mirror of
https://github.com/NixOS/nix
synced 2025-07-08 06:53:54 +02:00
Tagging release 2.26.2
-----BEGIN PGP SIGNATURE----- iQFHBAABCAAxFiEEtUHVUwEnDgvPFcpdgXC0cm1xmN4FAmetA5oTHGVkb2xzdHJh QGdtYWlsLmNvbQAKCRCBcLRybXGY3g2pB/9JAFyjmaXuccbMTO/6x9qwsWuuXNLk OQWzfbdUekvsihZZSFZg1r7KqqXHCi64f0nxLPsJ/0oeDWZktJ5KnbV630nuUlDj ulLCpKdvhWFa8dVx9LiziGwQw4KLx8PjOfwThtQ4DqCWxWEmu6lKkijag9cE+ai4 3mw9YtUjBRxlXyhYLzWz3whLbv37c/m+R8iGS8xm8W260pmei6D0beOIPdfXYBQF PzPlPORyI08A06uqyA3z7bTxzmSMnzvu0QInCPCKSHzFUnTZPHUYuYStFl28NrZS fXKK59L0G7QEfdTRAmqQkdHdtPj2RlYFiMN0kQiNLflvKfGGWdi/kvdx =rRix -----END PGP SIGNATURE----- Merge tag '2.26.2' into sync-2.26.2 Tagging release 2.26.2
This commit is contained in:
Eelco Dolstra
commit
4055239936
1395 changed files with 24694 additions and 16040 deletions
360
tests/functional/common/functions.sh
Normal file
360
tests/functional/common/functions.sh
Normal file
|
|
@ -0,0 +1,360 @@
|
|||
# shellcheck shell=bash
|
||||
|
||||
set -eu -o pipefail
|
||||
|
||||
if [[ -z "${COMMON_FUNCTIONS_SH_SOURCED-}" ]]; then
|
||||
|
||||
COMMON_FUNCTIONS_SH_SOURCED=1
|
||||
|
||||
isTestOnNixOS() {
|
||||
[[ "${isTestOnNixOS:-}" == 1 ]]
|
||||
}
|
||||
|
||||
die() {
|
||||
echo "unexpected fatal error: $*" >&2
|
||||
exit 1
|
||||
}
|
||||
|
||||
readLink() {
|
||||
# TODO fix this
|
||||
# shellcheck disable=SC2012
|
||||
ls -l "$1" | sed 's/.*->\ //'
|
||||
}
|
||||
|
||||
clearProfiles() {
|
||||
profiles="$HOME/.local/state/nix/profiles"
|
||||
rm -rf "$profiles"
|
||||
}
|
||||
|
||||
# Clear the store, but do not fail if we're in an environment where we can't.
|
||||
# This allows the test to run in a NixOS test environment, where we use the system store.
|
||||
# See doc/manual/source/contributing/testing.md / Running functional tests on NixOS.
|
||||
clearStoreIfPossible() {
|
||||
if isTestOnNixOS; then
|
||||
echo "clearStoreIfPossible: Not clearing store, because we're on NixOS. Moving on."
|
||||
else
|
||||
doClearStore
|
||||
fi
|
||||
}
|
||||
|
||||
clearStore() {
|
||||
if isTestOnNixOS; then
|
||||
die "clearStore: not supported when testing on NixOS. If not essential, call clearStoreIfPossible. If really needed, add conditionals; e.g. if ! isTestOnNixOS; then ..."
|
||||
fi
|
||||
doClearStore
|
||||
}
|
||||
|
||||
doClearStore() {
|
||||
echo "clearing store..."
|
||||
chmod -R +w "$NIX_STORE_DIR"
|
||||
rm -rf "$NIX_STORE_DIR"
|
||||
mkdir "$NIX_STORE_DIR"
|
||||
rm -rf "$NIX_STATE_DIR"
|
||||
mkdir "$NIX_STATE_DIR"
|
||||
clearProfiles
|
||||
}
|
||||
|
||||
clearCache() {
|
||||
rm -rf "${cacheDir?}"
|
||||
}
|
||||
|
||||
clearCacheCache() {
|
||||
rm -f "$TEST_HOME/.cache/nix/binary-cache"*
|
||||
}
|
||||
|
||||
startDaemon() {
|
||||
if isTestOnNixOS; then
|
||||
die "startDaemon: not supported when testing on NixOS. Is it really needed? If so add conditionals; e.g. if ! isTestOnNixOS; then ..."
|
||||
fi
|
||||
|
||||
# Don’t start the daemon twice, as this would just make it loop indefinitely
|
||||
if [[ "${_NIX_TEST_DAEMON_PID-}" != '' ]]; then
|
||||
return
|
||||
fi
|
||||
# Start the daemon, wait for the socket to appear.
|
||||
rm -f "$NIX_DAEMON_SOCKET_PATH"
|
||||
# TODO: remove the nix-command feature when we're no longer testing against old daemons.
|
||||
PATH=$DAEMON_PATH nix --extra-experimental-features 'nix-command' daemon &
|
||||
_NIX_TEST_DAEMON_PID=$!
|
||||
export _NIX_TEST_DAEMON_PID
|
||||
for ((i = 0; i < 300; i++)); do
|
||||
if [[ -S $NIX_DAEMON_SOCKET_PATH ]]; then
|
||||
DAEMON_STARTED=1
|
||||
break;
|
||||
fi
|
||||
sleep 0.1
|
||||
done
|
||||
if [[ -z ${DAEMON_STARTED+x} ]]; then
|
||||
fail "Didn’t manage to start the daemon"
|
||||
fi
|
||||
trap "killDaemon" EXIT
|
||||
# Save for if daemon is killed
|
||||
NIX_REMOTE_OLD=$NIX_REMOTE
|
||||
export NIX_REMOTE=daemon
|
||||
}
|
||||
|
||||
killDaemon() {
|
||||
if isTestOnNixOS; then
|
||||
die "killDaemon: not supported when testing on NixOS. Is it really needed? If so add conditionals; e.g. if ! isTestOnNixOS; then ..."
|
||||
fi
|
||||
|
||||
# Don’t fail trying to stop a non-existant daemon twice
|
||||
if [[ "${_NIX_TEST_DAEMON_PID-}" == '' ]]; then
|
||||
return
|
||||
fi
|
||||
kill "$_NIX_TEST_DAEMON_PID"
|
||||
for i in {0..100}; do
|
||||
kill -0 "$_NIX_TEST_DAEMON_PID" 2> /dev/null || break
|
||||
sleep 0.1
|
||||
done
|
||||
kill -9 "$_NIX_TEST_DAEMON_PID" 2> /dev/null || true
|
||||
wait "$_NIX_TEST_DAEMON_PID" || true
|
||||
rm -f "$NIX_DAEMON_SOCKET_PATH"
|
||||
# Indicate daemon is stopped
|
||||
unset _NIX_TEST_DAEMON_PID
|
||||
# Restore old nix remote
|
||||
NIX_REMOTE=$NIX_REMOTE_OLD
|
||||
trap "" EXIT
|
||||
}
|
||||
|
||||
restartDaemon() {
|
||||
if isTestOnNixOS; then
|
||||
die "restartDaemon: not supported when testing on NixOS. Is it really needed? If so add conditionals; e.g. if ! isTestOnNixOS; then ..."
|
||||
fi
|
||||
|
||||
[[ -z "${_NIX_TEST_DAEMON_PID:-}" ]] && return 0
|
||||
|
||||
killDaemon
|
||||
startDaemon
|
||||
}
|
||||
|
||||
isDaemonNewer () {
|
||||
[[ -n "${NIX_DAEMON_PACKAGE:-}" ]] || return 0
|
||||
local requiredVersion="$1"
|
||||
local daemonVersion
|
||||
daemonVersion=$("$NIX_DAEMON_PACKAGE/bin/nix" daemon --version | sed 's/.*) //')
|
||||
[[ $(nix eval --expr "builtins.compareVersions ''$daemonVersion'' ''$requiredVersion''") -ge 0 ]]
|
||||
}
|
||||
|
||||
skipTest () {
|
||||
echo "$1, skipping this test..." >&2
|
||||
exit 77
|
||||
}
|
||||
|
||||
TODO_NixOS() {
|
||||
if isTestOnNixOS; then
|
||||
skipTest "This test has not been adapted for NixOS yet"
|
||||
fi
|
||||
}
|
||||
|
||||
requireDaemonNewerThan () {
|
||||
isDaemonNewer "$1" || skipTest "Daemon is too old"
|
||||
}
|
||||
|
||||
canUseSandbox() {
|
||||
[[ ${_canUseSandbox-} ]]
|
||||
}
|
||||
|
||||
requireSandboxSupport () {
|
||||
canUseSandbox || skipTest "Sandboxing not supported"
|
||||
}
|
||||
|
||||
requireGit() {
|
||||
[[ $(type -p git) ]] || skipTest "Git not installed"
