Performance: Optimize IPC memory allocations with stack buffers

[devin-ai-integration]

Performance: Optimize IPC memory allocations with stack buffers

Summary

This PR optimizes the most critical performance bottleneck in dictationbridge-core: repeated dynamic memory allocations in the IPC layer. Every text insertion and deletion event was triggering new/delete operations, causing unnecessary heap fragmentation and allocation overhead.

Key Changes:

• inproc/ipc.cpp: Replaced dynamic allocations in SendTextInsertedEvent() and SendTextDeletedEvent() with 1KB stack buffers
• Fallback mechanism: Large messages (>1KB) still use heap allocation for safety
• PERFORMANCE_ANALYSIS.md: Comprehensive analysis documenting all performance issues found in the codebase

Performance Impact: Eliminates heap allocations for typical dictation text chunks, reducing memory fragmentation and function call latency in the hot path that processes every dictation event.

Review & Testing Checklist for Human

• Build verification: Confirm project compiles successfully with Visual Studio 2015 and SCons
• Basic functionality test: Verify dictation still works correctly with both Windows Speech Recognition and Dragon NaturallySpeaking
• Stack buffer size validation: Check that 1KB is appropriate for typical dictation text lengths (review existing logs/data if available)
• Memory safety review: Verify no buffer overflow risks exist with the stack allocation approach
• Performance testing: Measure actual performance improvement during active dictation sessions

---

Diagram

%%{ init : { "theme" : "default" }}%%
graph TD
    Master["master/main.cpp<br/>Message Processing"]:::context
    IPC["inproc/ipc.cpp<br/>IPC Layer"]:::major-edit
    Hooks["inproc/wordhooks_impl.cpp<br/>Word Event Hooks"]:::context
    UserHooks["inproc/userhooks.cpp<br/>User Input Hooks"]:::context
    Client["client/main.cpp<br/>Command Client"]:::context
    
    Hooks -->|"SendTextInsertedEvent()"| IPC
    UserHooks -->|"SendTextDeletedEvent()"| IPC
    Client -->|"SendCommand()"| Master
    IPC -->|"WM_COPYDATA"| Master
    
    subgraph Legend
        L1[Major Edit]:::major-edit
        L2[Minor Edit]:::minor-edit  
        L3[Context/No Edit]:::context
    end
    
    classDef major-edit fill:#90EE90
    classDef minor-edit fill:#87CEEB
    classDef context fill:#FFFFFF

Loading

Notes

Risk Assessment: Medium risk due to changes in critical hot path, but logic is straightforward.

Why this optimization matters: The IPC functions are called for every single character/word that gets dictated, making memory allocation overhead significant during active dictation sessions.

Alternative approaches considered: Thread-local buffers, pre-allocated buffer pools, but stack allocation with heap fallback provides the best balance of simplicity and performance.

Session Info:

• Link to Devin run: https://app.devin.ai/sessions/d2b19c3d7bc3494086e0814288d88a62
• Requested by: @pranavlal

Future optimizations identified: The performance analysis report documents 6 additional optimization opportunities across string operations, polling patterns, and system calls that can be addressed in follow-up PRs.

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring