Yezur Wiki:Daily routine/Report
Working report of the Alompan daily routine (how it works). Every finding lives as one row below until it is resolved: set a row's Status to give a verdict, and the next run carries it out and clears the row. Nothing on the wiki is changed except what is explicitly approved or amended.
The possible statuses:
pending— awaiting review; no action is taken (the default)approvedoraccepted— carry out the Suggested fix, as writtenamended— carry out the fix described in Notes insteaddeferred— real, but not now; moves to Dismissed flags and is not re-flaggedrejected— not an issue; moves to Dismissed flags and is not re-flaggedto to-do— real issue that should be marked on User talk:Yezur's to-do list instead for human interventionfixed— real issue, but was a simple fix for Yezur, or has by other means already been taken care of; confirm and clear
To revive a dismissed flag, move it back into the table as pending.
Flagged items
| # | Flagged | Type | Severity | Page(s) | Issue | Suggested fix | Status | Notes | |
|---|---|---|---|---|---|---|---|---|---|
| 47 | 2026-07-26 | structural | medium | MediaWiki:Sidebar, Yezur Wiki:Main Page, Yezur Wiki:Conventions | Two namespaces are documented as part of the wiki's structure and have never held a single page. Meta: — namespace 4, this wiki's project namespace — is named as live out-of-frame space in six places across three pages. Yezur Wiki:Conventions gives it a row of its own in the §Namespaces table ("Meta: — Out-of-frame — Design notes and rationale"), names it in §Where out-of-frame content goes ("It belongs on Talk pages, in the Meta: namespace, or in this Yezur Wiki: namespace"), and names it again in the same section's rule of thumb ("except the Meta:, Talk:, and Yezur Wiki: namespaces"). Yezur Wiki:Main Page lists it first under §Behind the scenes as "about the Yezur project itself; the place for Earth comparisons and design notes" and repeats it under §Conventions. And MediaWiki:Sidebar gives it a permanent navigation slot on every page of the wiki, as Project docs → Special:PrefixIndex/Meta:. But Meta: and Meta talk: hold 0 pages, and list=allrevisions returns nothing for either — no page has ever existed in them, so the sidebar link every reader sees leads to an empty index. The same is true of Help:, which Yezur Wiki:Main Page lists as the third out-of-frame namespace ("how-to guidance for editors") and which Yezur Wiki:Conventions's table omits entirely: also 0 pages, also no revision ever. Everything the two are described as holding is in fact on Talk pages and in the six-page Yezur Wiki: namespace. |
A decision, then mechanical edits. Say whether Meta: is still wanted as a namespace distinct from Yezur Wiki:. (a) If it is retired — the reading that fits what has actually been built, since every out-of-frame page written so far has gone to Yezur Wiki: or to a Talk page — then drop its row from the §Namespaces table, drop it from the two sentences in §Where out-of-frame content goes and the rule of thumb, drop the §Behind the scenes bullet and the §Conventions mention, and repoint the sidebar's Project docs link at Special:PrefixIndex/Yezur Wiki:, where the six project pages actually are. (b) If it is still wanted, it needs at least one page so the sidebar link is not a dead end, and the two project pages should be made to agree on what belongs there rather than in Yezur Wiki:. Either way the Help: mismatch wants settling in the same pass, since the two project pages disagree about whether it is part of the scheme at all. One caveat on scope: the sidebar sits in the MediaWiki: namespace, which needs the editinterface right — queried directly this run, Alompan does not hold it — so that one line needs an administrator even once approved. The two project pages the routine can do itself. |
amended | New 2026-07-26, from the first full read of the Module, Talk, User, MediaWiki and project namespaces. Same species as #45 — documentation that was never followed through — but load-bearing in a way #45 is not: this is the only navigation link on the wiki that leads somewhere empty, and it is on every page. Recorded as not a finding beside it: Yezur Wiki talk:, Dictionary talk: and Map talk: are empty too, but no page claims otherwise, and a talk namespace fills itself when there is something to discuss.
Re-confirmed 2026-07-28: | |
| 50 | 2026-07-26 | structural | low | Module:Auphen | An unreachable branch in the affix path, and an else-clause form that is silently dropped along with it. In Engine:handle_affix, the change slot is handled by two branches — one for a change containing #, and one for an empty change carrying only an @-index or span, which is what a pure indexed operation parses to, since the index-stripping helper removes the @n and leaves the change empty. The else slot copies that same two-branch shape, but wraps it in if ELS ~= then — so its second branch tests ELS == inside a block that has already required ELS ~= , and can never run. The same guard also skips the else clause's affix_prep call a few lines earlier. The consequence is not only dead code: an affix rule whose else clause is an @-index alone falls through both guards, the else never applies, and the rule returns the word unchanged with no warning — where the identical form in the change slot works. It is the same class of defect the talk page records as already fixed: D1 reads "the ELS @-index is parsed and used as a fallback instance filter (the JS assigned it to the CHG index, so ELS indices were dead)" — and in the affix path they are still dead. |
flag-only. This is a change to a live rule engine with an off-wiki test suite behind it, and the routine will not edit module code on the strength of its own reading. The shape of the fix is plain — hoist the ELS ~= guard so the index-only case reaches affix_prep and apply_affix_logic, mirroring the change slot exactly — but it wants a proving test case in that suite first, and the module is Khurouan's. Raised so it is recorded rather than found again later. Worth settling one question before any code moves: whether an index-only else clause on an affix rule is a form the rule notation actually admits. If it is not, the right fix is the opposite one — delete the dead branch and say in a comment why the else slot needs only the # case. |
rejected | New 2026-07-26, from the first full read of Module:Auphen — at 38 KB the largest page on the wiki outside this report. The only code defect the sweep found: the UTF-8 codepoint layer, the case tables, the escape layer, the category-explosion guard (opts.explodeCap, default 5,000,000, exactly as the talk page states) and the IPA transposition all check out against their documented behaviour, as do the two data pages' category sets.
Re-checked 2026-07-28, unchanged.
Re-checked 2026-07-30, unchanged — Module:Auphen's last revision is still that of 2026-07-18.
Re-checked 2026-07-31, unchanged — Module:Auphen's last revision is still that of 2026-07-18.
