Skip to content

fix(docs): restore Nightward-style homepage flow - #9

Merged
JSONbored merged 1 commit into
mainfrom
codex/docs-nightward-style-home
May 26, 2026
Merged

fix(docs): restore Nightward-style homepage flow#9
JSONbored merged 1 commit into
mainfrom
codex/docs-nightward-style-home

Conversation

@JSONbored

Copy link
Copy Markdown
Owner

Summary

  • restores the homepage to a Nightward-style VitePress flow instead of the custom oversized landing page
  • moves the Gittensor screenshot out of the hero and into a lower contextual section
  • reduces hero scale, keeps the MCP version widget, and preserves the dark Gittensor-adjacent visual language

What changed

  • replaced the custom giant homepage hero with VitePress home frontmatter and compact feature cards
  • added a compact start strip, path picker, audience table, and lower Gittensor context image section
  • removed the old custom hero/proof-panel CSS and tightened responsive typography
  • refreshed the Gittensor context image crop and fixed its intrinsic dimensions

Why

The previous homepage was too heavy: oversized text, an awkward hero image, and a flow that did not match the Nightward docs pattern. This keeps the same product positioning but makes the page scan like a docs site again.

Validation

  • npm run test:ci

Notes

  • No Playwright or frontend browser tests were added or run for this fix.
  • This does not change backend, MCP, npm package, or GitHub App behavior.

@JSONbored
JSONbored merged commit efec175 into main May 26, 2026
1 check passed
@JSONbored
JSONbored deleted the codex/docs-nightward-style-home branch May 26, 2026 17:12
@JSONbored JSONbored added the gittensor:bug Gittensor-scored bug fix — scores a 0.05x multiplier. label Jun 10, 2026
loopover-orb Bot pushed a commit that referenced this pull request Jul 14, 2026
…om maybePublishPrPublicSurface (#4607) (#5787)

`maybePublishPrPublicSurface` is 2,600+ lines, and #4607's stated cost is that "most cross-cutting
review-pipeline concerns live as unnamed inline blocks in one function rather than named,
independently-testable steps". This lifts one such block into a named step.

`derivePublicCommentMergeFacts` derives the five merge/disposition values the public PR comment
renders -- ciState, mergeStateLabel, mergeReadiness, heldForReview, neverClosed -- from the live CI
aggregate, the live merge-state refresh, the repo settings and the changed files. Its output IS
buildUnifiedCommentBody's input contract for those fields, so the seam is a real step rather than an
arbitrary cut. It is pure: the `incr()` metrics emit that sat in the middle of the block stays at the
call site, since it reads the gate conclusion the derivation never sees.

These flags exist so the COMMENT agrees with the ACTION the disposition planner takes -- heldForReview
for a guardrail-touching diff (#guarded-hold-comment), neverClosed for an owner/protected-automation
author (#8/#9). Until now they were only reachable by standing up a whole webhook delivery: the
renderer's own suite passes them in as hand-written literals, and the queue suites exercise them
transitively. They now have direct table tests, including the empty-changed-file-list fail-safe (an
unresolved file list HOLDS for review rather than claiming safe-to-merge), which had no test at all.

No behavior change: the 882 tests in the existing queue suites pass unchanged.
loopover-orb Bot pushed a commit that referenced this pull request Jul 24, 2026
…parse as the overflow check (#8477)

`resolveLinkedIssueHardRule` computed the overflow check from a fresh parse of the
PR's current `body` (`extractLinkedIssueNumbersWithOverflow`) but ran its per-issue
fact-fetch loop against a SEPARATE, caller-supplied `args.linkedIssues` array — a
field synced by an earlier body parse. If a contributor edits their PR description
to add a new closing reference between that sync and this evaluation, the overflow
check (fresh body) and the fact-fetch loop (stale list) disagree about which issues
are "currently linked", and the newly-added issue is silently never evaluated
against the configured hard rules until a later re-sync. That is a real
edit-driven detection-bypass window for a module whose whole purpose is "evaluate
every currently-linked issue's facts against the hard rules."

Fix: parse the body once and drive BOTH the overflow check and the fact-fetch loop
(and the empty early-out) from that single `extractLinkedIssueNumbersWithOverflow`
result, so the two can never disagree. Direction chosen (per the issue's second
option): keep the `linkedIssues` parameter for call-site compatibility but stop
reading it — a surgical change that leaves `src/queue/processors.ts` untouched; the
parameter is documented as retained-but-not-the-source-of-truth.

Add a regression test: a body linking #9 with a stale `linkedIssues: [1]` now
evaluates #9's facts (previously it fetched only #1, 404'd, and reported the wrong
issue). Existing tests that relied on `linkedIssues` driving the fetch were updated
to put the same references in `body` (their intent is unchanged); all still pass.

Closes #8354
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

gittensor:bug Gittensor-scored bug fix — scores a 0.05x multiplier.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant