Repository navigation
terminal output modeling for CI/interactive - #5887
Merged
Merged
Conversation
…e] read floor spam) The [file] read / [file] write / [rest] / [hermetic:mock] trace sites were unconditional eprintln!s — [file] read alone floods the CI floor log with hundreds of source-file reads (doubled by the within-run double-resolve). #5874's deferred slice: route them through the .dag output policy instead of suppressing in Rust. - gunbc.output_policy gains ChannelPolicy + resolve_channel_policy(verbose, quiet): it enumerates the channels and calls channel_decision once each, so the seed can evaluate the whole policy in one .dag call. channel_decision stays the single authority for channel x verbosity -> decision. - The seed (cli_run::install_output_policy) evaluates resolve_channel_policy from the .dag authority at the current cli_verbosity and installs the per-channel verdicts into a process-global (set once before discovery threads spawn). It only TRANSPORTS the evaluated OutputDecision variants across the boundary -- no Rust re-derivation of the policy logic. - v1_interpreter routes the four trace sites through trace_emit(channel, line): reads/mocks -> Instrumentation (Verbose-only), writes/rest -> ShellTrace (condensed at Normal). render_shell_trace now CONSUMES output_decision(ShellTrace) too, dissolving the cli_verbosity mirror #5860 left in it. - claim_executor + claim_batch install the policy at startup. Unset (other bins) falls back to Full = pre-funnel behavior, so nothing else regresses. Verified by execution: a WET filesystem_read witness prints 0 [file] read lines at Normal and 1 at Verbose, and still passes. output_policy_witness covers resolve_channel_policy. fmt + clippy --all-targets clean. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…5579) The WIP grouping model added // doc-comments to std.render / extdeps.render.surface / the witness — but .dag has had no comment syntax since #5579, so each // parsed as "expected item declaration", failing std.render's parse and cascading "unresolved type Frame/Span/…" through extdeps.render.terminal. That reds dsl_compile_clean_gate (the CI floor batch-1 gate) the whole tree over. Comments removed; the grouping seam stands: - std.render: Block = Leaf | Group { title, frame } + Document — the target-agnostic collapsible-section model (one concept, every surface). - extdeps.render.surface: RenderTarget = PlainTerminal | GithubActions dispatch + GroupSyntax authority; GitHub Actions -> ::group::/::endgroup:: (consuming log_annotations), plain terminal -> "── title ──" header. detect_target keys off GITHUB_ACTIONS, mirroring terminal_capability's env->cap. Verified: render_surface_witnesses passes by execution via the floor resolve path (resolve_entry_graph + run_claim); whole-tree `gunbc compile --target rust` emits 614 modules / 0 diagnostics. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…tch (option B) Wire the grouping seam into the floor: claim_executor installs the per-target GroupSyntax from extdeps.render.surface.resolve_group_syntax(GITHUB_ACTIONS) at startup, then brackets each batch's parallel walk in a host-effect group. The [file]/[rest]/[shell] trace lines (stderr, emitted live from worker threads) fold INSIDE the group; the PASS/FAIL summary (stdout) is deferred until AFTER group_end so it stays the scannable, ungrouped signal. One open/close on the main thread spans the whole batch, so it is sound under spawn_width > 1 (threads never interleave markers). Gated on host_trace_grouping_active() — no brackets when every trace channel is Suppressed (Quiet), so empty groups stay out of the log. Verified by execution (minimal 1-batch plan over filesystem_read witness): - GITHUB_ACTIONS=true: `::group::batch 1 host-effects` / `[file] read …` / `::endgroup::` then `✓ PASS` AFTER the close. - plain terminal: `── batch 1 host-effects ──` header, no workflow commands. - Quiet: no grouping at all. render_surface_witnesses (incl. resolve_group_syntax) green; fmt + clippy --bins clean. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
claude-opus-4-7 REQUEST_CHANGES on #5887: an auto-WIP commit swept in dead substrate — TypeEnv.parents + lookup_binding_in_scope / enumerate_bindings_in_scope in src/v1/04_env.dag with zero callers (~10 constructor sites all pass parents: []), unrelated to this PR's terminal-output topic and with no named dissolution trigger (DESIGN §6). Reverted src/v1/{04_env,04_infer,04_resolve,05_emit,05_emit_rust}.dag to origin/main; the render-surface grouping work is untouched. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.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
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.
Auto-opened by session-dashboard for session
bright-ibex-44.Pushing to
host-trace-funneladvances this PR.Worker attestation
Before flipping this PR to ready for review, confirm each item:
npm test,cargo test) and the result.Closes #Ndirective.Summary
TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.
Test plan