Cap the reconnect backoff at 60s everywhere - #104
Merged
Merged
Conversation
The 5 min cap is sized for the CLI and desktop clients, which sit through outages unattended for days. An iOS session is temporary by nature: it lives only as long as the app, and the user who started it is typically watching, so a wait of minutes between attempts reads as a hang rather than saving anything. The iOS build keeps a 60s cap; everything else in the policy (doubling series, jitter, foreground wake, rebuild gating) is unchanged. The backoff math takes the cap as a parameter so the unit test covers both values from any host, and the docs and FFI comments state the iOS cap. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The 5 min cap waits too long between attempts for every client, not just iOS: each probe is one bounded connect on an endpoint the client already holds, so the minutes bought almost nothing while delaying the notice of a server's return by that much. Go back to the single 60s cap this had before the reconnect revamp — one cap for desktop, CLI and mobile alike, so the platform split (and the test helper that took the cap as a parameter to cover both) goes away with it. Everything else in the policy is unchanged: the doubling series from 1s, jitter, retrying the first attempt, the network/foreground wake, and the endpoint rebuild after the third failure then at most every 30 minutes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The one-line description still described the client as a proxy only, which has been half the story since forwards landed: a client can run with no proxy listener at all and serve nothing but its declared port forwards. The crate description, the CLI's --help line, the two crate-level docs and the README's opening paragraph now name both. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
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.
The 5 min reconnect cap introduced with the reconnect revamp waits too long
between attempts for every client, not just iOS: each probe is one bounded
connect on an endpoint the client already holds, so the extra minutes bought
almost nothing while delaying the notice of a server's return by that much.
This goes back to the single 60s cap the client had before the revamp —
one value for CLI, desktop and mobile alike — so the
cfg(target_os = "ios")split (and the test helper that took the cap as a parameter to cover both
caps) goes away with it.
Everything else in the policy is unchanged: doubling from 1s, jitter, retrying
the first attempt, the network/foreground wake, and the endpoint rebuild after
the third failure then at most every 30 minutes. Docs, FFI comments and
ios/flextunnel.hfollow the value.Also bumps the workspace to 0.0.80 for the release.
🤖 Generated with Claude Code