Re-checked 2026-08-03, unchanged — Module:Auphen is still at the revision of 2026-07-18, and none of the 851 content edits this window touched the Module namespace except Module:Languages and the new Module:Translations. | |
| 51 | 2026-07-26 | naming | low | Yezur Wiki:Conventions, Category:Countries | The worked example for place-qualified category names matches neither the wiki's categories nor the same page's own next paragraph. Yezur Wiki:Conventions gives the form twice as "Polities of Andusia" — in §Categories ("a country in Andusia sits in a 'Polities of Andusia' category that is a child of both 'Andusia' and 'Polities'") and again in §Naming ("Keep narrower place qualifiers, such as 'Polities of Andusia'"). Two paragraphs after the first, the same section writes "Former countries in Andusia". Every place-qualified category that exists uses in, and none uses polities: Category:Cities in Hertia and Category:Ancient cities in Hertia, beside Category:Countries and Category:Former countries. The split runs deeper than the connector. The category side of the wiki says country — Category:Countries, one member, Engil — while the template and Cargo side says polity: Template:Infobox polity declaring and storing into the Polities table. Only Template:Infobox former country and Category:Former countries agree with each other across the divide. | Two parts, one mechanical and one a decision. (a) The connector is settled both by practice and by this page's own later sentence: make both worked examples read "in". Mechanical. (b) Whether the house word is country or polity is yours to say — and it is cheap now: Category:Countries holds one page and the Polities table holds none, so nothing has to be moved either way. If country wins, the examples become "Countries in Andusia" and Template:Infobox polity is the odd name out; if polity wins, Category:Countries wants renaming to match, and the routine holds the move right so it could do that on approval. Leaving both words in use is defensible too — but then the examples on the convention page should at least use a form that exists somewhere. |
amended | New 2026-07-26, from re-reading Yezur Wiki:Conventions in full beside all 24 category pages. Third row touching §Naming, with #28 and #45(a) — if more than one is approved, make them in one edit. Recorded as not a finding: neither Category:Andusia nor Category:Polities exists, though the worked example implies both — those are categories not built yet rather than an inconsistency, and the same page asks that new categories be cleared with the operator first. (Named in italics rather than linked, so this report does not itself create two more wanted categories.)
Re-checked 2026-07-28. Now the only open row touching §Naming — #28 and #45(a) were carried out this run, so part (a) here is a standalone one-line edit whenever you want it. The section is otherwise unchanged by those edits: #45(a) removed the retired Family A bullet and touched nothing else, and both "Polities of Andusia" wordings are still live, in §Categories and §Naming. Part (b) is cheaper than when filed: Category:Countries still holds one page and Polities still holds 0 rows. Re-checked 2026-07-30, unchanged: both "Polities of Andusia" wordings are still live, in §Categories and §Naming, and Category:Countries still holds one page. Part (b) is cheaper again than when last checked: the Polities Cargo table no longer exists at all, so nothing on the wiki stores the word polity except Template:Infobox polity's own name and documentation — and that template is now transcluded by no page, so whichever word wins, nothing has to be migrated. Re-checked 2026-07-31, unchanged: Yezur Wiki:Conventions was not edited this window, both "Polities of Andusia" wordings are still live, Category:Countries still holds one page, and Template:Infobox polity is still transcluded by no page at all. Re-checked 2026-08-03, unchanged: Yezur Wiki:Conventions was not edited this window, both "Polities of Andusia" wordings are still live, Category:Countries still holds one page, and Template:Infobox polity is still transcluded by no page at all — it is one of the three entries on Special:UnusedTemplates. Operator verdict 2026-08-04 (via Thiorosan): amended. (a) approved: in over of. (b) country is the house word, preferred over polity — the worked examples become "Countries in Andusia". | |
| 52 | 2026-07-26 | factual | low | Yezur Wiki:Conventions, Template:On the map | The convention set's list of map templates omits the one that articles actually use, and lists four that none do. Yezur Wiki:Conventions §Maps reads, in full: "Link to the Yezur map with {{Node}}, {{Way}}, and {{Relation}} (by object id), or {{Coords}} / {{Coordslink}} (by coordinates)." All five exist, and the transclusion counts show how they are used: {{Node}} by seven articles, and {{Way}}, {{Relation}}, {{Coords}} and {{Coordslink}} by no page at all. Meanwhile Template:On the map is not mentioned, though it is in use on Alompe and is the only one of the six that behaves structurally differently: it renders through MediaWiki's page-status-indicator mechanism, placing a 🌐 On the map link beside the page title rather than inline. Its own documentation carries two rules that are recorded nowhere else — "by convention, put it on the first line of the article", and "If the template appears more than once on a page, the last use wins" — and it states its own scope: points on the map, "not continents or countries, which cannot be reduced to a single point". |
Add one bullet to §Maps, restating only what Template:On the map's own usage note already says: that a subject resolving to a single point on the map takes {{On the map|lat|lon}} on its first line, which places a title-side indicator rather than an inline link and takes effect once per page; and that continents and countries, which cannot be reduced to a point, do not take it. Nothing new is asserted. Mechanical. The four unused templates are not proposed for removal — a documented template awaiting its first use is not a defect. |
approved | New 2026-07-26, from reading all 25 templates against the convention set. If #28, #45 or #51 are approved this can go into the same edit to Yezur Wiki:Conventions, though it falls in §Maps rather than §Naming. Related but deliberately not flagged: Talk:Alompe records that article's map node id (60111) as an open thread "until a settlement infobox exists to hold it" — a new infobox is a decision the convention set reserves to the operator, so the note is left where it is.
Re-checked 2026-07-28, unchanged; §Maps still lists the five and omits Template:On the map, whose transclusion count rose from one page to one page (Alompe only). Both edits to Yezur Wiki:Conventions this run fell in other sections, so this remains a single clean bullet insertion. Re-checked 2026-07-30, unchanged; §Maps still lists the same five templates and still omits Template:On the map, whose transclusion count is still one article (Alompe). The 2026-07-30 edit to Yezur Wiki:Conventions fell in other sections, so this remains a single clean bullet insertion. Re-checked 2026-07-31, unchanged; §Maps still lists the same five templates and still omits Template:On the map, whose transclusion count is still one article (Alompe). Re-checked 2026-08-03, unchanged; §Maps still lists the same five templates and still omits Template:On the map, whose transclusion count is still one article (Alompe). Operator verdict 2026-08-04 (via Thiorosan): approved. | |
| 53 | 2026-07-27 | structural | medium | File:Tacchoan-subfamilies.png, Tacchoan | The wiki's first image carries an out-of-frame attribution strip, and its colour key cannot distinguish two of the four branches it exists to distinguish. The file was uploaded on 2026-07-27 and placed in Tacchoan §Members; it is a 1283×785 map export, and both defects were read out of the pixels rather than inferred. (a) The attribution strip. A pale bar carrying the mapping software's "© OpenStreetMap contributors" credit is baked into the saved image at its bottom-right corner — rows 778–784 across columns 1159–1283, so the frame's own bottom edge cuts the line of text in half. It renders at full size in the article, where Yezur Wiki:Conventions §Namespaces holds that "everything a reader or editor sees is in-character". It is the most reader-visible break of frame currently on the wiki, and the only one that is not text. (b) The colour key. The caption assigns yellow twice — "yellow for the Kipoa languages, and a small yellow area to denote the Usy language" — and the image bears this out: two separate yellow fields, an eastern strip and a southern island, in the same hue. A reader cannot tell which is which, on a map whose whole purpose is to separate the four branches. Purple (north), green (centre) and the two yellows do match the four bullets' stated geography, and red does sit off to the southwest, away from Taccho. | Three parts, and only the third is mechanical. (a) is flag-only: a corrected export is the map's business, not the routine's, and a straight crop of the bottom seven pixels is not a safe substitute — mapped land reaches the bottom edge across the left of the frame (77–108 non-sea pixels in every one of rows 778–784), so cropping would clip the southwestern island chain. There is also a question the routine cannot answer: whether any of the framed data is OpenStreetMap-sourced, in which case the credit is a licence condition and must be re-placed rather than removed. (b) is likewise flag-only — giving Usy its own colour needs a new export. (c) is mechanical and can be done without touching the file: the caption opens on "Red is referring to the Herta-Olaic family" without saying that Herta-Olaic is not a Tacchoan branch at all but the neighbouring family, shown for contrast; one clause fixes that, asserting nothing that Herta-Olaic and Taccho do not already say. | deferred | New 2026-07-27. Verified against the file itself, not the caption: the strip's extent was measured, and the two yellow fields confirmed as the same hue. #54 covers the description page rather than the image. If (c) is approved it can go into the same edit as #55, #56 and the mechanical half of #57, which all touch the same six lines of Tacchoan.
