-
Backport six post-cut fixes onto release/v0.9.0 (#4596)
frostbyte_neo 发布于
2026-08-11 18:31:13 +00:00 -
revert(sessions): unwind the #2150 approval/attribution stack (#3446, #3422, #3416) (#4318)
-
revert(sessions): remove delegated approval authority (#3446)
Reverts the delegated approval feature from #3446, returning to
owner-only approval (the deny-by-default behavior from #3416). Owners
can no longer delegate a "can_approve" capability to shared editors;
approvals are again restricted to the session owner, while editors keep
reject/cancel.The change is a faithful inverse of #3446 rebased on current main:
files untouched since #3446 revert byte-identical to their pre-feature
state; files later commits also modified keep those newer changes and
drop only the approval lines.Migration handled non-destructively for deployed databases:
- The original additive migration (c4d5e6f7a8b9) is kept intact so
already-migrated databases still resolve their history. - A new forward migration (f7a8b9c0d1e2) drops the session_permissions
.can_approve column; its downgrade re-adds it.
Also removes a dangling import of _approval_access_from_grants in
sessions/__init__.py left by the later wildcard-import refactor (#3934),
which otherwise broke server import after the helper was reverted.Signed-off-by: Serena Ruan serena.rxy@gmail.com
- revert(sessions): remove shared-message attribution (#3422)
Reverts the model-visible shared-message authorship feature from #3422.
Messages no longer gain[author]:prefixes in the model prompt, the
SHARED_SESSION_AUTHORSHIP_INSTRUCTION framework instruction is removed,
and the OMNIGENT_SHARED_MESSAGE_ATTRIBUTION_ENABLED switch is gone.
Persistedcreated_byauthorship (a store-level column predating #3422)
is unaffected.Rebased on current main, keeping later independent work in the same
regions:- Smart Routing's conditional
model_overrideon the native-terminal
forward path is preserved. - The
host_storeparameter added to the event-forward path is kept. - The two
test_external_interrupt_*tests from #4160 (which overlap
#3422's added block in test_sessions_endpoints.py) are kept; only
#3422'stest_external_user_message_strips_model_author_prefixis
removed.
Also removes dangling imports of
_strip_pending_author_prefixin
orchestration.py and sessions/__init__.py left after the helper's
definition was reverted.Signed-off-by: Serena Ruan serena.rxy@gmail.com
- revert(sessions): restore editor approval authority (#3416)
Reverts the owner-only approval restriction from #3416. Approval events
and URL-based elicitation resolution are gated at LEVEL_EDIT again, so
shared editors — not only the owner — can resolve approvals.SECURITY REGRESSION (intentional, per request): #3416 was a security
fix. Shared-session tools execute with the session owner's runner
identity and ambient credentials, so a shared editor can once more
authorize owner-credentialed tool calls. This, together with the #3422
and #3446 reverts, fully unwinds the #2150 stack and re-opens #2150.Signed-off-by: Serena Ruan serena.rxy@gmail.com
Signed-off-by: Serena Ruan serena.rxy@gmail.com
(cherry picked from commit7efe05623b)
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- feat(web): make the header Chat/Terminal switcher a segmented toggle (#4385)
The header switcher hid both destinations behind a dropdown: a
MessagesSquare + chevron trigger you had to open before you could see
which view you were in or switch to the other one. Reading the current
view took a hover (the tooltip), and switching took two clicks.Replace it with a two-segment icon toggle in a shared track. Both
destinations are always on screen, the active one is filled, and
switching is a single click. Sits in the same header slot, immediately
left of Share, at the same 32px scale as the neighbouring controls
(size-6 segments in a p-0.5 track).Behavior is unchanged: the same TerminalFirstContext drives it, it
self-gates for non-terminal-first sessions, the iOS shell (native
Liquid Glass bar), and rail-opened shell views, and Terminal stays
disabled — with a spinner while a PTY is coming up — until one is
reachable. Each segment carries aria-pressed and a tooltip naming it,
so the icon-only control stays legible to pointer and AT users alike;
the Terminal tooltip doubles as the "starting up" explanation.Collapsing the menu drops the machinery it needed: the controlled
tooltip (two merged Slots on one node dropped its listeners), the
pointer-vs-keyboard close-refocus ref, and the e2e open-retry loop
that existed because a toggle-trigger click could net back to closed.Co-authored-by: Isaac
Signed-off-by: Dhruv Gupta dhruv.gupta@databricks.com
(cherry picked from commit95186250cb)
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(claude-native): keep MCP tool search on for gateway-backed native Claude (#4533)
The native-claude launch config unconditionally set
CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1 on the ucode and bedrock provider
paths. That flag disables all experimental betas, including MCP tool
search (which rides on theadvanced-tool-usebeta). With tool search off,
Claude Code loads every MCP tool schema eagerly, inflating the context
window — for an isaac-omni session with ~187 MCP tools that is ~88k tokens
spent up front instead of on demand.The disable flag existed to avoid the gateway 400ing on
invalid beta flag.
