the actual shm name, the file name corresponds to the user given shared memory name

In the file is then the triple stored ($PID$_$TIMESTAMP_SECS$_$TIMESTAMP_NSEC$). In the worst case scenario where PID + SEC is larger than 1.000.000.000 it would hit the max name length of mac os (32 > 31), but to our luck NSEC is always smaller than 1.000.000.000 giving us the extra character we need.

Pre-Review Checklist for the PR Author

1. Add sensible notes for the reviewer
2. PR title is short, expressive and meaningful
3. Relevant issues are linked
4. Every source code file has a copyright header with SPDX-License-Identifier: Apache-2.0 OR MIT
5. Branch follows the naming format (iox2-123-introduce-posix-ipc-example)
6. Commits messages are according to this guideline
   • Commit messages have the issue ID ([#123] Add posix ipc example)
   • Commit author matches Eclipse Contributor Agreement (and ECA is signed)
7. Tests follow the best practice for testing
8. Changelog updated in the unreleased section including API breaking changes
9. Assign PR to reviewer
10. All checks have passed (except task-list-completed)

Checklist for the PR Reviewer

• Commits are properly organized and messages are according to the guideline
• Unit tests have been written for new behavior
• Public API is documented
• PR title describes the changes

Post-review Checklist for the PR Author

1. All open points are addressed and tracked via issues

References

Use either 'Closes #123' or 'Relates to #123' to reference the corresponding issue.

Closes #ISSUE-NUMBER

[elBoberido]

I'm thinking whether it would make sense to keep the old value on Linux as canary check that something is broken?

[elfenpiff]

At this level I would like to have the same tests, behavior, limits etc. and say with confidence that every platform supports at least 50 subscribers.

I completely understand your rational, but when we go this route we may have a wild mixture of verified upper limits for every platform. I would rather go the path that we say 50 subscribers are supported for every platform, but an individual higher limit for a specific platform is maybe working.

Maybe we can introduce platform specific system limit tests where exactly those differences can be verified. What do you think?

[elBoberido]

Indeed. These limits might be a criteria for tier 2 and for tier 1 we could have tests for additional limits.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (5d54039) 77.90% compared to head (a8cb41f) 77.89%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #53      +/-   ##
==========================================
- Coverage   77.90%   77.89%   -0.02%     
==========================================
  Files         172      172              
  Lines       18664    18664              
==========================================
- Hits        14541    14539       -2     
- Misses       4123     4125       +2     

Files	Coverage Δ
iceoryx2-bb/posix/src/condition_variable.rs	75.23% <100.00%> (ø)
iceoryx2-bb/posix/src/thread.rs	55.75% <ø> (ø)
iceoryx2-cal/src/event/process_local.rs	85.82% <100.00%> (ø)

... and 1 file with indirect coverage changes