docs: plan for #303 — verify expected R/Python versions on Workbench#314
Conversation
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
| 1. **UNCONFIRMED: Unexpected versions flag** — The issue proposes an optional `unexpected_r_versions: bool` config flag (default `true`) to control whether "no unexpected versions are offered" is enforced. This would allow some customers to use an allow-list approach (strict) while others use additive (only verify expected versions are present, but don't fail if extras exist). The plan assumes this is added to `RuntimesConfig` in `src/vip/config.py`, but the implementation PR may defer this until a customer explicitly requests it. | ||
|
|
||
| 2. **UNCONFIRMED: Third scenario dependency** — The issue's third scenario ("An RStudio project opens with the expected R version") requires in-session code execution to read `R.version.string`. The issue notes this "depends on the in-session execution primitive (separate issue)". The plan assumes this scenario will be marked `@skip` or deferred to a follow-up PR until that primitive is available. | ||
|
|
||
| 3. **UNCONFIRMED: Workbench API alternative** — Workbench may expose runtime version information via `/api/server/settings` or a similar endpoint, which would allow simpler API-based verification instead of UI scraping. If such an API exists, the implementation should prefer it for the first two scenarios and only use Playwright for the third (in-session verification). The plan assumes no such API exists based on the Connect/Workbench client code review, but the implementer should confirm. |
There was a problem hiding this comment.
-
We should support both allow-list patterns and deny-list patterns. Make sure this is in the plan and not unconfirmed.
-
We need to scope this in the plan as well, we need to test that scenario.
-
No such API exists AFAIK.
There was a problem hiding this comment.
Addressed in commit 2cc89e5. The plan now explicitly scopes support for both allow-list and deny-list patterns, includes the in-session runtime verification scenario in scope, and removes the API-alternative assumption in favor of UI-based verification.
Co-authored-by: ian-flores <18703558+ian-flores@users.noreply.github.com>
|
@copilot resolve the merge conflicts in this pull request |
Co-authored-by: ian-flores <18703558+ian-flores@users.noreply.github.com>
Co-authored-by: ian-flores <18703558+ian-flores@users.noreply.github.com>
|
|
Pull request created: #327
|
|
🤖 Implementation PR opened: #aw_impl303 Implements the three scenarios from the plan:
Also adds
|
This plan proposes adding parallel runtime version verification for Workbench, mirroring the existing Connect tests in
src/vip_tests/connect/test_runtime_versions.feature. The implementation will reuse theexpected_r_versionsandexpected_python_versionsconfig fields for allow-list matching, add deny-list pattern checks for excluded versions, and use Playwright to verify runtime versions in the New Session dialog and in-session validation flow.Key components:
src/vip_tests/workbench/Merging this PR will trigger an implementation PR — comment to iterate on the plan first.
Plan
See
thoughts/shared/plans/2026-05-29-issue-303-workbench-runtime-versions.mdfor the full plan.> Generated by Triage open issues on
posit-dev/vipfor issue #303 · ● 7.6M · ◷