[shivam5643]

What is this about?

BROWSERSTACK_PROJECT_NAME and BROWSERSTACK_BUILD_NAME were ignored by this service. The launcher read projectName / buildName only out of the capabilities, so a value exported in the environment never reached the session — the build landed on the dashboard under whatever the config file said.

Every other SDK, and the binary's own env→config map, already give the environment precedence over the config file. The service was the one place that did not, so a user who overrides the build name per CI run (the standard pattern) silently got the config name instead.

The fix

A private helper, _applyEnvBuildNames(capability), reads BROWSERSTACK_PROJECT_NAME / BROWSERSTACK_BUILD_NAME (trimmed; blank values ignored) and writes them back into the capabilities, so the session itself carries them and this._buildName / this._projectName (Test Reporting) read the same values. It is called from all four capability arms in the constructor - array and multiremote, with and without bstack:options.

Knock-on: _handleBuildIdentifier — please read

Because the env var is now folded into the caps, a buildName is present whenever BROWSERSTACK_BUILD_NAME is exported (on a BrowserStack session, or whenever the caps carry bstack:options). _handleBuildIdentifier drops the identifier when !this._buildName, so that branch no longer fires in this case and the identifier is appended to the env-sourced name instead of being discarded.

This is the direction already taken and approved in this repo, not a new one:

• SDK-4748 (f1ebafc, 079f4b8, merged) removed || process.env.BROWSERSTACK_BUILD_NAME from that guard, so the env var stopped suppressing the identifier.
• SDK-7075 (this PR) makes the env var supply the build name.

Together they give the coherent result: the env names the build, and the identifier is appended to it. The previously open #192 (SDK-4165) asked for the same end state and was superseded by SDK-4748.

The one test that encoded the old edge case — env var set, caps carrying no buildName — asserted the identifier was dropped. Its premise is now self-contradictory, since the env var is a buildName source, so it is rewritten to assert the corrected contract. A buildName counts as absent only when neither the caps nor the environment supply one, which should not set buildIdentifier from env when buildName is absent already covers and which still passes unchanged. The SDK-4748 comments that asserted "this service never reads that var as a buildName source" are updated, since that is no longer true.

Scope

Capability shape	What the env value does
has bstack:options	overwrites bstack:options.projectName / buildName; any top-level project / build the caps already carry are aligned to the same value (never added there)
no bstack:options, extension caps (e.g. goog:chromeOptions), BrowserStack session	written into the bstack:options the service creates
legacy caps (no bstack:options, no : keys), BrowserStack session	written to the top-level project / build keys
non-BrowserStack session without bstack:options	nothing - caps untouched, and _buildName / _projectName keep the caps values

Top-level project / build are only ever added where the caps have no : extension keys - webdriver rejects a capability mixing : extension keys with non-W3C unprefixed keys, and any capability with a : key gets bstack:options instead.

testObservabilityOptions.projectName / buildName, when set, still take precedence for Test Reporting naming (util.ts, cli/cliUtils.ts) - unchanged by this PR. The environment always wins for the session itself.

Verification

End to end, against a real session. The config file and the environment deliberately disagreed:

	config file said	dashboard shows
project	sdk7075repro_project	sdk7075v9_project_env_vars
build	sdk7075repro_build	sdk7075v9_build_env_vars v9_identifier_repro

Run via the automate/wdio_cucumber BStackAutomation sample (binary/CLI flow) — Spec Files: 1 passed, 1 total, session status passed. Both env values won; before the fix both came through as the config values. The identifier is appended, as described above.

Capability shapes checked directly against the hub - raw W3C POST /session to hub-cloud.browserstack.com, then GET api.browserstack.com/automate/sessions/<id>.json:

Request	Recorded on the session
alwaysMatch {browserName, build, project} only - no bstack:options	the top-level build / project values: the hub honours legacy keys in a W3C request
bstack:options {buildName, projectName} and top-level build / project with different values	the bstack:options values: bstack:options wins, so the top-level alignment above is defensive

Unit tests. 11 constructor tests in a dedicated describe with afterEach env cleanup: prefer env, keep caps when unset, blank values ignored, padded values trimmed, build-only override, service-created bstack:options, legacy caps, aligned top-level keys, multiremote with and without bstack:options, and an untouched non-BrowserStack session; plus the rewritten _handleBuildIdentifier test. Against the base launcher.ts 8 of the 11 fail, so they are not vacuous; the other 3 (env unset, blank env, non-BrowserStack session) are no-regression guards that pass either way. launcher.test.ts: 143 passed.

Full suite on this branch: 61 files, 1394 tests passed (the 3 unhandled errors from uploadLogsArchive.test.ts are pre-existing - identical on the base commit), tsc and eslint clean.

Related Jira task/s

SDK-7075

Release (mandatory for every PR — required for the ready-for-review label)

Version bump: (required — tick exactly one)

• minor (backwards-compatible feature)
• patch (bug fix or other small change)

Release notes type:

• New Feature
• Bug Fix
• Other Improvement

Release notes (customer-facing):

• Fixed BROWSERSTACK_PROJECT_NAME and BROWSERSTACK_BUILD_NAME being ignored — the environment now takes precedence over the project and build names in your config file, matching the other BrowserStack SDKs.

