Stagger scoreboard sends to prevent buffer overflows on full servers - #313
Merged
Conversation
Scoreboard layout messages are now spread across multiple server frames instead of being sent to all clients simultaneously. This prevents packet buffer overflows that could crash the server or disconnect legacy clients with small reliable message limits (~1400 bytes). - Periodic scoreboard updates (every 3 seconds) are staggered per-client across the full cycle, so a 32-player server sends ~1 per frame instead of 32 at once. teams_changed still forces immediate update. - Intermission scoreboard sends are spread across 4 frames (~0.4s) instead of bursting all clients in a single frame. - On-demand score key (TAB) switched from reliable to unreliable send, since the periodic refresh catches any dropped packets within 3 seconds. - MAX_SCOREBOARD_SIZE raised from 1024 to 1300 bytes and MAX_PLAYERS_PER_TEAM raised from 8 to 10, using the headroom from reduced burst pressure while staying within legacy packet limits. - Document scoreboard delivery changes and the configurable scoreboard field codes (previously undocumented) in action.md. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Review feedback on the scoreboard stagger commit (ebf1254): 1. DeathmatchScoreboard (TAB key) was switched from reliable to unreliable with the rationale that "the periodic refresh catches drops". But the periodic refresh runs only once per 3 seconds, which is too slow to mask a dropped TAB response — a user pressing TAB on a lossy connection saw a blank scoreboard for up to 3 seconds. Single-client unicasts also don't contribute to the full-server burst problem the stagger was reducing. Revert that one site to reliable. 2. The per-client stagger formula used `i % (3 * HZ)` for slot assignment. At default HZ=10 the cycle is 30 frames; on servers with maxclients > 30 (common: 32-64), high-index clients collided modulo-30 with low-index clients while some frame slots received two updates and others stayed idle. Expand the cycle to max(3*HZ, maxclients) so every client gets a unique slot. Common case (maxclients <= 30) is unchanged; large servers trade slightly slower refresh (maxclients/HZ seconds) for perfect distribution. Note on the buffer-size review finding (MAX_SCOREBOARD_SIZE not propagated): verified false — A_NewScoreboardMessage already uses MAX_SCOREBOARD_SIZE for both `string` and `buf` declarations. The reviewer was looking at an older file state.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Scoreboard layout messages are now spread across multiple server frames instead of being sent to all clients simultaneously. This prevents packet buffer overflows that could crash the server or disconnect legacy clients with small reliable message limits (~1400 bytes).