perf: optimize getInbox usage and shared pointer handling in loops

[dudantas]

Description

This addresses performance issues by optimizing the use of shared pointers and removing redundant calls. Specifically, the getInbox call was removed from extensive loops to prevent unnecessary reference count increments, which were causing CPU overhead. Additionally, shared pointers used in recursive or iteration-heavy functions, such as the ContainerIterator, were changed to const& where applicable, reducing the impact of reference counting on performance.

Fixes this perf report: 23_11_20_18.txt

Behaviour

Actual

When iterating over container items or processing inbox elements in loops, excessive reference count updates for shared pointers caused significant CPU strain, leading to reduced performance under heavy load.

Expected

The changes reduce the number of reference count updates and improve performance, particularly in recursive contexts and large loops involving container operations.

Type of change

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested

The changes were tested by running performance benchmarks during typical server operations to ensure that the CPU usage was reduced in scenarios involving container iterations and recursive pointer accesses.

• Test A: Benchmark with container iteration to measure CPU usage before and after the change.
• Test B: Stress test with simulated load to verify the absence of performance degradation.

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I checked the PR checks reports
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works

[sonarqubecloud]

Quality Gate passed

Issues
4 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud