testsAndMisc/linux_configuration/misc/testsAndMisc-bash/lib/tests/test_transcribe_pkgmgr.sh
Krzysztof kuhy Rudnicki 4f2b0ccdf7 Dissolve linux_configuration/scripts/ and single_use/
Two levels of pure taxonomy carrying no information, and a 2-directory
cap leaves no room for them. Categories now sit directly under
linux_configuration/: features, fixes, utils, misc, fresh-install,
installers, diagnostics, gaming, boot_recovery, periodic_background, lib,
meta, tests. A script's own lib/ and tests/ are exempt from the cap, so
those stay put.

Depth violations 248 -> 139; roughly 40 of the remainder are inside
digital_wellbeing and hosts, deferred until the guard-lib repoint.

The moves are the easy half. What breaks is every path that counted
levels, and three classes of it had to be found by resolving paths rather
than reading diffs:

  - 52 relative `source` lines for lib/common.sh and lib/android.sh. The
    correct prefix differs per file: 40 were ../../ and 12 were ../../../,
    so a blanket replace would have fixed one set and broken the other.
  - REPO_ROOT in 7 fixes/lib/tests harnesses and setup_syncyomi_guard.sh,
    each walking up four levels from a directory now two levels shallower.
    This caused a real regression -- the fixes suite went red with
    "log_warn: command not found" -- caught by running it and comparing
    against the pre-move baseline worktree.
  - conftest.py's sys.path entries, which took the pytest gate down with
    14 collection errors, and the `# shellcheck source=` directives, which
    took shellcheck down with SC1091.

16 symlinks in $HOME broke, including the entire i3blocks bar. All
repointed, and cpu_monitor/memory/disk/time/volume were run to confirm
they emit real output rather than merely resolving.

Verified: 1798 tests pass at 100% coverage, 8 of 9 lib/tests suites pass
(the ninth needs the coverage jail, identically at baseline), pre-commit
is clean, and the two untracked secret configs are still ignored after
their gitignore paths changed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
2026-08-23 18:53:30 +02:00

