f173382fd3
Some checks failed
A tooling review predicted a PID-1 signal problem from two symptoms we could not
explain: `OOMKilled: true` with **ExitCode 0**, and `--continue` failing to find
a conversation that plainly existed. Traced it, and the prediction was right --
though the culprit is not PID 1, it is one level below.
The path is tini (PID 1) -> entrypoint.sh (exec'd) -> expect -> spawn -> claude
`spawn` CANNOT be an exec: expect has to stay alive to drive the pty. So expect
is the process Docker signals, and everything depends on it passing things on.
It did neither, in two lines:
1. NO SIGNAL FORWARDING, no trap of any kind. `docker stop` sent SIGTERM to
expect, which died and took the pty with it. Claude Code never got a SIGTERM,
so it never ran SessionEnd hooks and never wrote lastSessionId/history --
which are written ONLY at a graceful shutdown. That is the entire reason
`claude --continue` answered "No conversation found to continue" with 33 MB of
transcripts in the volume beside it, and why we resume by scraping a session
id off a transcript filename.
2. `eof { exit }` RETURNED 0 FOR EVERY DEATH. A bare `exit` in expect is exit
ZERO. When the OOM-killer took the child, expect saw EOF and reported a clean
exit. `OOMKilled: true` with `ExitCode 0` was never Docker being odd -- it was
this line. It also meant `--restart on-failure` would read a memory kill as
success, which is why the policy had to be `unless-stopped`.
Fixed and MEASURED, old against new, in a container:
child exits 7 old -> 0 (the bug) new -> 7
SIGTERM to wrapper old -> 143, child's trap NEVER RAN
new -> 42, child trapped and cleaned up
Same file in both images; they were byte-identical, so the port copy takes the
same change.
Consequences worth stating: a kill now reports 137 rather than 0, so exit codes
mean what they say; `docker stop` gives Claude Code a real SIGTERM, so it runs
SessionEnd and writes the session index -- which may make the transcript-filename
resume unnecessary. That is not assumed here: the resume path stays as it is
until it is verified redundant.