But in gateway-aware mode (CLAUDE_CODE_USE_GATEWAY=1) Claude Code negotiates
the anthropic-beta set with the gateway rather than sending every flag
blindly, and the Databricks AI Gateway now accepts the flags it sends
(verified end-to-end against a live gateway: a CLAUDE_CODE_USE_GATEWAY=1
turn sends advanced-tool-use-2025-11-20 / prompt-caching-scope-2026-01-05 /
advisor-tool-2026-03-01 and completes with no 400). So the workaround is no
longer needed when USE_GATEWAY=1.- _provider_config_for_native_claude (generic gateway path): already
guarded on CLAUDE_CODE_USE_GATEWAY (unchanged). - _ucode_config_for_profile: this path always launches in gateway mode
(it sets CLAUDE_CODE_USE_GATEWAY=1 itself), so drop the disable flag
outright rather than guard it. Restores the pre-#4074 behavior. - _bedrock_config_for_native_claude: add the same USE_GATEWAY guard the
generic gateway path uses, so a bedrock-style corporate gateway running
in gateway-aware mode keeps tool search on. Real AWS Bedrock (no
USE_GATEWAY) is unchanged — the flag still gets set.
Tests: update the ucode assertion, add positive coverage for the gateway
and bedrock paths under USE_GATEWAY=1, and make the env-sensitive tests
deterministic by clearing CLAUDE_CODE_USE_GATEWAY.Co-authored-by: harry-yao_data harry.yao@databricks.com
(cherry picked from commit55a270a2d8)
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(codex): speak codex's model vocabulary on a CLI login (#4558)
Codex's own backend (ChatGPT account or API key) names models with a
dotted version,gpt-5.6-sol. Databricks serving names the same model
with hyphens only,databricks-gpt-5-6-sol. Two places sent the wrong
one, so every codex dispatch on a CLI login failed at launch with a 400.The curated codex catalog carried the Databricks spelling, so selecting
any offered model was rejected. It now carries codex's own slugs, which
still fold to the same comparable spelling, leaving routed-arm matching
unchanged.The launch default resolved through the generic OpenAI catalog, whose
newest row is the bare family aliasgpt-5.6that codex rejects as a
family name. Only the Databricks-gateway branch consults that catalog
now; a codex CLI login defaults to a concrete variant from codex's own
catalog. The Databricks branch keeps its hyphenated ids.Co-authored-by: Isaac
Signed-off-by: Dhruv Gupta dhruv0811@gmail.com
(cherry picked from commit5857a2c3d6)
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(host): keep CLAUDE_CODE_USE_GATEWAY / ENABLE_TOOL_SEARCH in the runner env allowlist (#4553)
A launcher (e.g. Databricks' isaac) sets CLAUDE_CODE_USE_GATEWAY=1 and
ENABLE_TOOL_SEARCH=true in its process env so the native-claude harness keeps
MCP tool search on (schemas load on demand). But the host daemon env
(_build_host_daemon_env) and the runner env (_build_runner_env) are both
built from_RUNNER_ENV_ALLOWLIST, and neither var was on it — so they were
stripped at daemon spawn and never reached the runner process.The native-claude provider path (
_provider_config_for_native_claude,
_ucode_config_for_profile,_bedrock_config_for_native_claude) reads
CLAUDE_CODE_USE_GATEWAY from os.environ to decide whether to set
CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1. With it stripped, the runner saw it
absent, re-added the disable flag, and Claude Code turned tool search off —
loading every MCP tool schema eagerly (~88k tokens for ~190 MCP tools at
startup instead of on demand).Add both non-secret boolean flags to
_RUNNER_ENV_ALLOWLIST, beside the
existing CLAUDE_CODE_USE_BEDROCK / CLAUDE_CODE_SKIP_BEDROCK_AUTH flags (same
category). The single allowlist is consulted by both gates, so the vars now
survive daemon spawn and runner spawn and reach the guard.Tests: assert both vars survive
_build_host_daemon_env(local + remote) and
_build_runner_env. They fail before this change and pass after.Co-authored-by: harry-yao_data harry.yao@databricks.com
(cherry picked from commita2de2b44ac)
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com-
fix(routing): count managed-settings AIGW backing for claude-native (#4491)
-
fix(routing): count managed-settings AIGW backing for claude-native
claude_gateway_inference_backed() returned False whenever
resolve_native_claude_config yielded no config — the case for a
subscription (Claude Code login) provider. But Claude Code itself still
routes all inference through an AI Gateway when an enterprise managed
settings file pins ANTHROPIC_BASE_URL, so Smart Routing was being gated
off for a genuinely gateway-backed launch. Codex already reads its own
config.toml base_url; this brings Claude to parity.Add a fallback: read Claude Code managed settings and treat the launch as
gateway-backed when env.ANTHROPIC_BASE_URL is a Databricks AI Gateway URL
(validated with is_databricks_ai_gateway_url) and a credential is
delivered via top-level apiKeyHelper or a truthy env.CLAUDE_CODE_USE_GATEWAY.
