Compare commits

..
5 Commits
Author SHA1 Message Date
semantic-release-bot 50fa4320cd chore(release): 0.39.6 [skip ci]
## [0.39.6](https://github.com/sasjs/server/compare/v0.39.5...v0.39.6) (2026-07-14)

### Bug Fixes

* **api:** isolate drive.spec.ts's files folder from the real drive ([dec5191](https://github.com/sasjs/server/commit/dec51914911c10b66b43ebe95f77b0af63e2d03e))
* **api:** stop overwriting a failed JS/PY/R session back to completed ([4858245](https://github.com/sasjs/server/commit/4858245372fa19929b6e59c28148f5d004239c9d))
2026-07-14 09:05:49 +00:00
Allan BoweandGitHub 40b9fa8735 Merge pull request #391 from sasjs/issue-390
Issue 390
2026-07-14 10:02:31 +01:00
YuryShkoda fd31fe94ea Merge remote-tracking branch 'origin/main' into issue-390
# Conflicts:
#	api/src/controllers/internal/spec/Execution.spec.ts
2026-07-14 11:58:24 +03:00
YuryShkoda dec5191491 fix(api): isolate drive.spec.ts's files folder from the real drive
drive.spec.ts already isolates itself into a unique, timestamped
tmpFolder for getSasjsRootFolder()/getUploadsFolder(), cleaned up via
afterAll(() => deleteFolder(tmpFolder)). But getFilesFolder() - what
nearly every test in this file actually writes to - resolves through
a separate, unmocked function (getSasjsDriveFolder()/process.driveLoc),
so every run left real files/folders (e.g. 'level1', 'my/path/...')
behind in the shared api/sasjs_root/drive/files, causing later runs
to fail: folder-listing tests saw stale entries, and file-creation
tests got 409 Conflict against files a previous run already created.

Mock getFilesFolder() the same way getUploadsFolder() already is, so
it resolves inside the same isolated tmpFolder and gets cleaned up
by the existing afterAll. Verified by running the suite twice in a
row from a clean slate: both runs pass, and the real sasjs_root/drive
is never touched.
2026-07-14 11:32:01 +03:00
YuryShkoda 4858245372 fix(api): stop overwriting a failed JS/PY/R session back to completed
For JS/PY/R, processProgram sets session.state = failed (with
failureReason) itself when the interpreter process exits non-zero,
without throwing. ExecutionController.executeProgram then
unconditionally set session.state = completed right after
processProgram returned, silently overwriting that - so anything
downstream inspecting session.state (e.g. scheduleSessionDestroy's
expiresAfterMins branch) would see a crashed session mis-reported as
successful.

Guard the assignment so a failed state is never overwritten. No
effect on SAS, which sets state independently via its own spawned
process lifecycle.