126 lines
5.0 KiB
Plaintext
Executable File
126 lines
5.0 KiB
Plaintext
Executable File
#!/usr/bin/expect -f
|
|
# Start Claude Code for an unattended run, answering the first-run gates.
|
|
#
|
|
# Claude Code has three one-time interactive prompts, and every one of them is a
|
|
# silent, permanent hang for an agent with nobody at the keyboard — no error, no
|
|
# log line, just a container that looks healthy and does nothing:
|
|
#
|
|
# 1. the theme picker (first run, or whenever the installed version is
|
|
# newer than lastOnboardingVersion)
|
|
# 2. "do you trust this folder?" (per workspace)
|
|
# 3. the Bypass Permissions disclaimer (for --dangerously-skip-permissions)
|
|
#
|
|
# `seed-claude-config.py` pre-sets the config keys for 1 and 2. The disclaimer
|
|
# has no such key — it is meant to be accepted by a person once — so it is
|
|
# answered here instead. That is the honest reading of `sylph-agent loose`: the
|
|
# operator accepted it by choosing to run this, and the container is exactly the
|
|
# sandbox the warning asks for.
|
|
#
|
|
# ── Why the patterns are single words ──
|
|
# Claude Code draws its UI with ABSOLUTE COLUMN escapes between words, so the
|
|
# prompt arrives on the wire as
|
|
#
|
|
# 2.\x1b[8GYes,\x1b[13GI\x1b[15Gaccept
|
|
#
|
|
# A multi-word pattern like {Yes, I accept} therefore never matches, and the
|
|
# wrapper sits there looking like it is not running at all. Match one word.
|
|
|
|
set timeout 90
|
|
log_user 1
|
|
|
|
# Give the pty a wide, tall geometry. A detached `docker run -t` defaults to
|
|
# 80x24, and Claude Code hard-wraps to the terminal width — which truncates the
|
|
# Remote Control URL to "https://claude.ai/code/session_01…" in the one place
|
|
# you need to read it, and makes `docker logs` nearly unusable generally.
|
|
set stty_init "rows 50 cols 200"
|
|
|
|
set answered_theme 0
|
|
set answered_trust 0
|
|
set answered_bypass 0
|
|
|
|
spawn -noecho claude --dangerously-skip-permissions {*}$argv
|
|
set child_pid [exp_pid]
|
|
|
|
# 🔴 THIS WRAPPER USED TO SWALLOW BOTH THE SIGNAL AND THE EXIT STATUS, and those
|
|
# two omissions caused most of this project's multi-hour outages. Found
|
|
# 2026-09-03 by tracing the signal path, after a tooling review predicted exactly
|
|
# this from the symptoms.
|
|
#
|
|
# The path is: tini (PID 1) -> entrypoint.sh (exec'd) -> expect -> spawn -> claude
|
|
#
|
|
# `spawn` CANNOT be an exec: expect has to stay alive to drive the pty. So expect
|
|
# is the process Docker signals, and everything below it depends on expect
|
|
# passing things along. It did not.
|
|
#
|
|
# 1. NO SIGNAL FORWARDING. `docker stop` sent SIGTERM to expect, which died and
|
|
# took the pty with it. Claude Code never received a SIGTERM, so it never ran
|
|
# its `SessionEnd` hooks and never wrote `lastSessionId`/`history` to
|
|
# `~/.claude.json` -- which are written only at a GRACEFUL shutdown. That is
|
|
# the whole reason `claude --continue` answered "No conversation found to
|
|
# continue" with 33 MB of transcripts sitting in the volume beside it, and why
|
|
# we resume by scraping a session id off a transcript filename instead.
|
|
#
|
|
# 2. `eof { exit }` RETURNED 0 FOR EVERY DEATH. A bare `exit` in expect is exit
|
|
# ZERO. So when the kernel OOM-killer took the child, expect saw EOF and
|
|
# reported a clean exit -- `OOMKilled: true` with `ExitCode 0`, which is not
|
|
# Docker being odd, it is this line. It also meant `--restart on-failure`
|
|
# would have treated a memory kill as success, which is why the policy had to
|
|
# be `unless-stopped`.
|
|
#
|
|
# Both are fixed here. Signals are forwarded to the child and its real status is
|
|
# propagated, so a kill reads as 137, a clean stop lets Claude Code shut down
|
|
# properly, and the exit code means what it says.
|
|
proc forward {sig} {
|
|
global child_pid
|
|
catch { exec kill -$sig $child_pid }
|
|
}
|
|
trap { forward TERM } SIGTERM
|
|
trap { forward INT } SIGINT
|
|
trap { forward HUP } SIGHUP
|
|
|
|
expect {
|
|
-re {Choose} {
|
|
if {!$answered_theme} { set answered_theme 1; send "\r" }
|
|
exp_continue
|
|
}
|
|
-re {trust} {
|
|
if {!$answered_trust} {
|
|
set answered_trust 1
|
|
send_user "\n\[claude-autonomous] accepting the workspace trust prompt\n"
|
|
send "1\r"
|
|
}
|
|
exp_continue
|
|
}
|
|
-re {accept} {
|
|
if {!$answered_bypass} {
|
|
set answered_bypass 1
|
|
send_user "\n\[claude-autonomous] accepting the Bypass Permissions disclaimer\n"
|
|
send "2\r"
|
|
}
|
|
exp_continue
|
|
}
|
|
timeout {
|
|
# No new gate for a while: the session is up (or never had one). Stop
|
|
# matching so nothing later in the run can be answered by accident.
|
|
}
|
|
eof { exit }
|
|
}
|
|
|
|
# Hand the terminal over for the rest of the run.
|
|
interact
|
|
|
|
# Propagate the child's REAL exit status. `interact` returns when the child is
|
|
# gone; `wait` then yields {pid spawnid os_error status}. Without this the script
|
|
# simply ran off the end and returned 0 -- see the note at `spawn` above for what
|
|
# that cost.
|
|
catch wait result
|
|
set status 0
|
|
if {[info exists result] && [llength $result] >= 4} {
|
|
# os_error_flag (index 2) is -1 for a normal exit; anything else means the
|
|
# wait itself failed and the status field is not a status.
|
|
if {[lindex $result 2] == 0} {
|
|
set status [lindex $result 3]
|
|
}
|
|
}
|
|
exit $status
|