Managed settings win at the real launch, so this signal can flip the
answer to True even when the omnigent provider is subscription.Co-authored-by: Isaac
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(routing): validate the resolve-path base URL as a Databricks AIGW
The resolve-based branch of claude_gateway_inference_backed() returned
True on just ANTHROPIC_BASE_URL + api_key_helper being present, without
checking the URL is actually a Databricks AI Gateway. A bare
api.anthropic.com (or any non-Databricks Anthropic-compatible endpoint)
would qualify — but the external task_v1 router's picks are Databricks
catalog ids that endpoint cannot serve. Require
is_databricks_ai_gateway_url() on the resolved base URL too, matching the
managed-settings fallback and the Codex check.Co-authored-by: Isaac
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(routing): resolve cli-config codex base URL from the shared config.toml
native_codex_launch_base_url() returned None for a cli-config launch,
because such a launch pins only a model_provider name — the provider
table (with base_url) lives in the user's shared ~/.codex/config.toml,
which the launch never inlines. So codex_gateway_inference_backed()
reported a genuinely AIGW-routed cli-config provider as not backed,
gating Smart Routing off. This is the Codex analogue of the Claude
managed-settings gap.Read the shared config.toml in the final branch: extract the pinned
provider name (codex_session_meta_model_provider), locate the user's
CODEX_HOME config via codex_home_config_source_from_env, and return
model_providers..base_url with tomllib. openai (Codex's own login)
and omnigent_databricks (the profile branch's generated id) have no
user-config table, so they stay None. Any read/parse failure returns
None — an unreadable config is unknown, not backed. codex_gateway
inference_backed() is unchanged; it validates the URL as before.Co-authored-by: Isaac
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(routing): resolve codex config-default base URL for the empty-override launch
The prior commit covered a cli-config launch that pins a model_provider
name, but the user's Databricks-wide setup hits a different path: when no
omnigent provider resolves and the config default is not dismissed,
resolve_native_codex_launch leaves config_overrides empty on purpose so
Codex uses its own config.toml top-level model_provider default. On such
a machine that default is a Databricks AIGW provider, yet the probe saw
empty overrides and reported not-backed.Extend native_codex_launch_base_url: when a launch pins no model_provider
override and no profile, resolve the config.toml top-level model_provider
default's base_url (unless the user dismissed the default, which pins
Codex's built-in openai). An explicit model_provider="openai" override
(subscription / dismissed paths) still returns None — only a truly
unpinned launch reads the config default. Factor the shared table lookup
into _config_toml_provider_base_url, used by both the cli-config and
config-default paths.Co-authored-by: Isaac
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- fix(routing): count a resolvable launch base URL as codex readiness
_codex_auth_unavailable_reason() detected a provider-routed launch only
via a profile or a non-openai model_provider override. On a Databricks-
wide machine the launch pins neither — omnigent defers to Codex's own
config.toml top-level model_provider default — so readiness fell through
to the auth.json check, found no openai credential, and falsely reported
needs-auth even though barecodexworks. That gated the Smart Routing
harness row off in New Chat (it needs both claude-native and codex-native
ready).Broaden the predicate to also count a resolvable launch base URL
(native_codex_launch_base_url(launch) is not None), which now resolves
the config.toml provider default. This only adds a ready case: an
explicit model_provider="openai" pin still returns None from that helper,
so a genuinely logged-out openai user still reports needs-auth. Readiness
now agrees with the launch resolver and the gateway-inference check.Co-authored-by: Isaac
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com- test: wrap the codex config.toml fixture under the line limit
Split the three identical model_providers config-toml f-strings across two
adjacent literals so each line stays under 99 chars, clearing the ruff E501
that failed pre-commit.Co-authored-by: Isaac
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com
(cherry picked from commitff20407a2d)
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com
Signed-off-by: Serena Ruan serena.rxy@gmail.com
Signed-off-by: Bryan Qiu bryan.qiu@databricks.com
Signed-off-by: Dhruv Gupta dhruv.gupta@databricks.com
Signed-off-by: Dhruv Gupta dhruv0811@gmail.com
Co-authored-by: Serena Ruan 82044803+serena-ruan@users.noreply.github.com
Co-authored-by: Dhruv Gupta dhruv0811@gmail.com
Co-authored-by: Harry Yao harryyao13@gmail.com
Co-authored-by: harry-yao_data harry.yao@databricks.com下载附件
-