Revert the hidden section tabs, and revise all three docs - #68
Merged
Conversation
Wrong call, reverted at the author's direction. Removing the tab does not read as "this section is still filling up" — it reads as "this section was removed", and it left the nav with four items huddled beside the logo against a full-width hero, visibly off-balance. `Paper Summaries` and `Tech Guides` get `main_nav: true` and their original nav_order back, and lose `sitemap: false`. The empty-state line from _includes/category-posts.html is the honest signal that they are not filled yet; the tab is the statement that this blog writes those kinds of posts at all. Keeps the other half of the reverted commit — the <h2> that repeated each page's own title stays gone, since that was a rendering defect and not part of this. For the record on why they are empty: the automated PRs that would have filled them were closed unmerged — #37/#38 (MAI-Thinking-1) and #39-#43 (Argo CD). The MAI-Thinking-1 draft still exists on origin/paper-reviews/mai-thinking-1-… if it is ever wanted; as written it would not pass the writing gates (emoji headings, 평서체 instead of 입니다체, ~2.1k words against a 6k target).
Two corrections, both mine:
1. The MCP tag was renamed to Model-Context-Protocol. Renaming is removing, and
the approved scope was *adding* topic tags — so this was out of scope, and it
killed a live anchor: /tags/#mcp stopped existing in production while the PR
claimed "nothing is removed, so no live /tags/ anchor moves". MCP is back
alongside the long form, so the anchor resolves and the two posts still
connect. Whether to fold them later is a separate decision.
2. The docs carried post counts ("Paper Reviews (28편)", "35편 전부", "220 of 264
tags") and work history (PR numbers, a branch name, "this was once done and it
was wrong"). Counts are false the moment a post lands, and history belongs in
git. Both replaced with the rule and its reason: a tag phrased as one paper's
contribution can only apply to that paper, related_posts.rb needs a shared
tag, therefore a post tagged only that way gets no Related reading block.
search.json's size is stated as "grows linearly with post count" with the
command to measure it, instead of a figure that goes stale.
Applies the "no stateful facts, no work history" rule to all three documents, not just the passages edited most recently, and reconciles every claim against the code. Corrected, in rough order of how much damage each could do: - `_sass/_syntax.scss` was described as a generated file to be recreated with `rougify style monokai.sublime`. It is hand-edited: two Monokai colours were raised to clear 4.5:1 against the code background, with the original values and ratios in a comment at the top. Following the old instruction would have silently reverted the contrast fix. Now flagged do-not-regenerate. - Both READMEs still listed `archive` in the layout chain, which no longer exists. - The pre-push command block claimed "the same three checks" while listing four, built into `_site/`, and sat directly above a note explaining why you must not build into `_site/`. It now matches CI exactly and builds to /tmp/site-verify. - `test/` was described as covering three plugins; it covers all five, one file each. Stated that way so it stays true. - The tags section had been inserted into the middle of "Categories drive the URL", orphaning the output-path example under a heading about tags. Moved back, and the URL-immutability warning now sits with it. - A cross-reference pointed at a `TAG_WEIGHT` discussion that exists only in the plugin, not in this document. Now points at the plugin. - The plugin roll-call in three places named three of the five plugins; `related_posts.rb` also owns the prev/next links now. De-stated: post counts, tag counts, search.json byte sizes, "html-proofer caught 128 of these", "18 posts were affected". Each is either false after the next post lands or already false. Where the number carried the point, the point is now made structurally — search.json "grows linearly with post count", with the curl command to measure it. De-historied: "this was once X and it was wrong", closed-PR numbers, a branch name, "the repo has already been migrated". Rules and reasons kept, narrative dropped — git holds that. Also fixed: the search section documented `_config.yml` as unable to control the index only in passing, as a parenthetical about a block that had been deleted; it is now a standing warning, since that is the mistake it prevents. And the known search limitation (index is the full body, the card shows a 40-word head) is stated once in §5 and listed in §9 rather than half-explained in both.
bits-bytes-nn
changed the title
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.
Reverts the nav/sitemap half of
chore: hide the two empty section pagesat theauthor's direction.
Hiding a tab because its section is empty was the wrong read. A missing tab says
removed, not filling up — and it left four items huddled next to the logo
under a full-width hero, which is what prompted "탭이 묘하게 어색해졌는데".
paper-summaries.md/tech-guides.md:main_nav: trueand the originalnav_orderrestored,sitemap: falsedropped.docs/tech-doc.mdupdated back, now recording whyempty sections keep their tabs so this does not get "cleaned up" again.
<h2>that repeated each page's own title stays removed — that was aseparate rendering defect.
Why they are empty, for the record: nothing was lost here.
Paper Summarieshas never had a post on
main. The automated PRs that would have filled bothsections were closed unmerged — #37/#38 (MAI-Thinking-1) and #39–#43 (Argo CD).
The MAI-Thinking-1 draft is still on
origin/paper-reviews/mai-thinking-1-…; aswritten it would fail the writing gates (emoji in headings, 평서체 rather than
입니다체, ~2.1k words against the 6k Paper Summaries target).
Verified: 82 tests / 136 assertions, clean build, html-proofer successful, all 16
validate-site checks passed, nav renders six tabs, both pages back in
sitemap.xml.