Release notes (internal):

• New _applyEnvBuildNames helper, called from all four capability arms of the launcher constructor, folds BROWSERSTACK_PROJECT_NAME / BROWSERSTACK_BUILD_NAME into the caps ahead of the this._buildName / this._projectName assignment, so the session and Test Reporting both carry the env value.
• Knock-on: a buildName is now present whenever BROWSERSTACK_BUILD_NAME is set, so _handleBuildIdentifier keeps the identifier rather than dropping it. Completes the direction SDK-4748 started; the test encoding the old edge case is rewritten and the stale comments corrected.
• Every capability shape is covered: bstack:options (incl. the one the service creates for extension caps), and top-level project / build for legacy caps on BrowserStack sessions. Non-BrowserStack sessions are left untouched.

Checklist

• Ready to review
• Has it been tested locally?

PR Validations

Run Tests: Comment RUN_TESTS to trigger sanity tests.

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

• Configuration used: Central YAML (base), Organization UI (inherited), Workspace UI (inherited)
• Review profile: ASSERTIVE
• Plan: Enterprise
• Run ID: 5c8c499f-4107-4ca4-9f4f-f3ee6ca490d9

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

_{Comment @coderabbitai help to get the list of available commands.}

[shivam5643]

v8 backport: #283 — same change, same tests, verified the same way. Please review the two together.

[github-actions]

🔴 SDK PR Review gate is red. Pending:

• The SDK PR Review Agent has not been run on the current head commit yet — run the SDK PR Review Agent (its verdict is advisory; this gate only requires that it ran on the latest commit).

It turns green once the SDK PR Review Agent has run on the current head commit (any verdict — the gate only requires that the review ran). A native reviewer approval is separately required by branch protection before merge.

[github-actions]

🔴 SDK PR Review gate is red. Pending:

• The SDK PR Review Agent has not been run on the current head commit yet — run the SDK PR Review Agent (its verdict is advisory; this gate only requires that it ran on the latest commit).

It turns green once the SDK PR Review Agent has run on the current head commit (any verdict — the gate only requires that the review ran). A native reviewer approval is separately required by branch protection before merge.

[rounak610]

@shivam5643 Please test a scenario where you set different project names and build names using environment variables and in the config file. Then check whether the O11y build and the Automate/App Automate build contain the names set using the environment variables.

[shivam5643]

@shivam5643 Please test a scenario where you set different project names and build names using environment variables and in the config file. Then check whether the O11y build and the Automate/App Automate build contain the names set using the environment variables.

Yes working, have verified.
attached here: https://browserstack.atlassian.net/browse/SDK-7075?focusedCommentId=2459470

[github-actions]

🔴 SDK PR Review gate is red. Pending:

• The SDK PR Review Agent has not been run on the current head commit yet — run the SDK PR Review Agent (its verdict is advisory; this gate only requires that it ran on the latest commit).

It turns green once the SDK PR Review Agent has run on the current head commit (any verdict — the gate only requires that the review ran). A native reviewer approval is separately required by branch protection before merge.

[shivam5643]

✅ Good to go

Per-File Confidence

File	Status
.changeset/pr-282.md	✅ All Clear
packages/browserstack-service/src/launcher.ts	✅ All Clear
packages/browserstack-service/tests/launcher.test.ts	✅ All Clear

Change map (generated deterministically from the diff)

graph LR
  subgraph nwdio_service["wdio-service"]
    npackages_browserstack_service_src_launcher_ts["⚠ launcher.ts<br/>~113 lines"]
    npackages_browserstack_service_tests_launcher_test_ts["⚠ launcher.test.ts<br/>~113 lines"]
    n_changeset_pr_282_md["pr-282.md<br/>~5 lines"]
  end

Loading

↻ This verdict comment is the review anchor — it's updated in place on each run (the gate posts its status separately).

— SDK PR Review Agent

[github-actions]

🟢 SDK PR Review gate is green — the SDK PR Review Agent has run on the current head commit (verdict: success).

This gate confirms a review ran on the latest commit. The verdict itself is advisory — read the findings and use your judgement; it does not block merge. A native GitHub reviewer approval is still separately required by branch protection before this PR can merge.

[github-actions]

🟢 SDK PR Review gate is green — the SDK PR Review Agent has run on the current head commit (verdict: success).

This gate confirms a review ran on the latest commit. The verdict itself is advisory — read the findings and use your judgement; it does not block merge. A native GitHub reviewer approval is still separately required by branch protection before this PR can merge.

[shivam5643]

RUN_TESTS

[minionhelperappqa]

[SDK Wdio Test] TRA build state: failed | Stability 98% — verdict: success. Passed: 51, Failed: 1, Aggregate: 52. TRA: https://observability.browserstack.com/builds/sktqkkdampcskw3nnggpiuetif2qqf1m87wdwfvs

[github-actions]

🔴 SDK PR Review gate is red. Pending:

• The SDK PR Review Agent has not been run on the current head commit yet — run the SDK PR Review Agent (its verdict is advisory; this gate only requires that it ran on the latest commit).

It turns green once the SDK PR Review Agent has run on the current head commit (any verdict — the gate only requires that the review ran). A native reviewer approval is separately required by branch protection before merge.