| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
| |
Every integration test is tagged :slow, so the default run selected none of them and then printed "All integration tests passed" over four "Ran 0 tests" lines.
The target now sums the tests actually run. Zero means it says skipped and names make test-all; a real run reports its count. A failing file still exits non-zero.
This is the third false-green in this Makefile. The other two read $? after a pipeline (7e03862).
|
| |
|
|
|
|
|
|
|
|
| |
Both targets read $? after a pipeline, so they got the exit status of the last command in it rather than of Emacs.
test-file piped Emacs through tee, so it printed "All tests passed" over a wall of FAILED lines and exited 0. It now redirects to the log, saves the status, and cats the log back.
validate piped Emacs through grep, and grep matched the failure line it was looking for, so a file with unbalanced parens printed its error in green and counted as validated. It now captures the output and the status separately. Verified against a deliberately broken file: before the fix, "All 78 files validated successfully" and exit 0.
make test and make test-all were never affected. They shell out to test-unit and test-integration without a pipeline.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Took the package-lint commit's foundation and finished the job. Three
groups of changes:
1. **checkdoc** is now clean across all three package files. I worked
through every warning: 13 missing-double-space-after-period fixes, 4
message-capitalization fixes, 1 trailing-period-on-error fix, 6
Lisp-symbol-should-be-quoted fixes, and one rephrase from "starts" to
the imperative "Log the start of an async check". One `chime-debug.el'
docstring also exceeded 80 columns and got split into two lines.
2. **`make lint'** ran elisp-lint but had been silently broken. The shell
glob `test-*.el' wasn't expanding (CWD ended up at project root, not
tests/), and the noisy validators flagged everything. I rewrote the
target to lint package files explicitly via absolute paths, preload
chime.el so the byte-compiler sees cross-file symbols when checking
chime-debug.el, and disable the validators that conflict with project
style or duplicate other lint steps:
- `--no-checkdoc' (covered by `eask lint checkdoc')
- `--no-package-lint' (covered by `eask lint package'; running it on
auxiliary files re-flags them as standalone packages)
- `--no-indent-character' (project uses spaces per
`.claude/rules/elisp.md'; the validator defaults to tabs)
- `--no-fill-column' (project allows up to 80; validator defaults to 70)
- `--no-indent' (false-positives on dash threading macros `->'/`->>')
The validators-disabled list is documented in the recipe header.
3. **Generated autoload files** (`chime-autoloads.el',
`tests/tests-autoloads.el') are now gitignored. `tests/tests-autoloads.el'
was tracked from an earlier commit; I removed it from the index.
Eask regenerates these on every `eask compile' / `eask install-deps'
run, so they don't belong in the tree.
Verified: `make compile' clean (with byte-compile-error-on-warn t),
`make test-all' green at 677 tests, `eask lint package' clean,
`eask lint checkdoc' clean, `make lint' clean. Full lint backlog is now
zero across all three checks.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I tagged the 15 end-to-end tests across `test-integration-chime-mode.el`,
`test-integration-recurring-events-tooltip.el`, and
`test-integration-startup.el` with `:tags '(:slow)`. The default `make test`
now runs only the 652 fast unit tests; `make test-all` runs the full 667.
While verifying, I hit a pre-existing bug in `tests/Makefile` where
`test-all`'s `--eval '(ert-run-tests-batch-and-exit)'` came back from eask's
argv passthrough with its quotes stripped, which broke shell parsing. Passing
the explicit `t` selector dodges whatever eask does to no-arg single-quoted
forms. The bug only mattered once integration tests were `:slow` — before
that, `test-all` overlapped fully with `test-unit` plus `test-integration`,
so nothing actually depended on it working.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I switched the test runner from `~/.emacs.d/elpa`-grep to eask. With
this, `make setup` installs every dep into a project-local `.eask/`,
so test runs don't depend on whatever's in my global elpa. It also
lets us catch missing Package-Requires entries before MELPA submission.
New targets:
- `make setup` — runs `eask install-deps --dev`
- `make compile` — byte-compiles chime.el and surfaces warnings that
checkdoc and elisp-lint don't catch
- `make coverage` — runs the unit suite under undercover and writes a
simplecov JSON to `.coverage/simplecov.json`
- `make test-all` — runs every test, including `:slow` tagged
- `-include makefile-local` in both Makefiles, for per-machine knobs
I added `ERT_FAST_SELECTOR` so `make test`, `test-unit`,
`test-integration`, and `test-file` exclude tests tagged `:slow`. When
we tag end-to-end integration tests as `:slow`, they'll stay out of
the fast feedback loop until someone explicitly asks for them via
`make test-all`.
Eask treats CWD as its workspace. So all eask invocations now run
from project root, with `(cd "tests/")' as the first `--eval' to
restore Emacs's default-directory. That preserves the relative loads
the existing test files and test-bootstrap.el rely on, without
touching either.
I updated `.gitignore` for `.eask/`, `.coverage/`, and the optional
`makefile-local` files.
|
| |
|
|
|
|
| |
Root Makefile now delegates to tests/Makefile. Fixed stale
UNIT_TESTS/INTEGRATION_TESTS definitions, added test-name target,
switched to per-file test execution for better isolation.
|
| |
|