Re-checked 2026-07-28, unchanged. Re-checked 2026-07-30, unchanged. Re-checked 2026-07-31, unchanged — File:Tacchoan-subfamilies.png is still at the revision of 2026-07-27 and Tacchoan at that of 2026-07-30. Re-checked 2026-08-03, unchanged — File:Tacchoan-subfamilies.png is still at the revision of 2026-07-27, is still the wiki's only file, and Tacchoan is still at the revision of 2026-07-30. New this run: #67 proposes rewriting the same §Members section to add seven Tacchoan languages; if it and (c) here are both approved they belong in one edit, together with #55, #56 and the mechanical half of #57. Operator verdict 2026-08-04 (via Thiorosan): deferred — the operator will upload a better export in time. The licence question is closed: the map is not OSM-derived, it only uses the OSM software, so no credit is required. | |
| 54 | 2026-07-27 | structural | medium | Yezur Wiki:Conventions, File:Tacchoan-subfamilies.png | The wiki now holds files, and has no conventions for them. Until 2026-07-27 the site statistics read images 0, and the last report recorded the unused-images check as clean on the grounds that "the wiki holds no files at all". That is no longer true: there is one file, and nothing governs it. Yezur Wiki:Conventions has no §Files — its eleven sections cover voice, namespaces, naming, formatting, categories, infoboxes, open threads, maintenance, maps and sourcing, and say nothing about what a file description page carries, whether files are categorised, or how a source or licence is recorded. The consequences are already visible on the one file that exists: it is the sole entry in Special:UncategorizedImages; no File: namespace category exists for it to join; and its description page is the article caption copied verbatim — the same colour key, word for word, in two places that nothing keeps in step, so approving any caption change (see #53) silently puts them out of sync. |
flag-only on the substance. What a file description page should carry is a convention decision, and the convention set reserves those to the operator. Three questions settle it: whether files are categorised at all (and if so, whether the category is in-character like every other, per §Naming); whether a source line is recorded, given that these are exports of the map and the question of licence raised in #53; and whether a description page may restate an article caption or should point at it. Once decided the mechanical work is small — one category, one short description — and the routine can carry it out. | amended | New 2026-07-27, from the first upload in the wiki's history. Deliberately not proposed here: any particular category name, or a licence template. #14 (deferred) is the neighbouring case for pages — uncategorised front pages, where the operator accepted the finding but had no decided approach yet; this is the same shape one namespace over, and may want the same answer.
Re-checked 2026-07-28, unchanged.
Re-checked 2026-07-30, unchanged: still one file on the wiki, still the sole entry in Special:UncategorizedImages, still no | |
| 55 | 2026-07-27 | voice | low | Tacchoan | Four wordings in the new §Members text sit below the encyclopedic register. Yezur Wiki:Conventions §Editorial voice asks for "neutral and third-person, definition-first, like any good reference work". (a) The caption describes the key rather than the subject: "Red is referring to the Herta-Olaic family, while green is the Esselis branch, purple the Rumesselis branch..." — a reference work captions what the map shows, not the act of matching colours to names. (b) "Esselis, the one spoken by the highest amount of population" — amount takes a mass noun, and "the one spoken by" is a conversational stand-in for a plain superlative. (c) "a subfamily that is spoken in the south-east of the continent, as well as spoken on small areas of Lovvocia" — spoken twice in one sentence, and on where the sense is in. (d) "subdivisible into four categories" — category has a working meaning on this wiki, and the same six lines call these things branch or subfamily three times over. | Mechanical, and asserts nothing new — every fact stays exactly as written, including which colour marks which branch and which branch sits where. (a) recast as a description: the four Tacchoan branches, named with their colours, and Herta-Olaic identified as the neighbour (see #53(c)). (b) → "the most widely spoken of the four, in the centre of the continent". (c) → "spoken in the south-east of the continent, and in small areas of Lovvocia through colonial ties". (d) → "divided into four branches". | approved | New 2026-07-27, from reviewing the day's one article edit against the convention set in full. Not flagged in the same passage, and recorded as passing: the lead's "one of the five major Yezuri language families" is the reflexive planet-qualifier already carried by #38, which lists Tacchoan among its pages; the new Kipoa bullet's colonial claim is confirmed by Lovvocia, which independently records Tacchoan languages "in smaller areas — particularly the central region adjoining Vetsia"; and "the continent", used three times unnamed, resolves correctly to Taccho from the first sentence.
Re-checked 2026-07-28, unchanged. Re-checked 2026-07-30, unchanged. Tacchoan was edited this window (2026-07-30, rev 1618), but only to replace the empty Cargo member query in §Members with the plain sentence "No member languages are currently listed."; all four wordings are untouched. Re-checked 2026-07-31, unchanged — Tacchoan was not edited this window. Re-checked 2026-08-03, unchanged — Tacchoan was not edited this window, and all four wordings stand. See #67 for a second, larger edit now proposed to the same section. Operator verdict 2026-08-04 (via Thiorosan): approved, all four. | |
| 56 | 2026-07-27 | factual | low | Tacchoan | Tacchoan calls one of its own branches an isolate, against the wiki's own definition of the word. The fourth bullet reads "Usy, an isolate subfamily spoken on the island of Usings". Category:Language isolates defines the term on this wiki, in full: "Languages with no demonstrated kinship to any language family." Its one member, English language, uses it in exactly that sense — "It is a language isolate: it has no demonstrated kinship with any of the major language families". A branch of Tacchoan has demonstrated kinship by definition: that is what makes it a branch. As written the sentence claims both at once. | Mechanical, a two-word swap. "an isolate subfamily" → "the family's only single-language branch", which is what the passage means and what the caption already implies by naming "the Usy language" in the singular. Nothing else in the bullet changes. | amended | New 2026-07-27. The distinction matters beyond the wording: Category:Language isolates is a live category with a member, so leaving isolate in place invites Usy into it when the article is written, which would put it in two mutually exclusive places at once.
