DOC-6968 Omit a clients-example tab when that client lacks the requested step

[andy-stark-redis]

DOC-6968 — omit a clients-example tab when the client lacks the requested step

29 lines in one partial. It stops 63 panes site-wide from rendering a client's entire example file where a snippet belongs — the worst being 258 lines of Go (package, imports and several unrelated examples) on commands/scan.md.

This is the fix the rest of DOC-6968 was clearing the way for. The coverage work that had to land first is already on main (#3809).

The condition, and why each clause matters

step != ""  AND  step not in named_steps  AND  len(named_steps) > 0

The shortcode's else branch was conflating three situations. Only the middle one is a defect:

Case	Panes	Behaviour
step="" — explicitly empty	12	Unchanged. The whole file is the example (data-store.md)
File has other steps, not this one	63	Omit the tab — an authoring gap
File has no steps at all	25	Unchanged. Predates STEP markers; the whole file is still the closest thing to an example

Dropping the len > 0 clause would strip every client tab from commands/set.md and commands/get.md — 12 each — because those files have no STEP markers. That's a regression, not a fix, so the clause is load-bearing rather than defensive.

It deliberately tests membership in named_steps rather than $sliceable: a step that exists but whose line range is malformed or out of bounds should still fall through to the whole-file path, which is exactly the drift $sliceable was written to absorb.

The continue fires before any content is built, not just before the tab is appended — the legacy path mutates .Page.Store and inlines the file once per page, so building it for a tab we then drop would leave orphaned markup and a wasted copy of the file in the output.

Loud, not silent

Each omission emits a warnf naming page, client, step and set, so the gaps become a worklist rather than an invisible absence:

WARN  [tabbed-clients-example] ".../content/commands/scan.md": client "Go" has no step
      "scan1" in set "cmds_generic" — tab omitted. Add the step to that client's example
      file, or name the supporting clients in lang_filter.

Self-extinguishing — it drops to zero as the examples get written. CI runs plain hugo -d output with no --panicOnWarning, and the build still exits 0, so this surfaces the gaps without failing anything. Deliberately not warning on the 25 zero-step files, which would produce permanent warnings nobody can action.

Verification

Two builds of the identical tree, with and without the change:

	Before	After
Tab panes	2711	2648 (−63)
Legacy-fallback panes	100	37 (−63)
Legacy clone panes	23	0
Omission warnings	—	63

The 63 warnings correspond one-to-one with the 63 removed panes, and what remains on the legacy path is exactly the two protected classes (25 cosmetic + 12 by-design).

Structural checks, because a change like this fails quietly:

• All 594 tab groups survive with ≥1 pane — nothing renders empty.
• Selector options still match panes on the worst-hit group (scan1: 6 and 6, identical sets) — no tab button left without a pane.
• commands/incr.md untouched at 12 clients, thanks to the examples added in DOC-6968 Fix two node-redis step names, and add the missing incr step for nine clients #3809.
• commands/set.md / commands/get.md untouched, 12 client tabs each.
• data-store.md (step="") untouched.
• hashes.md hexpire keeps all 11 clients including Ruby; hexpireat keeps its 10 (Ruby excluded by lang_filter, so it never reaches this check).
• vector3/vector4 degrade to a CLI-only tab as agreed — a 1-pane shape that already exists elsewhere on the site (nredisstack/prob).

What this does not fix

The 63 panes stop rendering wrongly; the missing coverage is still missing. commands/scan.md drops from 12 client tabs to 5, the h* pages from 14 to 12, and del/expire/ttl from 12 to 7. The four cmds_generic scan steps are 28 of the 63 and remain the biggest lever — see the DOC-6968 status comment for the full per-page table. The warnf output is the worklist.

🤖 Generated with Claude Code

---

Note

Low Risk
Docs layout-only change with explicit guards for legacy no-step files; build still succeeds and tab groups retain at least one pane per verification notes.

Overview
Client example tabs in tabbed-clients-example.html now skip a language tab when the shortcode asks for a named step that client’s example file does not define, but that file does define other steps.

That stops the legacy whole-file path from dumping full multi-example sources (imports, unrelated snippets) into a pane meant for one command—63 panes site-wide, per the PR verification.

Unchanged behavior: empty step still uses the whole file; files with no STEP markers at all still get a tab (so pages like set/get keep all client tabs). Malformed or out-of-bounds ranges still use the existing whole-file fallback via $sliceable, not this omission rule.

Each omission logs a warnf with page, client, step, and example set. The loop continues before building tab content so legacy .Page.Store inlining does not run for tabs that are dropped.

^{Reviewed by Cursor Bugbot for commit 94cb5ab. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

DOC-6968

[github-actions]

🧠 Redis Memory

Found 5 related items from repository history (5 new this commit):

• PR — DOC-6968 Fix two node-redis step names, and add the missing incr step for nine clients 🆕

  DOC-6968 — two node-redis step names that never matched their sets

• PR — Fix oversized client-example pages: slice snippets, embed each file once (#3398) 🆕

  Fixes Streams data-type page is ~41 MB — the client-example partial embeds the whole source file on every step #3398.

• PR — DOC-6905 Include clients-example and jupyter-example code in JSON/Markdown output 🆕

  Problem

• PR — DOC-6954 Give the Client tools page its categories back 🆕

  Fixes DOC-6954. One deleted line.

• Commit — DOC-6968 Add the missing incr step to nine cmds_string client examples 🆕

  DOC-6968 Add the missing incr step to nine cmds_string client examples

Memory updated at 94cb5ab

[dwdougherty]

GAH! ZOMG! So many new warnings! OCD triggered bigly. 😲

That said, please proceed.

[andy-stark-redis]

GAH! ZOMG! So many new warnings! OCD triggered bigly. 😲

Thanks for the review @dwdougherty ! Yeah, I prefer to be clear of warnings too - next PR is all about fixing them :-)