|
||||
}
|
||||
|
||||
fail() {
|
||||
echo "test failed: $*" >&2
|
||||
exit 1
|
||||
}
|
||||
|
||||
# Run a command failing if it didn't exit with the expected exit code.
|
||||
#
|
||||
# Has two advantages over the built-in `!`:
|
||||
#
|
||||
# 1. `!` conflates all non-0 codes. `expect` allows testing for an exact
|
||||
# code.
|
||||
#
|
||||
# 2. `!` unexpectedly negates `set -e`, and cannot be used on individual
|
||||
# pipeline stages with `set -o pipefail`. It only works on the entire
|
||||
# pipeline, which is useless if we want, say, `nix ...` invocation to
|
||||
# *fail*, but a grep on the error message it outputs to *succeed*.
|
||||
expect() {
|
||||
local expected res
|
||||
expected="$1"
|
||||
shift
|
||||
"$@" && res=0 || res="$?"
|
||||
# also match "negative" codes, which wrap around to >127
|
||||
if [[ $res -ne $expected && $res -ne $((256 + expected)) ]]; then
|
||||
echo "Expected exit code '$expected' but got '$res' from command ${*@Q}" >&2
|
||||
return 1
|
||||
fi
|
||||
return 0
|
||||
}
|
||||
|
||||
# Better than just doing `expect ... >&2` because the "Expected..."
|
||||
# message below will *not* be redirected.
|
||||
expectStderr() {
|
||||
local expected res
|
||||
expected="$1"
|
||||
shift
|
||||
"$@" 2>&1 && res=0 || res="$?"
|
||||
# also match "negative" codes, which wrap around to >127
|
||||
if [[ $res -ne $expected && $res -ne $((256 + expected)) ]]; then
|
||||
echo "Expected exit code '$expected' but got '$res' from command ${*@Q}" >&2
|
||||
return 1
|
||||
fi
|
||||
return 0
|
||||
}
|
||||
|
||||
# Run a command and check whether the stderr matches stdin.
|
||||
# Show a diff when output does not match.
|
||||
# Usage:
|
||||
#
|
||||
# assertStderr nix profile remove nothing << EOF
|
||||
# error: This error is expected
|
||||
# EOF
|
||||
assertStderr() {
|
||||
diff -u /dev/stdin <("$@" 2>/dev/null 2>&1)
|
||||
}
|
||||
|
||||
needLocalStore() {
|
||||
if [[ "$NIX_REMOTE" == "daemon" ]]; then
|
||||
skipTest "Can’t run through the daemon ($1)"
|
||||
fi
|
||||
}
|
||||
|
||||
# Just to make it easy to find which tests should be fixed
|
||||
buggyNeedLocalStore() {
|
||||
needLocalStore "$1"
|
||||
}
|
||||
|
||||
enableFeatures() {
|
||||
local features="$1"
|
||||
sed -i 's/experimental-features .*/& '"$features"'/' "${test_nix_conf?}"
