Repository navigation
chore(sync): merge upstream through 143bf5f - #111
Merged
Merged
Conversation
…dev#5308) OpenSSH for Windows hands the SSH command argument to the login shell wrapped in double quotes, which expanded the `$`/`$(...)` in Herdr's POSIX discovery and API bridge scripts before `/bin/sh` ran them. A saved machine whose default shell is WSL bash is detected as POSIX, so `herdr --machine` then failed with `remote binary discovery failed`. Deliver the discovery script on stdin (`/bin/sh -s`) and base64-encode the cached POSIX API bridge command so the whole script, including any executable path, stays out of the command argument. The decoded bridge reads the API payload from fd 3 so decoding the script cannot consume it. refs herdrdev#5132
…page (herdrdev#5303) * fix: stop server crash when a cursor move carries a link onto a full page refs herdrdev#5296 * fix: keep cursor link when a style page split moves the cursor refs herdrdev#5296
…5236) * fix: keep the macos server running when its terminal quits refs herdrdev#4920 * test: cover the macos server leaving its launchd job refs herdrdev#4920 * test: wait for the launcher to exit before checking the detached server refs herdrdev#4920
) A window drag can momentarily make the terminal report a zero-sized grid. `terminal_grid_size()` turns that into an error, and the resize watcher treated any error as a dead terminal: it detached and exited the client, which ends the session on a remote host. Keep the last known geometry and the pending resize signal, then resample. Real terminal failures still detach, and startup still rejects a zero-sized grid. This prevents a zero-sized-grid-induced detach. Whether it fully explains the reported Git Bash over SSH session loss is unconfirmed: the pasted ssh log shows a normal session teardown, and the report's leftover mouse reporting is not reproduced here. refs herdrdev#4105
Owner
Author
|
I don't think this PR caused it:
I've re-run the job once. If it fails the same way again, I'll treat it as a real regression and look at herdrdev#5308's bridge change ( Generated by Claude Code |
cameronsjo
marked this pull request as ready for review
October 19, 2026 09:14
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
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.
Before: the fork was 4 commits behind herdrdev/herdr master.
After: the fork includes upstream through
143bf5f:How:
git merge --no-ff upstream/mastermerged cleanly with no conflicts. Upstream changed nothing under.github/,scripts/,build.rsorcrates/*/build.rs. Only two of the changed files are also fork-modified,src/client/tests/mod.rsandsrc/remote/attach.rs, and upstream only added code there. The ledger line is indocs/fork/CHANGES.md.Local checks:
ZIG=/bin/true cargo clippy --all-targets --locked -- -D warnings: clean.cargo fmt --all --check: clean.Generated by Claude Code