perf(network): Reduce loop complexity for transport messages and remove unused static variables

[Caball009]

This PR makes modest performance improvements:

1. Improves loop logic in a couple of places (early breaks and keep track of the current index to avoid restarting inner loops).
2. Removes unused static variables.

See commits for cleaner diffs.

TODO:

• Verify everything works ok.

[stephanmeesters]

I think it would be nice if we can change the signature of NetPacket from class NetPacket : public MemoryPoolObject to struct NetPacket. The memory pooling right now doesn't really get used, especially with your change? By the looks of it this would mostly need changes around NetPacket::ConstructBigCommandPacketList

[Caball009]

I think it would be nice if we can change the signature of NetPacket from class NetPacket : public MemoryPoolObject to struct NetPacket. The memory pooling right now doesn't really get used, especially with your change? By the looks of it this would mostly need changes around NetPacket::ConstructBigCommandPacketList

Perhaps I can just put the relevant logic from Connection::sendNetCommandMsg in a callback and pass that to NetPacket::ConstructBigCommandPacketList.

[Caball009]

I think it would be nice if we can change the signature of NetPacket from class NetPacket : public MemoryPoolObject to struct NetPacket. The memory pooling right now doesn't really get used, especially with your change? By the looks of it this would mostly need changes around NetPacket::ConstructBigCommandPacketList

I think it's best to do this in a separate pull request because it involves another set of changes.

[greptile-apps]

Greptile Summary

This PR makes modest but sound performance improvements across the networking layer: it removes two unreferenced static counters (numPackets, numCommands) from ConnectionManager::doRelay(), and refactors several buffer-scanning loops to add an early break on the first empty slot and to carry a bufferIndex forward across iterations rather than restarting from slot 0 every time.

• Early-break optimisation (ConnectionManager, LANAPI, NAT): adds else { break; } when m_inBuffer[i].length == 0, relying on the invariant that the buffer is always packed contiguously. This invariant holds because doRecv() fills from slot 0 forward (via bufferIndex that starts at 0 each call) and all consumers clear slots sequentially from slot 0, so the first empty slot is always the boundary of unprocessed data.
• bufferIndex carry-forward (Transport::doRecv, Transport::doSend latency path): replaces restarting the inner scan from i=0 for each packet with a single monotonically-advancing index, reducing worst-case work from O(N²) to O(N) across a batch of packets.
• MAX_MESSAGES → ARRAY_SIZE and int → size_t for loop variables: minor modernisation; the existing DEBUG_ASSERTCRASH(bufferIndex < MAX_MESSAGES, ...) (previously commented out) is now active, providing a debug-build warning when the buffer is full.

Confidence Score: 5/5

Safe to merge. The changes are a focused optimisation to loop traversal; all correctness-critical invariants hold.

The early-break optimisation is valid because doRecv() always fills m_inBuffer contiguously from slot 0 using a bufferIndex that resets at the start of each call, and all consumers clear slots sequentially from slot 0, preserving a packed-buffer layout. The bufferIndex carry-forward in doRecv() and the doSend() latency path is correct for the same reason.

No files require special attention. Transport.cpp has the most structural change but is correct for all practical configurations.

Important Files Changed

Filename	Overview
Core/GameEngine/Source/GameNetwork/Transport.cpp	Most complex change: introduces shared bufferIndex in doRecv() and doSend() to avoid restarting inner scans from 0; also activates the previously-commented-out DEBUG_ASSERTCRASH for buffer-full detection. Logic is correct given m_inBuffer is always contiguously packed.
Core/GameEngine/Source/GameNetwork/ConnectionManager.cpp	Removes dead static counters, refactors while-loop-with-manual-increment to for-loops, and adds early break on first empty inBuffer slot.
Core/GameEngine/Source/GameNetwork/LANAPI.cpp	Switches loop bounds from MAX_MESSAGES to ARRAY_SIZE, preserves the !LANbuttonPushed guard, and adds early break. All LAN message types fall through to m_inBuffer[i].length = 0 outside the switch, maintaining the packed-buffer invariant.
Core/GameEngine/Source/GameNetwork/NAT.cpp	Adds early break and switches to ARRAY_SIZE. PROBE and KEEPALIVE packets are always cleared; unrecognised packets are not cleared (pre-existing behaviour) but remain filled and thus never trigger the early break.

_{Reviews (5): Last reviewed commit: "Used '>' and '<=' instead of '==' and '!..." | Re-trigger Greptile}

[xezon]

This pull does a number of things.

[Caball009]

This pull does a number of things.

I can put the loop logic in another PR and put #2866 in here to replace all memory pooled uses of NetPacket (either with a single dynamic allocation or stack allocation if possible).

[xezon]

Whatever makes most sense.

[xezon]

Makes sense as far as I can tell.

[Caball009]

Rebased.

I noticed no issues in my testing. Can be merged as far as I'm concerned.