* fix(cli): only preserve case for raw-app runnableIds, not app/flow summaries
PR #8940 stopped lowercasing in sanitizeForFilesystem to fix#8939, where
a raw-app runnableId like CamelCaseTSRunnable produced a CamelCase YAML
metadata file but a lowercased code file, making them desync and
register as duplicate runnables on push.
That fix overshot. sanitizeForFilesystem is also reached by
newPathAssigner, which serves normal apps and flows where the input is
the script's human summary ("Get Users Data") rather than an identifier.
There the on-disk filename is the only artifact — there's no companion
YAML to keep in sync — so lowercasing was the right behavior. Removing
it changed both the on-disk filename and the !inline reference in
app.yaml / flow.yaml from get_users_data.inline_script.ts to
Get_Users_Data.inline_script.ts on the next pull, surfacing as
unwanted case churn for users updating to 1.693.x.
Add a preserveCase option to sanitizeForFilesystem (default false →
lowercase). newRawAppPathAssigner opts in; newPathAssigner stays on
the default. Update unit tests accordingly and add an end-to-end
raw-app round-trip in raw_app_sync.test.ts that pushes a CamelCase
backend runnable, pulls it back, and asserts both YAML and code file
preserve case with no lowercase orphan.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(cli): use readdir for exact-case orphan check on Windows
The CamelCase round-trip test used fileExists("camelcasetsrunnable.ts")
to assert no lowercase orphan was produced, which false-positives on
Windows since the filesystem is case-insensitive and resolves the
lookup to the existing CamelCaseTSRunnable.ts. Switch to readdir +
toContain so the exact on-disk casing is compared identically on Linux
and Windows.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(cli): preserve case in raw-app runnable filenames
The path assigner lowercased filesystem-safe names, so a raw-app
runnable id like CamelCaseTSRunnable produced a YAML metadata file
keeping the original case but a code file lowercased to
camelcasetsrunnable.ts. On the next push, loadRunnablesFromBackend
paired the two by case-sensitive name match, failed, and registered
the lowercase code file as a separate empty runnable — surfacing as
duplicate runnables in the app editor.
Stop lowercasing in sanitizeForFilesystem and dedupe path assigners
case-insensitively so case-only collisions still get a counter on
case-insensitive filesystems. Also match content/lock files and the
processed-id set case-insensitively in loadRunnablesFromBackend so
repos already pulled by the buggy CLI can recover on the next push
without re-pulling.
Fixes#8939
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* refactor(cli): extract readSiblingLock helper, drop test #ref comments
Address review:
- Both call sites in loadRunnablesFromBackend now use the same
case-insensitive lock-file lookup, so a partial-legacy repo with a
mixed-case code file and lowercase lock (or vice versa) works in the
orphan-code branch too, not just the YAML branch.
- Remove `Regression for #8939 …` comments from the new tests; CLAUDE.md
bans referencing the current task/issue from source comments. Test
names already describe the behavior under test.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(cli): update assigner-path assertions for case preservation
Three tests in inline_scripts_failure_preprocessor_unit.test.ts asserted
that the assigner produced lowercased filenames (get_users_data, step_b)
from mixed-case summaries — that was the buggy lowercasing behavior.
With case preserved end-to-end, the assigner now returns Get_Users_Data
and Step_B.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix(cli): preserve inline script files during flow generate-locks
Three bugs caused `wmill flow generate-locks` to destroy inline script
content and rename files:
1. YAML parser stripped unquoted `!inline` tags (treated as YAML tag,
not string prefix), leaving just the filename as script content.
Fix: register custom YAML tags for `!inline` and `!inline_fileset`.
2. Inline script files were renamed based on step summaries because
`extractInlineScriptsForFlows` was called with empty mapping `{}`.
Fix: call existing `extractCurrentMapping()` before replacement and
pass the mapping to preserve original filenames.
3. Lock file paths were derived from the assigner instead of the mapped
content path, causing inconsistent naming.
Fix: derive lock base path from mapped content path when available.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* test(cli): add unit tests for !inline YAML tag and mapping preservation
- YAML tag tests: unquoted/quoted !inline parsing, !inline_fileset,
nested structures, round-trip stability
- Mapping tests: path preservation with mapping, fallthrough without
mapping, lock path derivation from mapped content path, mixed
mapped/unmapped modules, dotted path handling
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(cli): correct yaml parse type cast and inline prefix check
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(cli): harden lock path for extensionless files and merge customTags
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(cli): separate unit tests from integration tests and fix test cleanup
- Rename 14 non-backend test files to *_unit.test.ts convention
- Add UNIT_ONLY env var guard in setup.ts to skip cargo build/backend startup
- Add test:unit and test:integration scripts to package.json
- Use setsid on Linux for process group management so stop() kills both
cargo and the windmill child process
- Fix exit handler to kill process group instead of just the direct child
- Add cleanupStaleTestResources() to drop orphaned windmill_test_* databases
and kill orphaned backend processes on startup
- Rewrite TESTING.md with current bun-based instructions
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(cli): fix process group approach - kill by db name instead of setsid
The setsid approach didn't work because setsid forks, making the PID
we get from Bun.spawn ephemeral. Instead, kill orphaned windmill child
processes by matching our unique database name in /proc/pid/environ.
Also add afterAll hook in setup.ts so full async cleanup (process kill
+ database drop) runs when all tests complete normally, not just on
SIGINT/SIGTERM.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(cli): address PR review feedback
- Remove duplicate cleanupStaleTestResources() call in getTestBackend()
(already called in setup.ts)
- Add regex guard on database names before SQL interpolation
- Extract shared killWindmillProcessesByEnvMatch() helper to deduplicate
process-killing logic
- Remove redundant test:integration script (test already runs everything)
- Flip setup.ts to if/else pattern for readability
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
hugocasa
Ruben Fiszel
centdix