Re-checked 2026-07-28, unchanged. Re-checked 2026-07-30, unchanged — the bullet still reads "an isolate subfamily", and Category:Language isolates still has exactly one member, English language. Re-checked 2026-07-31, unchanged — the bullet still reads "an isolate subfamily", and Category:Language isolates still has exactly one member, English language. Re-checked 2026-08-03, unchanged — the bullet still reads "an isolate subfamily", and Category:Language isolates still has exactly one member, English language. Sharper than when filed: List of languages now names seven Tacchoan languages (see #67), so the family whose branch is called an isolate here is no longer an empty one. Operator verdict 2026-08-04 (via Thiorosan): amended — not the suggested wording. Usy is isolated from the other branches, but it is part of the family; find a proper way of conveying exactly that. | |
| 57 | 2026-07-27 | structural | medium | Tacchoan, Esselis, Rumesselis, Kipoa, Usy, Usings | Five subjects were named into existence and none has an article; and the four branches are linked in the caption but left as plain text in the bullets that define them. The 2026-07-27 edit to Tacchoan introduces four branch names — Esselis, Rumesselis, Kipoa, Usy — and one place, the island of Usings. All five are red. They account for the entire rise in wanted pages this run, 31 → 36: four at two incoming links each (the article and the file description page, which repeats the caption) and Usings at one. Separately, the linking is inconsistent within the same six lines: all four branches are linked from the image caption, and all four are plain text in the bullet list immediately below — the bullets being the passage that actually defines each one. | Two halves. The mechanical half is the linking: link the four branch names at their bullets, where they are defined. The flag-only half is the articles. Four branch stubs could be written from what Tacchoan already asserts — each branch's position on Taccho, its relation to the others, and Kipoa's colonial reach into Lovvocia — restating wiki-asserted facts only and inventing nothing; Usings cannot, since the article gives it no property beyond being an island where Usy is spoken. Whether those stubs are wanted is the operator's call. | deferred | New 2026-07-27. #11 (deferred) is not this: it covers the empty Cargo member query on this and two sibling articles, which is a language-level gap and stays empty either way, because these five are branches and places rather than languages. Which raises a structural question the routine is not filing as its own row: the branch layer has nowhere to live in the structured data — LanguageFamilies holds the family, Languages holds languages, and neither carries a field for the layer between, so once branch articles exist nothing will relate them to Tacchoan but prose. Adding a field is a schema change and needs the administrative rebuild noted in §Infoboxes and structured data. Talk:Tacchoan does not exist; if the stubs are declined, §Open threads there is where the four names belong. The five names are written unlinked throughout this report, as Template:He-decl and Dictionary:dog were last run: linking them here would add a third incoming link to each and make this page a cause of the very counts it reports.
Re-checked 2026-07-28, unchanged. Re-checked 2026-07-30, unchanged: the five names are still red, still account for five of the 36 wanted pages, and the four branches are still linked in the caption and plain text in the bullets that define them. The 2026-07-30 edit to Tacchoan replaced the empty Cargo member query and touched nothing else. One part of this note is now moot: the structural question it raised — that the branch layer had nowhere to live in the structured data — has been settled by the retirement of Cargo from the encyclopedia. Member lists are hand-maintained now, so branch articles would be tied to Tacchoan by prose and category whether or not they are written, and no schema change is in the way. Re-checked 2026-07-31, unchanged: the five names are still red, they still account for five of the 31 wanted pages, and the four branches are still linked in the caption and plain text in the bullets that define them. Tacchoan was not edited this window. Re-checked 2026-08-03, unchanged: the five names are still red — Esselis, Rumesselis, Kipoa and Usy at two incoming links each, Usings at one — and the four branches are still linked in the caption and plain text in the bullets that define them. One name has gained weight elsewhere: List of languages now gives Esselis as the region of the ancient Tacchoan language Esselitian, so the branch name is being used as a placename in a second article while still having no page of its own. Operator verdict 2026-08-04 (via Thiorosan): deferred. | |
| 58 | 2026-07-28 | structural | low | Talk:Family A, Talk:Family B, Talk:Family C, Talk:Family D, Talk:Family E | Five pages are now queued for deletion and only an administrator can clear them. Carrying out #46(a) this run placed {{delete}} on the five talk pages that outlived the language-family rename, so Category:Pages flagged for deletion has gone from empty to five members. Each carries the same stated reason: "Talk page of an article deleted in the language-family rename; its one open thread was settled when the family was named." The underlying facts were confirmed when #46 was filed — Family A through Family E were deleted by Yezur on 2026-07-19 and their talk pages were simply never moved with them. |
flag-only. The editing bots hold no delete right, so this is listed for the administrator to action, as the task list requires of the deletion queue. Delete the five and the row clears itself; the routine will keep listing them every run until then. Removing a tag is not something the routine will do — if any of the five should be kept after all, say so and the row becomes the instruction to untag. |
deferred | New 2026-07-28, and the direct product of an approved fix rather than a fresh defect. The queue was empty on every previous run.
Re-listed 2026-07-30, unchanged: the same five pages, each carrying the same stated reason, are still the whole of Category:Pages flagged for deletion. Re-listed 2026-07-31, unchanged: the same five pages, each carrying the same stated reason, are still the whole of Category:Pages flagged for deletion. Third run this row has been listed; it clears itself the moment the five are deleted. | Re-listed 2026-08-01, unchanged: the same five pages, each with the same stated reason, are still the whole of Category:Pages flagged for deletion. Fourth run this row has been listed. Re-listed 2026-08-03, unchanged: the same five pages, each with the same stated reason. Fifth run this row has been listed — but for the first time it is no longer the whole queue: 37 superseded Proto-Andusian entries were tagged this window, taking Category:Pages flagged for deletion from five members to 42. They are filed separately as #70 because they are a different cause; both clear the moment an administrator deletes them. Operator verdict 2026-08-04 (via Thiorosan): deferred — the operator will get to the deletions in due time. | |
| 60 | 2026-07-30 | structural | high | the Map: namespace (33 pages), Map:Features, Map:Main Page |
Thirty-three of the thirty-five pages in the Map: namespace can no longer be reached by name, and the index built last run to gather them is now almost entirely red. The namespace's title-casing rule has changed: Map: (namespace 3004) now reports case=first-letter, so a title is capitalised after the prefix. Thirty-three of its pages were created under lowercase titles — Map:place, Map:highway, Map:name, Map:natural=coastline and the rest — and that rule can no longer address them. The pages themselves are intact: addressed by page id, Map:place is still 2,062 bytes of wikitext, and all thirty-three still hold their content. Nothing can reach it. A reader who opens one is bounced: /wiki/Map:place answers 301 to /wiki/Map:Place, which answers 404 with the "no text in this page" notice. The API behaves the same way — action=raw returns an empty body, and a title query silently normalises Map:place to Map:Place and reports it missing. Map:Features now renders 34 of its 35 links red: every key and value page it was built to index, plus Map:Conventions, which was already wanted. Only two pages in the namespace still work, and only because their titles begin with a capital anyway — Map:Features and Map:Main Page, which means the map portal's own navigation now leads nowhere. When it happened: the last report recorded those same 33 links as resolving on 2026-07-28, and this page's five Map: links were re-recorded in capitalised form when it was saved on 2026-07-30, so the change falls between 2026-07-28 20:40 and 2026-07-30 14:09. What it was: the Dictionary: namespace still reports case=case-sensitive, which is what Map: must have held when these pages were created — a first-letter namespace will not accept a lowercase title at all, so they could not have been made under the present setting. |
(a) The recommended fix is one line of configuration, and only you can make it: restore the case override for namespace 3004 in LocalSettings.php — the same setting that still keeps Dictionary: case-sensitive — and, for consistency, 3005 (Map talk:, which holds no pages). That brings all thirty-three back at once, with every page name, every link and the whole Map:Features index correct as already written, and nothing on the wiki to edit afterwards. (b) The fallback, which the routine could carry out, is to move all thirty-three to capital-initial titles and rewrite Map:Features and the key pages' internal links to match; Alompan holds the move right and can address the pages by page id, which is the only handle that still works. It is the worse answer — it renames a documented tag set away from the lowercase form the tags themselves take — and it is offered only in case the configuration cannot be recovered. Do not do both. Until one of them happens, nothing new should be written into the namespace. |
fixed | New 2026-07-30, and the first high-severity row since #26'. Read out of the wiki rather than inferred: the case setting was queried from siteinfo, the surviving content was fetched by page id, and the 301→404 was followed end to end. Filed as one row because it is one cause. Not a defect in anyone's editing' — no page content is wrong, and Map:Features was correct on the day it was written. #33 (dismissed) named two of the same value pages for a different reason and is not re-opened here. Note on this report: its own five links into the namespace were the sole cause of five wanted pages, so they are written as plain code from this revision on.