Added Execution.spec.ts covering both the failure case (state stays
failed) and the success case (state still becomes completed), to
guard against regressing in either direction.
2026-07-14 11:29:30 +03:00
4 changed files with 101 additions and 23 deletions
+8
View File
@@ -1,3 +1,11 @@
## [0.39.6](https://github.com/sasjs/server/compare/v0.39.5...v0.39.6) (2026-07-14)
### Bug Fixes
* **api:** isolate drive.spec.ts's files folder from the real drive ([dec5191](https://github.com/sasjs/server/commit/dec51914911c10b66b43ebe95f77b0af63e2d03e))
* **api:** stop overwriting a failed JS/PY/R session back to completed ([4858245](https://github.com/sasjs/server/commit/4858245372fa19929b6e59c28148f5d004239c9d))
## [0.39.5](https://github.com/sasjs/server/compare/v0.39.4...v0.39.5) (2026-07-14) ## [0.39.5](https://github.com/sasjs/server/compare/v0.39.4...v0.39.5) (2026-07-14)
@@ -144,7 +144,16 @@ export class ExecutionController {
: '' : ''
// it should be deleted by scheduleSessionDestroy // it should be deleted by scheduleSessionDestroy
//
// Guarded: for JS/PY/R, processProgram sets state to `failed` itself
// (without throwing) when the interpreter process exits non-zero - if
// we unconditionally set `completed` here we'd silently overwrite that,
// and anything downstream inspecting session.state (e.g.
// scheduleSessionDestroy's expiresAfterMins branch in Session.ts) would
// see a crashed session mis-reported as successful.
if ((session.state as SessionState) !== SessionState.failed) {
session.state = SessionState.completed session.state = SessionState.completed
}
const resultParts = [] const resultParts = []
@@ -14,7 +14,7 @@ const preProgramVariables: PreProgramVars = {
httpHeaders: [] httpHeaders: []
} }
describe('ExecutionController.executeProgram (SAS failure path)', () => { describe('ExecutionController.executeProgram', () => {
let session: Session let session: Session
beforeEach(() => { beforeEach(() => {
@@ -33,6 +33,57 @@ describe('ExecutionController.executeProgram (SAS failure path)', () => {
await deleteFolder(session.path) await deleteFolder(session.path)
}) })
// Regression coverage: for JS/PY/R, processProgram sets session.state to
// `failed` itself (without throwing) on a non-zero interpreter exit.
// ExecutionController.executeProgram used to unconditionally overwrite
// that back to `completed` right after processProgram returned, silently
// losing the failure for anything downstream that inspects session.state.
describe('JS/PY/R failure path', () => {
it('does not overwrite a failed session state back to completed', async () => {
// mirrors processProgram's real JS/PY/R branch on a non-zero exit:
// it sets state/failureReason itself and resolves - it does not throw
jest
.spyOn(ProcessProgramModule, 'processProgram')
.mockImplementation(async () => {
session.state = SessionState.failed
session.failureReason = 'Error: process exited with code 1'
})
const controller = new ExecutionController()
await controller.executeProgram({
program: 'throw new Error("boom")',
preProgramVariables,
vars: {},
session,
runTime: RunTimeType.JS
})
expect(session.state).toBe(SessionState.failed)
})
it('still marks a genuinely successful session as completed', async () => {
jest
.spyOn(ProcessProgramModule, 'processProgram')
.mockImplementation(async () => {
session.state = SessionState.completed
})
const controller = new ExecutionController()
await controller.executeProgram({
program: 'console.log("hello")',
preProgramVariables,
vars: {},
session,
runTime: RunTimeType.JS
})
expect(session.state).toBe(SessionState.completed)
})
})
describe('SAS failure path', () => {
it('throws a SessionExecutionError carrying the complete log when the session fails', async () => { it('throws a SessionExecutionError carrying the complete log when the session fails', async () => {
const logPath = path.join(session.path, 'log.log') const logPath = path.join(session.path, 'log.log')
const logContent = const logContent =
@@ -63,3 +114,4 @@ describe('ExecutionController.executeProgram (SAS failure path)', () => {
}) })
}) })
}) })
})
+9
View File
@@ -28,6 +28,15 @@ jest
.spyOn(fileUtilModules, 'getUploadsFolder') .spyOn(fileUtilModules, 'getUploadsFolder')
.mockImplementation(() => path.join(tmpFolder, 'uploads')) .mockImplementation(() => path.join(tmpFolder, 'uploads'))
// getFilesFolder() resolves via getSasjsDriveFolder()/process.driveLoc, a
// separate root from getSasjsRootFolder()/process.sasjsRoot above - without
// this, every test in this file that creates content under the drive (e.g.
// 'level1', 'my/path/...') writes into the real, shared sasjs_root/drive
// instead of this run's isolated tmpFolder, leaking state into later runs.
jest
.spyOn(fileUtilModules, 'getFilesFolder')
.mockImplementation(() => path.join(tmpFolder, 'drive', 'files'))
import appPromise from '../../../app' import appPromise from '../../../app'
import { import {
UserController, UserController,