Stop ug configure erroring on feature-disabled workspaces; hide unramped managed commands - #500
Open
david-siqi-liu wants to merge 4 commits into
Open
Stop ug configure erroring on feature-disabled workspaces; hide unramped managed commands#500david-siqi-liu wants to merge 4 commits into
ug configure erroring on feature-disabled workspaces; hide unramped managed commands#500david-siqi-liu wants to merge 4 commits into
Conversation
…amped managed commands Customers running `ug configure` on a workspace where workspace-managed coding-agent configs are disabled server-side hit an alarming, circular "ERROR ... Use `ug configure`" message and dead-ended. Root cause: refresh_managed_config treated FEATURE_DISABLED like a transient read failure and fell back to a stale cached config, masking the disabled state and routing an admin into the managed setup flow, which then raised the message. - managed_config.py: treat FEATURE_DISABLED as authoritative. Return (None, feature_disabled=True) and clear the persisted cache, so `ug configure` stays on the per-user flow, a launch does not re-apply a policy the workspace turned off, and a later transient failure cannot resurrect it. - cli.py: initialize coding_agent_config_feature_disabled on the --dry-run path (pre-existing UnboundLocalError when the local cache was empty). - Remove the managed-config authoring surface (ug setup / ug publish) from the public README and hide those commands from `ug --help` while the feature is still in development and unramped. They stay runnable behind ENABLE_MANAGED_AGENT_CONFIG; `ug export` stays documented. Co-authored-by: Isaac <no-reply@databricks.com>
david-siqi-liu
marked this pull request as ready for review
September 4, 2026 19:52
lilly-luo
reviewed
Sep 4, 2026
lilly-luo
reviewed
Sep 4, 2026
lilly-luo
reviewed
Sep 4, 2026
Comment on lines
+2426
to
+2430
| # --dry-run avoids the fetch but still applies the last saved config. It doesn't reach the | ||
| # gateway, so the feature-disabled signal is unknown here — default it to False so the guidance | ||
| # check below is well-defined (without this, a dry run with no cached config raised | ||
| # UnboundLocalError). | ||
| coding_agent_config_feature_disabled = False |
Collaborator
There was a problem hiding this comment.
can you just return here and comment out everything below? or at least make this a hidden env var that we can use to test/
Per review, remove the explanatory inline comments on the FEATURE_DISABLED branch (the behavior is already in the docstring) and on the hidden `ug setup` / `ug publish` registrations, and trim the --dry-run comment to one line. Co-authored-by: Isaac <no-reply@databricks.com>
The no-config guidance path checks workspace admin status, which shells out to the `databricks` binary. That binary is absent in CI, so the new dry-run test raised FileNotFoundError there (it passed locally only because the CLI was present). Stub the token and admin calls, matching the sibling bare-ug tests. Co-authored-by: Isaac <no-reply@databricks.com>
…code file The state-isolation fixture rebinds STATE_PATH and APP_DIR but not managed_config.MANAGED_STATE_PATH, which is bound from APP_DIR at import time, so save_managed_state wrote to the developer's real ~/.ucode/managed-state.json. This PR's FEATURE_DISABLED change made test_feature_disabled_sets_flag_when_there_is_no_fallback hit that unmocked write. Rebind MANAGED_STATE_PATH in the fixture, matching the existing STATE_PATH defense-in-depth. Co-authored-by: Isaac <no-reply@databricks.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
Customers running
ug configureon a workspace where workspace-managed coding-agent configs are disabled server-side (FEATURE_DISABLED) hit an alarming, circular error and then dead-ended:Root cause:
refresh_managed_configtreated FEATURE_DISABLED like a transient read failure and fell back to a stale cached config, returning it with the feature-disabled flag unset. That masked the disabled state, so an admin was routed into the managed setup flow, which re-read the workspace, got FEATURE_DISABLED, and raised the message.Changes
managed_config.py: treat FEATURE_DISABLED as authoritative rather than transient. Return no config with the feature-disabled flag set, and clear the persisted cache. Soug configurestays on the normal per-user flow, a launch does not re-apply a policy the workspace has turned off, and a later transient read or token failure cannot resurrect the disabled policy through the fallback.cli.py: initialize the feature-disabled flag on the--dry-runpath. This was a pre-existing UnboundLocalError when the local cache was empty, surfaced by review.ug setupandug publish) from the public README, and hide those commands fromug --help, while workspace-managed configs are still in development and unramped. They remain runnable behindENABLE_MANAGED_AGENT_CONFIG.ug exportstays documented.Testing
Review
Ran a simplify pass (no changes needed) and a Codex adversarial review. The two P1 findings it surfaced, persisted-cache resurrection and the pre-existing dry-run crash, are both fixed here.
This pull request and its description were written by Isaac.