Compare commits

..
Author SHA1 Message Date
ChuckBuildsandClaude Opus 5 cbfb0e035d fix(web): stop the installer chmod stripping exec bits on every update
git tracks five scripts as mode 644 that first_time_install.sh then chmods to
755 (start_display.sh, stop_display.sh, the two install_*_service.sh, and
one-shot-install.sh does the same to first_time_install.sh). With
core.fileMode true, the default on Linux, git reports all five as modified
from then on, in files the user never touched.

The update button stashes local changes before pulling, so it is not blocked
by this. But it never pops that stash -- stash pop and stash apply appear
nowhere in the update flow -- so the mode change is stashed away and left
there, and the files revert:

    === file modes after the update button's stash ===
      664  first_time_install.sh      <- installer had made these 755
      664  start_display.sh
      664  stop_display.sh
      664  scripts/install/install_service.sh

So every web-UI update silently strips the executable bit from the installer's
own scripts, and leaves a stash entry holding the difference. start_display.sh
and stop_display.sh stop working from the shell afterwards.

A manual `git pull --rebase` over SSH fails outright, since nothing stashes for
it: "cannot pull with rebase: You have unstaged changes". That is the likely
source of the reports, since plenty of people update that way.

Tracking the five as 755 -- what they should always have been, as the
installer chmodding them attests -- removes the spurious mode change
entirely: nothing to stash, nothing stripped, no stash entry, and manual
pulls work.

The pull also passes --autostash, for the case the code explicitly tolerates:
when the stash fails it logs a warning and pulls anyway, and that pull is what
then fails. Autostash also pops what it stashes, which the manual stash does
not.

Note that `git add -A` after `git update-index --chmod=+x` silently reverts
the index to the on-disk mode, so the modes here were set by chmodding the
files themselves.

