compute: publish maintained indexes into the sharing registry - #38389
Draft
antiguru wants to merge 1 commit into
Draft
compute: publish maintained indexes into the sharing registry#38389antiguru wants to merge 1 commit into
antiguru wants to merge 1 commit into
Conversation
antiguru
force-pushed
the
mh/interactive-04-publish
branch
from
August 21, 2026 11:23
d31414b to
d9ddff4
Compare
antiguru
force-pushed
the
mh/interactive-04-publish
branch
2 times, most recently
from
August 21, 2026 13:42
d754ecc to
1303035
Compare
antiguru
force-pushed
the
mh/interactive-04-publish
branch
from
August 21, 2026 14:31
1303035 to
dffb179
Compare
Both export paths now publish their `oks`/`errs` arrangements into the per-process registry when the runtime's role publishes, and the two re-export arms register the alias so a read waiting on the aliased id's seal is woken by the original's publisher. Logging indexes publish the same way, gated strictly on `Maintenance`: an interactive runtime reads maintenance's slot, and its own copy would clobber it, while `Solo` has no registry peer at all. `ComputeRuntimeRole::Interactive` stops being test-only. Nothing constructs it yet, but `publishes()` has to name it, and `pub mod server` keeps the variant reachable so dead-code analysis is satisfied without an attribute. The stale `owns_process_globals` note claiming every constructible role owns the globals goes with it. Carrying the role and the registry to the render path is what the rest of this change is: `Config` and `Worker` gain both, `ComputeState` stores them and exposes `role()`, and clusterd builds one registry per process. Per process, not per runtime, because a reader on one runtime looks up the slot a publisher on another filled. No behavior change. `Solo` is the only role anything constructs and it does not publish, so every added block is skipped and no dataflow gains an operator. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
antiguru
force-pushed
the
mh/interactive-04-publish
branch
from
August 21, 2026 17:54
dffb179 to
7d70e54
Compare
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.
Fourth of eight PRs splitting #37770. Stacks on #38388. Tracked by CPU-215.
Both export paths publish their
oks/errsarrangements into the per-process registry when the runtime's role publishes, and the two re-export arms register the alias so a read waiting on the aliased id's seal is woken by the original's publisher. Logging indexes publish the same way, gated strictly onMaintenance: an interactive runtime reads maintenance's slot and its own copy would clobber it, whileSolohas no registry peer.ComputeRuntimeRole::Interactivestops being test-only, becausepublishes()has to name it.pub mod serverkeeps the variant reachable, so no#[allow]is needed even though nothing constructs it yet.No behavior change:
Solois the only role anything constructs and it does not publish, so every added block is skipped and no dataflow gains an operator. That is also why no goldens move here. They move in the last PR of the stack, which turns the flag on in CI.