Repository navigation
Conversation
`cargo check -p apr-cli --no-default-features --lib` failed with 23 errors: items that use realizar or axum, or call inference-only helpers, were not behind `cfg(feature = "inference")`. The default build hid them, and the workspace-wide no-default check unifies `inference` back into apr-cli. - Gate `serve::route_index`, the `Devices` dispatch arm, the runtime thinking-ON leg, the four thinking-budget helpers and chat's realizar import. - `forced_accelerator_refusal` without inference refuses any forced accelerator by name; its realizar-table tests now need inference, and a no-inference module pins the new behaviour. - The chat banner names its template through `banner_template_family`, which says plainly when the build has no detector. - The no-inference `ChatSession::new` stub takes `force_cpu`, like its callers. The default build is unchanged. A planted undefined call in a no-inference-only arm turns the check RED (rc 101) while the default build stays green; reverted, it is green again. Clippy `-D warnings` is clean in both feature sets. Refs #4972 Agent: la-0.72b Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> (cherry picked from commit 9447c19)
|
§13.11 rung 1 — quorum shadow verdict Shadow mode: this records a verdict and merges nothing. A refusal |
…rrent lines (Refs #4972) The inference gates in this branch moved dispatch_analysis.rs, so G2.1 freshness asks for the ledger rows that cite it. Each of the 14 rows now names the line of its anchor on this head: the eval task match arms, the chat accelerator check, and the parity per-op dispatch. Most of these citations had already drifted on main; no row changes anything but its line number. Agent: la-0.72b Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
Author
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
1 participant
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.
apr-clidoes not compile with--no-default-featuresonmain:cargo check -p apr-cli --no-default-features --libstops with 23 errors (E0061, E0277, E0282, unresolved items), in chat, golden-output, output-verification, run, serve and dispatch code that reachesrealizarwithout theinferencegate.This puts the gate back where it belongs and adds no new behaviour under default features:
realizarchat-template import, thinking-on budget helpers, the runtime thinking-on leg,serve::route_index, theDevicesdispatch arm) onfeature = "inference"not(inference)fallbacks: the banner names no template family, and a forced accelerator is refused for every architecture (there is no forward on any device)not(inference)ChatSession::newstub's signature with its inference twinMeasured on
main11f844a plus this diff:cargo fmt --check,cargo checkandcargo clippy -D warningswith and without default features all exit 0. Without the diff, the no-default check exits 101 with 23 errors.Not yet measured: the row-4 tests.
cargo test -p apr-cli --no-default-features --libdoes not compile on this head: 70 errors, none in an item this diff touches. Test code elsewhere in the crate assumesinference, for example theTunecommand and the golden-output helpers. Until the lib test target builds withoutinference, no job compiles or runs row 4. Making that target build is follow-up work, not part of this PR.Refs #4972
keep-open: #4972 is the LAB check that apr-cli builds without inference; it closes once that check runs green at night, not on this merge.
🤖 Generated with Claude Code