Regression test asserts the five stay tracked executable; reverting any one
of them fails it.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01STMbQE4YctTacQXfbYqKuW
2026-08-20 13:53:50 -04:00
9 changed files with 49 additions and 184 deletions
Regular → Executable
View File
View File
View File
+1 -75
View File
@@ -111,12 +111,6 @@ class RestoreOptions:
"""Which sections of a backup should be restored."""
restore_config: bool = True
#: Whether to take the backup's display.hardware block as well.
#: Off by default: that block describes the panel physically wired to
#: *this* device -- its size, chain length, mapping, multiplexing and
#: refresh cap. A backup carries the panel of the machine it was taken
#: on, and restoring one onto a different rig drives the wrong geometry.
restore_hardware: bool = False
restore_secrets: bool = True
restore_wifi: bool = True
restore_fonts: bool = True
@@ -555,72 +549,6 @@ def _copy_file(src: Path, dst: Path) -> None:
raise
_HARDWARE_PATH = ("display", "hardware")
def _restore_config_preserving_hardware(src: Path, dst: Path, keep_hardware: bool) -> None:
"""Copy a backed-up config.json, optionally keeping the local panel block.
display.hardware describes the panel physically attached to this device:
cols, rows, chain_length, hardware_mapping, panel_type, multiplexing and
the refresh-rate cap. None of that travels with a configuration -- it is a
property of the machine. Restoring a backup taken on a 512x64 rig onto a
128x32 one used to overwrite the smaller panel's geometry with the larger
one's, which is not a setting the user can see going wrong; the display
simply stops being right.
Falls back to a plain copy when either file cannot be parsed, so a restore
never fails because of this.
"""
if not keep_hardware:
_copy_file(src, dst)
return
try:
incoming = json.loads(src.read_text(encoding="utf-8"))
local = json.loads(dst.read_text(encoding="utf-8")) if dst.exists() else {}
except (OSError, ValueError) as exc:
logger.warning(
"[Backup] Could not merge local panel config (%s); restoring the "
"backup's config.json as-is", exc)
_copy_file(src, dst)
return
section, key = _HARDWARE_PATH
local_hw = (local.get(section) or {}).get(key)
if not isinstance(local_hw, dict) or not local_hw:
_copy_file(src, dst)
return
if not isinstance(incoming.get(section), dict):
incoming[section] = {}
incoming_hw = incoming[section].get(key)
incoming[section][key] = local_hw
if isinstance(incoming_hw, dict) and incoming_hw != local_hw:
logger.info(
"[Backup] Kept this device's display.hardware; the backup's panel "
"was %sx%s chain %s, this one is %sx%s chain %s",
incoming_hw.get("cols"), incoming_hw.get("rows"),
incoming_hw.get("chain_length"),
local_hw.get("cols"), local_hw.get("rows"),
local_hw.get("chain_length"))
# Write the merged result to a scratch file and hand it to _copy_file
# rather than renaming it into place here. _copy_file preserves the
# destination's mode and owner on purpose -- these config files are
# installed root-owned while the web interface that runs a restore is not
# root -- and a bare write would have replaced a root-owned config.json
# with one owned by the web user at whatever the umask allows.
scratch = dst.with_suffix(dst.suffix + ".merged-tmp")
try:
scratch.write_text(json.dumps(incoming, indent=2) + "\n", encoding="utf-8")
_copy_file(scratch, dst)
finally:
try:
scratch.unlink()
except OSError:
pass
def restore_backup(
zip_path: Path,
project_root: Path,
@@ -656,9 +584,7 @@ def restore_backup(
# Main config.
if options.restore_config and (tmp_dir / _CONFIG_REL).exists():
try:
_restore_config_preserving_hardware(
tmp_dir / _CONFIG_REL, project_root / _CONFIG_REL,
keep_hardware=not options.restore_hardware)
_copy_file(tmp_dir / _CONFIG_REL, project_root / _CONFIG_REL)
result.restored.append("config")
except OSError as e:
logger.error("[Backup] Failed to restore config.json: %s", e, exc_info=True)
Regular → Executable
View File
Regular → Executable
View File
+37 -3
View File
@@ -58,7 +58,7 @@ def repos(tmp_path):
def test_branch_with_upstream_uses_a_plain_pull(repos):
args, note, error = resolve_pull_command(str(repos))
assert error is None
assert args == ['git', 'pull', '--rebase']
assert args == ['git', 'pull', '--rebase', '--autostash']
assert note == ''
@@ -73,7 +73,7 @@ def test_branch_without_upstream_falls_back_to_origin_branch(repos):
args, note, error = resolve_pull_command(str(repos))
assert error is None
assert args == ['git', 'pull', '--rebase', 'origin', 'audit']
assert args == ['git', 'pull', '--rebase', '--autostash', 'origin', 'audit']
assert 'audit' in note
@@ -155,7 +155,7 @@ def test_switching_attaches_tracking_so_pull_needs_no_fallback(repos):
args, note, error = resolve_pull_command(str(repos))
assert error is None
assert args == ['git', 'pull', '--rebase']
assert args == ['git', 'pull', '--rebase', '--autostash']
assert note == ''
@@ -200,3 +200,37 @@ def test_stash_option_lets_the_switch_through_and_keeps_the_work(repos):
assert _git('branch', '--show-current', cwd=repos).stdout.strip() == 'other'
# The edit is not lost — it is on the stash.
assert 'switch to other' in _git('stash', 'list', cwd=repos).stdout
class TestInstallerDoesNotBlockTheUpdateButton:
"""first_time_install.sh chmods scripts that git tracked as 644.
With core.fileMode true -- the default on Linux -- that leaves five
permanently modified tracked files on every machine that ran the
installer, and `git pull --rebase` refuses to start:
error: cannot pull with rebase: You have unstaged changes.
Tracking them as executable makes the installer's chmod a no-op.
"""
CHMODDED = [
'first_time_install.sh',
'start_display.sh',
'stop_display.sh',
'scripts/install/install_service.sh',
'scripts/install/install_web_service.sh',
]
def test_scripts_the_installer_chmods_are_tracked_executable(self):
import subprocess
from pathlib import Path
root = Path(__file__).resolve().parent.parent
out = subprocess.run(['git', 'ls-files', '-s', *self.CHMODDED],
capture_output=True, text=True, cwd=str(root)).stdout
