Probed the two remaining candidates at HEAD instead of guessing:
nvidia_troubleshoot.sh is ruled OUT. Its trace dies at step 3 because
backup_file writes /etc/profile.backup.<stamp>, a new SIBLING in unbound /etc
— a file bind cannot cover siblings and bare /etc is refused. 239 of its 336
lines never execute, so keeping the unproven ones in the entry (the rule the
pacman split established) leaves the entry at ~265, still over the cap. The
file cannot satisfy both the rule and the goal.
install_leechblock.sh is the best-conditioned target left: exit 0, 93 files
written, and all 93 hashes byte-identical across two consecutive runs. Only
the echoed prefix path and curl's progress meter vary.
But it fires an unstubbed `pkill` at every Chromium-family browser. pkill is
not in the stub set and --prefix cannot contain it, because signals are not
filesystem writes. Two probe runs fired it, harmless only because no browser
happened to be open. That is now documented as its own trap, and the brief
tells the next session to pass --stub pkill and check for a live victim first.
Also recorded: for a target embedding $(date), diff the written files' CONTENTS
with the date normalized rather than falling back to comparing sizes — that
still catches a dropped line, which a size comparison does not.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011w2Pj4Ek3ZNMQVYcFHgXDY
`--bind-abs /usr/bin` overlays the directory bwrap needs to exec anything, so
the run dies with `execvp bash: No such file or directory` before the script
starts. A target whose last act replaces a binary there traces fine with every
other path bound and stops at that line with exit 1 on EPERM — usable as a
baseline (it is deterministic across runs), but blind to statements at or
after the failure, which is a constraint on where the seams can go.
Also noted: the absolute-path scanner misses a literal `ln -sf` target,
because it follows variables into `cat >`/`cp`.
The brief now targets install_pacman_wrapper.sh (316) with its baseline
already captured: 16 file hashes, one stubbed call, exit 1 at line 304.
Checked ahead of time that its source.sha256 drift manifest hashes only the
seven copied source files, not the installer, so a split leaves it stable.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011w2Pj4Ek3ZNMQVYcFHgXDY
Last session landed the --prefix harness but split nothing, by design. This
points the next one at install_usage_monitoring.sh (290, over by 40) with the
facts needed to do it rather than rediscover them: the section-by-section seam
map, the five globals that cross those seams, and the 30-line reference trace
with the content hashes the split must reproduce.
Three hazards are file-specific and each was measured, not guessed:
- linux_configuration/tests/test_usage_monitoring_installer_efficiency.sh
greps the installer's SOURCE TEXT for the pmon heredoc. Moving that heredoc
into a lib breaks the test, and it is not in the pre-commit set, so nothing
reports it.
- REPO_DIR walks up five levels from $0 and is interpolated into the generated
catch-up script, so recomputing it at a different depth silently points the
generated file at the wrong repo while the trace still exits 0.
- Line 222 is an unquoted heredoc whose internal escaping is load-bearing.
The brief's own verification procedure had to be corrected: tracing the
baseline from a detached worktree makes REPO_DIR resolve to the worktree, which
changes the generated file's content and produces a convincing 2-line diff out
of nothing. Normalising the trace text does not help, because the path is
inside the hashed content. The baseline is captured in place instead, and that
workflow is verified to diff empty pre-split.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011w2Pj4Ek3ZNMQVYcFHgXDY
Six remaining over-cap files write to /etc, $HOME/.local/bin and systemd
unit dirs by shell redirection, which no PATH stub can intercept -- so
they currently cannot be verified at all, only inspected. That blocks
most of what is left, including several of the named blockers.
The brief scopes one tooling job (a prefix redirect in
trace_shell_split.sh), states the constraints that must not be violated,
and records what was measured today: destinations are mostly $HOME-rooted
variables, so a HOME redirect covers a large share without touching the
scripts. Two cases that a bare HOME redirect will NOT catch are named
explicitly -- setup_thorium_startup.sh writes through `sudo -u mkdir -p`,
which the existing sudo stub swallows, and install_leechblock.sh uses
`rsync --delete`, which is destructive against a wrong root.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011w2Pj4Ek3ZNMQVYcFHgXDY