RESOLVED 2026-07-31, by fix (a). The namespace has been put back: Re-verified 2026-08-03 and it still holds: | |
| 61 | 2026-07-30 | factual | medium | Yezur Wiki:Main Page, Module:Languages | Two pages still describe Cargo as the engine behind the encyclopedia's infoboxes and lists, after it was retired from exactly that role. Yezur Wiki:Conventions now reads "Infoboxes are display only — they file nothing into a database. Lists, member rosters, and cross-references are written and maintained by hand"; the five encyclopedia infoboxes have had their #cargo_store and #cargo_declare calls removed; the family Members tables and the two category listings have been converted to hand-maintained text; and Special:CargoTables is down to the single Lexemes table. Two pages were not brought along. (a) Yezur Wiki:Main Page §How it runs still introduces the software as "extended with Cargo (the structured data behind the infoboxes and the self-building lists)", and its §Conventions still advises "Prefer templates, Lua, and Cargo over repetition." Neither holds for the encyclopedia any more, and the first contradicts the convention page a reader is sent to from the same screen. (b) Module:Languages's header comment ends "Keep this table in step with the Languages Cargo table" — the Languages table no longer exists, and the dependency now runs the other way: Template:Infobox language's documentation was rewritten on 2026-07-30 to say "keep it in step with Module:Languages". As written, the module points at nothing and the template points at the module. |
Mechanical, and asserts nothing new. (a) Narrow the §How it runs parenthetical to what Cargo still does — the Dictionary's Lexemes table — and drop Cargo from the §Conventions bullet, leaving "Prefer templates and Lua over repetition." The see-also link to Special:CargoTables is not proposed for removal: the page still exists and still lists a table. (b) Replace the module's closing comment with the direction that now holds — keep the table in step with the code field of Template:Infobox language — matching the wording that template's own documentation already carries. Two edits, neither touching behaviour. |
approved | New 2026-07-30, from reading every page outside the Dictionary namespace against the ruling; these two are the only survivals of it. Cross-reference: #47 also proposes edits to Yezur Wiki:Main Page (§Behind the scenes and §Conventions), so if both are approved they belong in one edit. Recorded as not a finding: Talk:English language's open thread asks whether the isolate's Cargo row should store a queryable value — a talk-page thread the ruling has simply answered, and a stale thread on a talk page is not a defect. Template:Head and Yezur Wiki:Dictionary structure still describe Cargo at length and are correct to: the Dictionary is the deliberate exception.
Re-checked 2026-07-31, unchanged and confirmed line by line: Yezur Wiki:Main Page still reads "extended with Cargo (the structured data behind the infoboxes and the self-building lists)" and still advises "Prefer templates, Lua, and Cargo over repetition"; Module:Languages line 9 still reads "Keep this table in step with the Languages Cargo table". Neither page has been edited since the row was filed.
Part (b) has been carried out, outside this routine. Module:Languages was rewritten three times on 2026-08-03 (revisions 1935, 1939 and 2211) and its closing header line now reads "Keep this table in step with the | |
| 62 | 2026-07-30 | voice | low | 210 Dictionary: entries, Yezur Wiki:Dictionary structure |
The definition sweep of 2026-07-30 puts several senses on one numbered line, and quotes glosses with straight quote marks where the wiki asks for italics. Khurouan enriched the definition line of 210 entries — every Proto-Herta-Olaic lemma but one — turning a bare gloss into a described semantic range: "# blood" became "# blood; the life-fluid; by extension kinship and killing". The work is careful and the content is not in question; two points of form are. (a) Senses per line. Yezur Wiki:Dictionary structure states the rule twice — "Definitions are numbered wikitext list items (#), one sense per line", and "numbered senses" among the habits adopted as-is. All 210 lines carry two to four senses separated by semicolons, and 86 add an em-dashed comparative note. (b) Quoted glosses. 31 of the lines name other concepts inside straight single quotes — "overlaps 'see' and 'remember'", "not clearly separated from 'foot'" — 40 distinct terms in all. Yezur Wiki:Conventions §Text formatting gives that duty to italics: "Italics carry their usual duties: words as words, native-language forms, titles of works". (Two further lines contain an apostrophe only, in digit's and beast's, and are not affected.) |
(a) is a decision, not a defect to correct. A reconstructed lemma arguably has one sense whose range is being described, rather than several senses to number — and many of the lines say exactly that ("not clearly separated from 'foot'"). If that reading is right, the fix is to record the exception on Yezur Wiki:Dictionary structure — proto-language entries gloss a semantic range on a single line — rather than to split 210 entries; if it is not, the lines want splitting into numbered senses, which is a large authorial edit and belongs with Khurouan rather than with the routine. (b) is mechanical: italicise the 31 quoted spans in place of the straight quotes, changing no wording and no fact. The routine can do it in one pass on approval. | fixed | New 2026-07-30, from reviewing the largest content change since the last run against the convention set in full. Two things recorded here rather than filed as rows of their own. (i) Two of the new lines use human in character for the first time on the wiki — "a human being; someone; one of us as against a beast" and "possibly 'human' at large". Every earlier use of the word anywhere on the wiki is out-of-frame, describing the human editor. Whether the peoples of Yezur are called human in the encyclopedia's own voice is a canon question, not a maintenance one; noted so it is not found again as a defect. (ii) One Proto-Herta-Olaic lemma was missed by the sweep and still carries its bare gloss: xiəltə "four", conspicuous because three and five were both enriched. Otherwise the sweep is clean: no Earth reference, no break of frame, every gloss= parameter left in step with its short sense, and Lexemes unchanged at 526 rows with no field lost.