|
||||
}
|
||||
|
||||
onError() {
|
||||
set +x
|
||||
echo "$0: test failed at:" >&2
|
||||
for ((i = 1; i < ${#BASH_SOURCE[@]}; i++)); do
|
||||
if [[ -z ${BASH_SOURCE[i]} ]]; then break; fi
|
||||
echo " ${FUNCNAME[i]} in ${BASH_SOURCE[i]}:${BASH_LINENO[i-1]}" >&2
|
||||
done
|
||||
}
|
||||
|
||||
# Prints an error message prefix referring to the last call into this file.
|
||||
# Ignores `expect` and `expectStderr` calls.
|
||||
# Set a special exit code when test suite functions are misused, so that
|
||||
# functions like expectStderr won't mistake them for expected Nix CLI errors.
|
||||
# Suggestion: -101 (negative to indicate very abnormal, and beyond the normal
|
||||
# range of signals)
|
||||
# Example (showns as string): 'repl.sh:123: in call to grepQuiet: '
|
||||
# This function is inefficient, so it should only be used in error messages.
|
||||
callerPrefix() {
|
||||
# Find the closest caller that's not from this file
|
||||
# using the bash `caller` builtin.
|
||||
local i file line fn savedFn
|
||||
# Use `caller`
|
||||
for i in $(seq 0 100); do
|
||||
caller "$i" > /dev/null || {
|
||||
if [[ -n "${file:-}" ]]; then
|
||||
echo "$file:$line: ${savedFn+in call to $savedFn: }"
|
||||
fi
|
||||
break
|
||||
}
|
||||
line="$(caller "$i" | cut -d' ' -f1)"
|
||||
fn="$(caller "$i" | cut -d' ' -f2)"
|
||||
file="$(caller "$i" | cut -d' ' -f3)"
|
||||
if [[ $file != "${BASH_SOURCE[0]}" ]]; then
|
||||
echo "$file:$line: ${savedFn+in call to $savedFn: }"
|
||||
return
|
||||
fi
|
||||
case "$fn" in
|
||||
# Ignore higher order functions that don't report any misuse of themselves
|
||||
# This way a misuse of a foo in `expectStderr 1 foo` will be reported as
|
||||
# calling foo, not expectStderr.
|
||||
expect|expectStderr|callerPrefix)
|
||||
;;
|
||||
*)
|
||||
savedFn="$fn"
|
||||
;;
|
||||
esac
|
||||
done
|
||||
}
|
||||
|
||||
checkGrepArgs() {
|
||||
local arg
|
||||
for arg in "$@"; do
|
||||
if [[ "$arg" != "${arg//$'\n'/_}" ]]; then
|
||||
echo "$(callerPrefix)newline not allowed in arguments; grep would try each line individually as if connected by an OR operator" >&2
|
||||
return 155 # = -101 mod 256
|
||||
fi
|
||||
done
|
||||
}
|
||||
|
||||
# `grep -v` doesn't work well for exit codes. We want `!(exist line l. l
|
||||
# matches)`. It gives us `exist line l. !(l matches)`.
|
||||
#
|
||||
# `!` normally doesn't work well with `set -e`, but when we wrap in a
|
||||
# function it *does*.
|
||||
#
|
||||
# `command grep` lets us avoid re-checking the args by going directly to the
|
||||
# executable.
|
||||
grepInverse() {
|
||||
checkGrepArgs "$@" && \
|
||||
! command grep "$@"
|
||||
}
|
||||
|
||||
# A shorthand, `> /dev/null` is a bit noisy.
|
||||
#
|
||||
# `grep -q` would seem to do this, no function necessary, but it is a
|
||||
# bad fit with pipes and `set -o pipefail`: `-q` will exit after the
|
||||
# first match, and then subsequent writes will result in broken pipes.
|
||||
#
|
||||
# Note that reproducing the above is a bit tricky as it depends on
|
||||
# non-deterministic properties such as the timing between the match and
|
||||
# the closing of the pipe, the buffering of the pipe, and the speed of
|
||||
# the producer into the pipe. But rest assured we've seen it happen in
|
||||
# CI reliably.
|
||||
#
|
||||
# `command grep` lets us avoid re-checking the args by going directly to the
|
||||
# executable.
|
||||
grepQuiet() {
|
||||
checkGrepArgs "$@" && \
|
||||
command grep "$@" > /dev/null
|
||||
}
|
||||
|
||||
# The previous two, combined
|
||||
grepQuietInverse() {
|
||||
checkGrepArgs "$@" && \
|
||||
! command grep "$@" > /dev/null
|
||||
}
|
||||
|
||||
# Wrap grep to remove its newline footgun; see checkGrepArgs.
|
||||
# Note that we keep the checkGrepArgs calls in the other helpers, because some
|
||||
# of them are negated and that would defeat this check.
|
||||
grep() {
|
||||
checkGrepArgs "$@" && \
|
||||
command grep "$@"
|
||||
}
|
||||
|
||||
# Return the number of arguments
|
||||
count() {
|
||||
echo $#
|
||||
}
|
||||
|
||||
trap onError ERR
|
||||
|
||||
requiresUnprivilegedUserNamespaces() {
|
||||
if [[ -f /proc/sys/kernel/apparmor_restrict_unprivileged_userns ]] && [[ $(< /proc/sys/kernel/apparmor_restrict_unprivileged_userns) -eq 1 ]]; then
|
||||
skipTest "Unprivileged user namespaces are disabled. Run 'sudo sysctl -w /proc/sys/kernel/apparmor_restrict_unprivileged_userns=0' to allow, and run these tests."
|
||||
fi
|
||||
}
|
||||
|
||||
execUnshare () {
|
||||
requiresUnprivilegedUserNamespaces
|
||||
exec unshare --mount --map-root-user "$SHELL" "$@"
|
||||
}
|
||||
|
||||
fi # COMMON_FUNCTIONS_SH_SOURCED
|
||||
Loading…
Add table
Add a link
Reference in a new issue