modes = {line.split()[3]: line.split()[0] for line in out.strip().split('\n') if line}
non_exec = sorted(f for f, m in modes.items() if m != '100755')
assert not non_exec, (
f"{non_exec} are chmodded by the installer but tracked non-executable, "
"so every install leaves the working tree dirty and the update "
"button cannot pull")
-104
View File
@@ -1,104 +0,0 @@
#!/usr/bin/env python3
"""A restore must not repoint this device at another machine's panel.
display.hardware describes the panel physically wired to this device -- cols,
rows, chain_length, hardware_mapping, panel_type, multiplexing, the refresh
cap. A backup carries the panel of the machine it was taken on. Restoring a
512x64 rig's backup onto a 128x32 one used to overwrite the smaller panel's
geometry with the larger one's, and nothing on screen explains why: the
display just stops being right.
That is not hypothetical. It happened, and the rig it happened to had to be
reflashed.
"""
import json
import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
from src.backup_manager import _restore_config_preserving_hardware # noqa: E402
BIG = {"display": {"hardware": {"cols": 128, "rows": 64, "chain_length": 4,
"hardware_mapping": "adafruit-hat-pwm"},
"runtime": {"gpio_slowdown": 4}},
"timezone": "America/New_York", "some-plugin": {"enabled": True}}
SMALL = {"display": {"hardware": {"cols": 64, "rows": 32, "chain_length": 2,
"hardware_mapping": "regular"},
"runtime": {"gpio_slowdown": 2}},
"timezone": "UTC"}
def _run(tmp, keep):
src = tmp / "backup_config.json"; src.write_text(json.dumps(BIG))
dst = tmp / "config.json"; dst.write_text(json.dumps(SMALL))
_restore_config_preserving_hardware(src, dst, keep_hardware=keep)
return json.loads(dst.read_text())
def test_local_panel_survives(tmp_path):
out = _run(tmp_path, keep=True)
hw = out["display"]["hardware"]
assert (hw["cols"], hw["rows"], hw["chain_length"]) == (64, 32, 2), (
"the restore repointed this device at the backup's panel")
assert hw["hardware_mapping"] == "regular", "panel wiring came from the backup"
def test_everything_else_is_restored(tmp_path):
out = _run(tmp_path, keep=True)
assert out["timezone"] == "America/New_York", "config was not restored"
assert out["some-plugin"] == {"enabled": True}, "plugin config was not restored"
assert out["display"]["runtime"] == {"gpio_slowdown": 4}, (
"only display.hardware should be held back")
def test_opting_in_takes_the_backups_panel(tmp_path):
out = _run(tmp_path, keep=False)
hw = out["display"]["hardware"]
assert (hw["cols"], hw["rows"], hw["chain_length"]) == (128, 64, 4)
def test_a_device_with_no_local_hardware_takes_the_backups(tmp_path):
src = tmp_path / "b.json"; src.write_text(json.dumps(BIG))
dst = tmp_path / "c.json"; dst.write_text(json.dumps({"timezone": "UTC"}))
_restore_config_preserving_hardware(src, dst, keep_hardware=True)
out = json.loads(dst.read_text())
assert out["display"]["hardware"]["cols"] == 128, (
"nothing local to preserve, so the backup's panel should be used")
def test_unparseable_local_config_still_restores(tmp_path):
src = tmp_path / "b.json"; src.write_text(json.dumps(BIG))
dst = tmp_path / "c.json"; dst.write_text("{ not json")
_restore_config_preserving_hardware(src, dst, keep_hardware=True)
assert json.loads(dst.read_text())["timezone"] == "America/New_York", (
"a restore must never fail because of this merge")
def test_the_destination_mode_is_preserved(tmp_path):
"""config.json is installed with a deliberate mode; a merge must not widen it.
_copy_file preserves the destination's mode and owner on purpose -- these
files are root-owned while the web interface running the restore is not.
Writing the merged result directly would have replaced that with whatever
the umask allowed.
"""
import os
import stat as statmod
src = tmp_path / "b.json"; src.write_text(json.dumps(BIG))
dst = tmp_path / "c.json"; dst.write_text(json.dumps(SMALL))
os.chmod(dst, 0o600)
_restore_config_preserving_hardware(src, dst, keep_hardware=True)
mode = statmod.S_IMODE(os.stat(dst).st_mode)
assert mode == 0o600, f"restore widened config.json from 0600 to {oct(mode)}"
assert json.loads(dst.read_text())["display"]["hardware"]["cols"] == 64
def test_no_scratch_file_is_left_behind(tmp_path):
src = tmp_path / "b.json"; src.write_text(json.dumps(BIG))
dst = tmp_path / "c.json"; dst.write_text(json.dumps(SMALL))
_restore_config_preserving_hardware(src, dst, keep_hardware=True)
leftovers = [p.name for p in tmp_path.iterdir() if "tmp" in p.name]
assert not leftovers, f"scratch files left in place: {leftovers}"
+11 -2
View File
@@ -1657,13 +1657,22 @@ def resolve_pull_command(project_dir):
backup, or following an install guide that names one. The update button
then reports a failure the user cannot act on.
``--autostash`` is passed for the same reason. Rebase refuses to start
when any tracked file is modified, and on these installs something always
is: first_time_install.sh chmods five scripts that git tracked as 644, so
every machine that ran the installer carries five permanent mode changes
and the update button reports "cannot pull with rebase: You have unstaged
changes". Those modes are corrected in this commit, but a user cannot pull
the correction while the pull is what is blocked, and any other local edit
would reproduce it anyway. Autostash reapplies the changes afterwards.
Returns ``(args, note, error)``. When ``origin/<branch>`` exists the pull
is made explicit against it, so the update proceeds and the branch is
given tracking information afterwards.
"""
upstream = _git_upstream(project_dir)
if upstream:
return ['git', 'pull', '--rebase'], '', None
return ['git', 'pull', '--rebase', '--autostash'], '', None
branch = _git_current_branch(project_dir)
if not branch:
@@ -1673,7 +1682,7 @@ def resolve_pull_command(project_dir):
)
if _git_remote_branch_exists(project_dir, branch):
return (
['git', 'pull', '--rebase', 'origin', branch],
['git', 'pull', '--rebase', '--autostash', 'origin', branch],
f"Branch '{branch}' had no upstream; pulled from origin/{branch} and set it as the upstream.",
None,
)