Re-checked 2026-07-31. Part (a) has been half carried out, outside this routine. On 2026-07-30, between 20:31 and 20:34, Khurouan split packed senses onto separate numbered lines across 87 Proto-Herta-Olaic entries (revisions 1843–1929, summary "Enumerate senses as separate definitions"). The sweep is faithful — every one of the 87 was compared word for word against its previous revision, and nothing was lost beyond the connectives a split makes redundant. It is also incomplete: 124 of the 211 Proto-Herta-Olaic entries were not reached, and 118 of those still carry a packed first line, so the condition this row describes is now true of roughly three-fifths of the language rather than all of it. The two other languages were never affected — all 212 Ancient Hertic and all 102 Proto-Njeshan entries have exactly one definition line each. Part (b) has moved a little: straight-quoted glosses are down from 31 to 27 definition lines, four of them having become parenthetical labels in the sweep; the 27 are listed by the same method as before and 25 of them sit on entries the sweep did not touch. Two new rows come out of the sweep rather than being folded in here, because they are about how it was written rather than whether it should happen: #63 (the labels it introduced) and #64 (the separator it used). | Re-measured 2026-08-01 from the wikitext of all 527 Re-measured 2026-08-03 from a fresh 1,027-page corpus. The Proto-Herta-Olaic figures are unchanged — of the 211 entries, 87 carry more than one numbered definition line and 124 carry one, 118 of those still joining senses with | |
| 63 | 2026-07-31 | structural | medium | 87 Dictionary: entries (Proto-Herta-Olaic) |
Seventy-five sense labels were written as plain text, on a wiki that has a built template for exactly them. The definition sweep of 2026-07-30 introduced a leading parenthetical on 75 definition lines across 87 entries — (by extension) 71 times, plus (relative) twice and (indefinite) and (adjectivally) once each — all typed directly into the wikitext, as in # (by extension) the front, the chief, the beginning. Template:Lb exists and is built for this: its own documentation opens "Inline sense labels, placed at the very start of a definition line — the italic, parenthesised context/register/domain tags". Yezur Wiki:Dictionary structure names it twice, in §What the Yezur Dictionary adopts as {{lb|code|label…}}, "inline sense labels, italic and parenthesised before a sense (built)", and again in the passage describing definitions: "Context labels — {{lb|en|nautical|slang}} — prefix a sense". Module:Labels' own header names this very label among the ones that work without further setup: "semantic qualifiers (figurative, idiomatic, by extension, …) just work as plain text". The plain-text form renders upright, the template renders italic, so the wiki now shows two appearances for one thing. The template is barely used at all: 5 pages in a 527-page Dictionary call it, all of them Ancient Hertic, and all 75 plain-text labels date from this one sweep. |
Mechanical, one pass, no judgement. Rewrite each leading parenthetical as a label call: # (by extension) X → # {{lb|YAP-XX|by extension}} X, and likewise for relative, indefinite and adjectivally. The routine checked what this produces before proposing it: {{lb|YAP-XX|by extension}} renders (by extension) in italics and files the page into no category at all, because none of the four is in Module:Labels' categorising vocabulary — so the edit changes appearance and nothing else, and adds no wanted categories. Nothing else on the affected lines changes. |
approved | Re-measured 2026-08-01, unchanged: 75 plain parenthetical sense labels on the 87 pages — 71 (by extension), two (relative), one (indefinite), one (adjectivally) — and five pages on the whole wiki use {{lb}}, none of them a Proto-Herta-Olaic entry.
Overtaken by practice on 2026-08-03, exactly as this row's note predicted it would be. The note warned that if the sense-splitting sweep went further before this was settled, "the split runs the other way and the newer entries are the plain-text ones". What happened is the mirror image and is worse for the older corpus: the 270 new Proto-Andusian entries use | |
| 64 | 2026-07-31 | voice | low | 13 Dictionary: entries (Proto-Herta-Olaic) |
The same sweep separated near-synonyms two different ways, in the same pass, in the same language. Of the 87 entries it rewrote, 74 now join the terms of the first sense with commas — # bird, a flying creature, # fruit, the edible growth of a plant, # fear, be afraid, dread — while 13 kept the semicolon they had before: # nose; the smelling part, # head; the top of a person, # hand; the grasping part, # heart; the beating organ, # mouth; the opening, # tongue; the organ of taste and speech, # tooth; a biting part, # knee; the joint of the leg, # flesh; the edible substance of a body, # root; the underground part, # bite; seize with the teeth, # suck; draw in with the mouth, and # seed, grain; the sown thing, which uses both. The thirteen are not a distinguishable class — they are body-part nouns and their verbs, which the comma form covers elsewhere in the same sweep (# bone, the hard frame, # ear, hearing, # wing, the flying limb) — so the split reads as an artefact of the pass rather than a distinction being drawn. |
Pick one and apply it to the thirteen; either direction is a single mechanical pass and the routine holds no opinion on which. Comma matches the 74 and makes the whole sweep uniform. Semicolon is defensible if the intent was gloss; explanation — a short equivalent, then a fuller definition — in which case the 74 are the ones out of step and the fix is larger. A third answer is that the difference is deliberate and this row should be rejected. Whatever is chosen belongs in the same pass as #63 and, if it is carried further, #62(a). | fixed | New 2026-07-31. Filed narrowly. The larger fact — that the sweep covered 87 of 211 entries — is recorded against #62 rather than here, because it is that row's condition partly resolved, not a new one. This row is only about the punctuation the 87 were left with, which is new this window and did not exist to be flagged before. Low, and genuinely low: nothing is wrong or unclear on any of the thirteen pages read on its own.
Re-measured 2026-08-03 name by name and unchanged: of the 87 Proto-Herta-Olaic entries the sweep rewrote, 13 still open with a semicolon and 72 with a comma (two open with neither). The thirteen are the same thirteen. Now decidable against a larger practice: the 270 Proto-Andusian entries written this window use the comma in 116 of the 117 first senses that separate anything at all, which is the direction comma in this row's fix. Operator verdict 2026-08-04 (via Thiorosan): amended — semicolon wins: the intended form is gloss; explanation, so the thirteen are the model and the comma-opening first senses of the 2026-07-30 sweep are the ones out of step; bring them in line. Scope is this row's corpus — whether the ruling extends to Proto-Andusian awaits a separate operator word; do not extend the pass there yet. Operator second word, 2026-08-04 (via Thiorosan, superseding the verdict note above): deferred — the question is settled, differently from the note above, and recorded at Yezur Wiki:Dictionary structure: within a definition line the comma joins glosses and the semicolon introduces explanation (the model is "# seed, grain; the sown thing" — so neither mark wins wholesale, each has a role). Application across the Dictionary is assigned to Thiorosan directly, in the same audit pass as #62, once the Descendants edits currently underway are finished. No routine edits under this row. Carried out 2026-08-04 by Thiorosan, on the operator's direct instruction (chat): fixed. The convention recorded at Yezur Wiki:Dictionary structure — commas join glosses, a semicolon introduces explanation, "# seed, grain; the sown thing" the model — was applied across the whole Dictionary in one pass with #62: 163 definition lines on 158 pages (Proto-Herta-Olaic and Proto-Andusian; Ancient Hertic and Proto-Njeshan were checked and already conformed, as did English's definitional style). The thirteen semicolon entries stand unchanged as the model. Nothing left to apply — confirm and clear. | |
| 65 | 2026-08-01 | structural | low | Map:Main Page, Map:Features | The map portal's front page has two navigation tiles and one of them has never led anywhere. Map:Main Page presents the namespace through exactly two cards — Map features and Conventions — and the second points at Map:Conventions, a page that has never existed. It has been red since the card was written on 2026-07-15, and Map:Features added a second link to it on 2026-07-28 ("how features are tagged, named, and classified", in its §See also). Both linking pages describe the same absent subject in almost the same words, and the front page's own lead sentence promises it: "This namespace documents … the conventions for tagging and naming features." Half of what the portal advertises is not there. Map:Conventions is one of the 31 wanted pages, at two incoming links. |
Two ways, and the choice is yours because only one is mechanical. (a) Write the page. The material largely exists already, scattered through the 33 key and value pages — which values are preferred, how names are given, when a feature is classified rather than named — but collecting it is an editorial act, not a mechanical one, and the routine will not guess at map policy. (b) Remove the two links until the page is written: drop the Conventions card from Map:Main Page (leaving Map features as the single tile) and the bullet from Map:Features. Mechanical, reversible, and it takes Map:Conventions off the wanted list. (a) is the better end state; (b) is what to do if it will be a while. |
deferred | New 2026-08-01. Not a new breakage — it has been true for seventeen days and no run filed it, because every run so far attributed the red link to #26, the row that created Map:Features. #26 was applied and removed on 2026-07-28, so since then the attribution has pointed at nothing and this red link has had no open row. Filing it now gives it one. Not part of #60 — that row was about the namespace being unreachable by name, which is fixed; this page is simply absent, and was absent before the breakage and after the repair. Read against Map:Features: 34 of its 35 links are blue and this is the 35th, so it is now the only red link in the Map: namespace apart from the two lowercase place values of dismissed #33. Low: nothing is wrong on either page, and one dead tile on a two-tile portal is a gap rather than an error.