121 lines
4.8 KiB
Bash
Executable File

#!/usr/bin/env bash
# Tests for transcribe_pkgmgr.sh — package-manager detection and the CUDA check.
#
# No jail: every branch turns on `command -v <manager>` and every install goes
# through sudo, both of which the harness's PATH stub dir intercepts. The
# suite asserts on RECORDED ARGV, because which packages a branch would
# install is the part a presence test cannot see.
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
# shellcheck source=./transcribe_harness.sh
. "$SCRIPT_DIR/transcribe_harness.sh"
_t_setup_env
trap _t_teardown EXIT
# Globals the entry script defines above its source line.
OFFLINE=0
VENV_DIR="$TEST_TMPDIR/venv"
PY_HELPERS="$TEST_TMPDIR/helpers.py"
# shellcheck source=../transcribe_pkgmgr.sh
. "$TRANSCRIBE_LIB_DIR/transcribe_pkgmgr.sh"
# --- detect_pkg_mgr: none of them present -----------------------------------
_t_eq "none" "$(detect_pkg_mgr)" "detect_pkg_mgr reports 'none' with no manager on PATH"
# --- detect_pkg_mgr: each manager, in the documented priority order ---------
# Later entries are added on top of earlier ones to prove the ORDER, not just
# that each name is recognised: apt must win over dnf, dnf over yum, and so on.
_t_stub zypper
_t_eq "zypper" "$(detect_pkg_mgr)" "detect_pkg_mgr finds zypper"
_t_stub pacman
_t_eq "pacman" "$(detect_pkg_mgr)" "pacman takes priority over zypper"
_t_stub yum
_t_eq "yum" "$(detect_pkg_mgr)" "yum takes priority over pacman"
_t_stub dnf
_t_eq "dnf" "$(detect_pkg_mgr)" "dnf takes priority over yum"
_t_stub apt-get
_t_eq "apt" "$(detect_pkg_mgr)" "apt-get takes highest priority and reports 'apt'"
# --- has_libcublas12: absent everywhere -------------------------------------
cublas_rc=0
has_libcublas12 || cublas_rc=$?
_t_eq "1" "$cublas_rc" "has_libcublas12 fails when no CUDA library exists"
# --- has_libcublas12: found in a venv's nvidia site-packages ----------------
mkdir -p "$VENV_DIR/bin"
printf '#!/usr/bin/env bash\nprintf "3.11\\n"\n' >"$VENV_DIR/bin/python"
chmod +x "$VENV_DIR/bin/python"
mkdir -p "$VENV_DIR/lib/python3.11/site-packages/nvidia/cublas/lib"
: >"$VENV_DIR/lib/python3.11/site-packages/nvidia/cublas/lib/libcublas.so.12"
cublas_rc=0
has_libcublas12 || cublas_rc=$?
_t_eq "0" "$cublas_rc" "has_libcublas12 finds the library inside the venv's nvidia packages"
# --- ensure_cuda_runtime: already present is a no-op ------------------------
_t_reset_calls
ensure_cuda_runtime
_t_lacks "$(_t_calls)" 'sudo' "ensure_cuda_runtime installs nothing when cuBLAS is already present"
# --- ensure_cuda_runtime: offline and missing exits 6 -----------------------
rm -f "$VENV_DIR/lib/python3.11/site-packages/nvidia/cublas/lib/libcublas.so.12"
offline_rc=0
(
OFFLINE=1
ensure_cuda_runtime
) >/dev/null 2>&1 || offline_rc=$?
_t_eq "6" "$offline_rc" "ensure_cuda_runtime exits 6 when offline and CUDA is missing"
# --- ensure_cuda_runtime: no sudo means the install is skipped -------------
no_sudo_out="$( (ensure_cuda_runtime) 2>&1)" || true
_t_has "$no_sudo_out" 'sudo not found' "ensure_cuda_runtime reports a missing sudo instead of failing on it"
# --- ensure_cuda_runtime: the pacman branch ---------------------------------
# sudo is a pass-through so the manager stub below records the real argv.
printf '#!/usr/bin/env bash\nexec "$@"\n' >"$TEST_TMPDIR/bin/sudo"
chmod +x "$TEST_TMPDIR/bin/sudo"
_t_unstub apt-get
_t_unstub dnf
_t_unstub yum
_t_reset_calls
pac_rc=0
(ensure_cuda_runtime) >/dev/null 2>&1 || pac_rc=$?
_t_has "$(_t_calls)" 'pacman -Sy --noconfirm cuda cudnn' "the pacman branch installs cuda and cudnn"
_t_eq "6" "$pac_rc" "a still-missing cuBLAS after the install attempt exits 6"
# --- ensure_cuda_runtime: the apt branch ------------------------------------
_t_stub apt-get
_t_reset_calls
(ensure_cuda_runtime) >/dev/null 2>&1 || true
_t_has "$(_t_calls)" 'apt-get install -y nvidia-cuda-toolkit' "the apt branch installs the CUDA toolkit"
# --- ensure_cuda_runtime: dnf and yum share one branch ----------------------
_t_unstub apt-get
_t_stub dnf
_t_reset_calls
(ensure_cuda_runtime) >/dev/null 2>&1 || true
_t_has "$(_t_calls)" 'dnf install -y cuda cudnn' "the dnf branch installs cuda and cudnn"
_t_unstub dnf
_t_stub yum
_t_reset_calls
(ensure_cuda_runtime) >/dev/null 2>&1 || true
_t_has "$(_t_calls)" 'yum install -y cuda cudnn' "the shared dnf|yum branch also drives yum"
# --- ensure_cuda_runtime: the zypper branch ---------------------------------
_t_unstub yum
_t_unstub pacman
_t_stub zypper
_t_reset_calls
(ensure_cuda_runtime) >/dev/null 2>&1 || true
_t_has "$(_t_calls)" 'zypper install -y cuda cudnn' "the zypper branch installs cuda and cudnn"
# --- ensure_cuda_runtime: an unknown manager warns --------------------------
_t_unstub zypper
unknown_out="$( (ensure_cuda_runtime) 2>&1)" || true
_t_has "$unknown_out" 'cannot install CUDA automatically' "an unknown package manager warns rather than guessing"
_t_report "test_transcribe_pkgmgr.sh"