Re-checked 2026-08-03, unchanged: | |
| 66 | 2026-08-02 | structural | low | Main Page, Yezur Wiki:Main Page, Dictionary:Main Page, Map:Main Page, Template:Delete | Every front page the wiki has sets a background colour and no text colour, so each one reads correctly only by borrowing the skin's. Fifteen inline style attributes declare a background without declaring a color beside it: five on Main Page (the From the world panel and the four Explore the world cards), three each on Yezur Wiki:Main Page, Dictionary:Main Page and Map:Main Page, and one on Template:Delete's notice. Each box therefore fixes its own panel colour — #fbfcfd, #f8f9fa, #fdeaea — and takes its text colour from whatever the skin is serving. Today that is one palette only: night mode is switched off in this wiki's Vector configuration, so every box renders dark text on a light panel and nothing is wrong on any page as it stands. The moment night mode is enabled the skin turns its text light while these fifteen backgrounds stay light, and the boxes go light-on-light — all four portals and the deletion notice at once, from one configuration switch. This is exactly the defect MediaWiki's night-mode-unaware-background-color linter category exists to name, and the wiki's own linter reports 0 of them: the category records nothing while night mode is off. The routine found these by reading the wikitext of all 714 pages, not from the linter. |
Mechanical, and it asserts nothing about the palette. (a) Add an explicit color: to each of the fifteen style attributes, set to the colour the box already renders at — #202122, the skin's own body text colour — so each panel keeps exactly the appearance it has today and stops depending on the skin to supply half the pair. Fifteen insertions across five pages, no visual change now, and the night-mode switch becomes safe to flip. (b) is the larger option and yours to weigh: the four portals repeat nearly the same card styling fifteen times inline, and the wiki already serves TemplateStyles (Template:Infobox/styles.css). Moving the portal card rules into a stylesheet settles the colour pair once instead of fifteen times, and any later change to the portals' look becomes one edit. That is a rework of how the front pages are built rather than a maintenance fix, so it is filed as an option, not a recommendation. |
approved | New 2026-08-02. Found by hand, because the check that should have found it is dormant — see Latest run. Not urgent and not a present defect: every page renders correctly today and will keep doing so until night mode is enabled, which is why this is low. It is filed anyway because the trigger is a single setting rather than an edit, and because it lands on all four front pages simultaneously — the pages a reader meets first. Distinct from #14 (the same three portal pages, uncategorised) and #54 (file conventions), which touch these pages for unrelated reasons; neither covers styling.
Re-scanned 2026-08-03, and the count in this row is corrected upward: the scan that produced it looked at inline | |
| 67 | 2026-08-03 | factual | medium | Herta-Olaic, Njeshan, Tacchoan, List of languages | Twelve languages joined List of languages and not one reached the family article that is supposed to list it — and on one of those articles the roster is contradicted outright. The list grew on 2026-08-03 from 10 entries to 22, restructured into Modern and Historical tables. The three families that gained members carry a hand-maintained §Members table apiece, and none was brought along. Herta-Olaic lists 2 (Ancient Hertic, Gaillean); the roster assigns it 6, so Airananwe (modern), High Gaillean and Middle Airananwe (medieval) and Skennic (ancient) are missing. Njeshan lists 7 against 8: Anilonic (modern) is missing. Tacchoan lists none at all — its §Members ends "No member languages are currently listed." — while the roster now assigns it 7: Esselitian, Kankai, Old Ousperian, Old Usinger, Senitian, Terjiquan and Usinger. That sentence is not merely stale but false, and it is the one the routine itself asked for on 2026-07-30. The rest of the chain is sound, checked rather than assumed: all 23 codes in the list resolve in Module:Languages and agree with it on name and family; the three reconstructed codes are absent from the list by its own lead; the ten articles carrying {{Infobox language}} all agree with the registry on family; and Andusian and Shung-Worrehan list no members because the roster gives them none. |
Mechanical, and asserts nothing new. Every §Members table carries the same three columns the roster does — Language, Native name, Era — so each missing row is a copy: Airananwe / Airánanue / Modern, Skennic / Sgeeni / Ancient, High Gaillean / Gwalea / Medieval, Middle Airananwe / Airannawê / Medieval on Herta-Olaic; Anilonic / Aniljew / Modern on Njeshan; and on Tacchoan, replace the "No member languages" sentence with the seven-row table — Senitian / Sennic / Modern, Usinger / Yszènnu / Modern, Esselitian / Ezjec / Ancient, Old Ousperian / Ösuppə / Ancient, Kankai / — / Medieval, Old Usinger / Yszènqè / Medieval, Terjiquan / Tertzikkè / Medieval. Names go in as plain text where no article exists, exactly as the roster writes them, so no red link is created. | approved | New 2026-08-03, from the hand-maintained-lists check, which has agreed with itself on every previous run and has now found the thing it exists to find. Note on Tacchoan: #53(c), #55, #56 and the mechanical half of #57 all touch the same §Members section; if any of them are approved with this, they belong in one edit. Recorded as not a finding: the roster's Region and Notes columns are empty for most of the twelve, which is a gap in an article rather than a drift between two.
Operator verdict 2026-08-04 (via Thiorosan): approved. | |
| 68 | 2026-08-03 | factual | medium | Yezur Wiki:Dictionary structure | The Dictionary's blueprint has been overtaken by the Dictionary, in three places. (a) Its lead still reads "Nothing below is built yet — it documents the target so that words can be added in a structured, precise way from the very first entry." The Dictionary now holds 836 pages in six languages and files 800 rows into Lexemes; the same page marks six of its own templates built. (b) Its adoption list ends "No translation tables yet. With six languages and no lexicon, glosses carry the load; revisit the hub-language mechanism once entries exist in several languages for the same concepts." The hub-language mechanism was built on 2026-08-03 — Template:Translations, Module:Translations and a stylesheet — and is live on Dictionary:warm, the wiki's first English lemma, where one box carries five languages across three families. The condition the bullet set for revisiting has been met and the revisiting has happened. (c) The heading skeleton the page adopts gives a heading per relation (====Synonyms====, ====Antonyms====, ====Derived terms====, ====Related terms====). The entries written this window instead gather every relation under one ====Related terms==== heading as labelled bullets — Antonym: 30, Coordinate term: 14, Compare: 7, Near-synonym: 4, Derived term: 1, across 54 entries — on the operator's instruction, recorded in the edit summaries. The practice is settled; the page that is supposed to record it does not. |
Mechanical throughout, and asserts nothing new — every replacement restates something the wiki already does. (a) Replace the "nothing below is built yet" clause with what is now true: the skeleton is in use, and the page documents the form entries take. (b) Replace the no translation tables bullet with the built form, in the words Template:Translations's own documentation already uses — one box per sense under an English sense, keyed by registry code, grouped by family through the registry, hand-curated and display-only. (c) Record the labelled-bullet form in the adoption list beside the headings it replaces. | approved | New 2026-08-03. Not a criticism of any of the work — (a) and (b) are the blueprint being outrun, which is what a blueprint is for, and (c) is an operator ruling that simply has not been written down. Recorded as not a finding: the registry bullet still names eight codes where Module:Languages now holds 26, but it reads "the same codes the language infoboxes already carry" and gives examples rather than a list, so it is not wrong. #62 is the neighbouring row and stays separate: it is about what the entries say, this is about what the blueprint says.
Operator verdict 2026-08-04 (via Thiorosan): approved. | |
| 69 | 2026-08-03 | structural | low | Template:Translations/testcases | Nine of the forty-three wanted pages were created by one worked example. The new testcases page demonstrates the translations box with placeholder words — word1 … word9 — under known registry codes, and Module:Translations links every word under a known code to its Dictionary entry, so the page renders nine red links: Dictionary:word1 through Dictionary:word9. (word10 sits under the deliberately unrecognized code YQQ-ZZ, renders as an error and is not linked, so it makes no wanted page.) The page is candid about it — "All words below are placeholders unless they link blue" — so this is a known consequence rather than an oversight. It is filed because the wanted-pages list is a maintenance instrument the routine reads every run, and one page now supplies 21% of it. |
Mechanical either way, and the routine holds no opinion beyond preferring the choice be made once. (a) Swap the placeholders for words that exist: the Dictionary holds 800 lexemes, Dictionary:warm's own box already names five real ones across three families, and a testcase that shows the links resolving demonstrates more than one that shows them red. (b) Leave it and record it, as the worked example on Template:D (Dictionary:dog) has been recorded since 2026-07-25 — in which case nine entries on the wanted list are permanent and known. | amended | New 2026-08-03. Not proposed: removing the testcases page, which is the only documentation of the box's error rendering and homonym handling. Recorded as not a finding: the same page is unused and uncategorised, as Template:Auphen/testcases has been since it was written — a testcases page is not meant to be transcluded.
Operator verdict 2026-08-04 (via Thiorosan): amended — swap in real lemmas where possible; where that cannot be done, make the testcases page more sober instead. | |
| 70 | 2026-08-03 | structural | low | 37 Dictionary: pages (Proto-Andusian) |
The deletion queue has gone from five pages to forty-two. An operator ruling on 2026-08-03 retired the implosives from Proto-Andusian (ɓ → bʰ, ɗ → dʰ). The 37 entries whose titles carried them were re-created at the new spellings and the old pages blanked to a {{delete}} tag apiece, each with the same stated reason: "Superseded by [[Dictionary:<the new spelling>]] — implosives retired from Proto-Andusian (ɓ→bʰ, ɗ→dʰ)". The pages, with their replacements: aɓfiq, fuɗ, gbeɗ, kuɗum, lɛɓ, mudiɗiᵑg, muɓe, mɛɗ, qoɗ, ruɓ, saɗ, ɓafgi, ɓara, ɓat, ɓek, ɓet, ɓeɓe, ɓolo, ɓuɣɛsox, ɓɔŋ, ɗes, ɗex, ɗiki, ɗirpeᵐb, ɗiv, ɗixwa, ɗo, ɗudo, ɗukif, ɗum, ɗɔkfo, ɗɔt, ɗɔɣ, ɗɛ, ɗɛw, ɲevaɗ, ᵐbɛɓ. The retirement itself is complete, checked rather than assumed: every one of the 37 replacement pages exists; no page anywhere on the wiki still contains ɓ or ɗ outside the flagged titles themselves, so nothing links or refers to the old spellings; and none of the 37 still files a Lexemes row, the table holding 273 Proto-Andusian rows across 270 pages (the three extras being noun/verb pairs on one page each). Only the deletion is outstanding. |
flag-only. The editing bots hold no delete right, so this is listed for the administrator, as the task list requires of the deletion queue. Delete the 37 and the row clears itself; the routine will keep listing them every run until then. Removing a tag is not something the routine will do — if any should be kept, say so and the row becomes the instruction to untag. |
deferred | New 2026-08-03, and like #58 the product of deliberate work rather than a defect. Filed as its own row rather than folded into #58 because it is a different cause with a different reason, and because it clears in one action: the five talk pages and these 37 are unrelated except in needing the same right. Recorded as not a finding: the 37 new pages sit on Special:LonelyPages and Special:DeadEndPages, as every Dictionary entry does.
Operator verdict 2026-08-04 (via Thiorosan): deferred — as #58, the operator will get to the deletions in due time. | |
| 71 | 2026-08-03 | structural | low | List of languages | One of the twelve new roster entries links to an article that does not exist; the other eleven are plain text. Of the twelve languages added on 2026-08-03, eleven are written as plain text — Anilonic, Senitian, Usinger, Esselitian, High Gaillean, Kankai, Middle Airananwe, Old Ousperian, Old Usinger, Skennic, Terjiquan — and none of the eleven has an article. The twelfth is written [[Airananwe language|Airananwe]], and that article does not exist either: it is the only red link the article namespace gained this window and the only new mainspace entry on the wanted-pages list. The same table is inconsistent in a second place: the Family column links Herta-Olaic, Njeshan and Tacchoan on most rows but leaves the same three names plain on three (Senitian and Usinger → Tacchoan, Skennic → Herta-Olaic). |
Both halves mechanical, and (a) is a choice of direction rather than a defect to correct. (a) Either unlink Airananwe so the twelve read alike, or leave it as the one deliberate invitation to write that article — but not one of twelve by accident. (b) Link the three plain family names to the family articles they name, matching every other row in both tables. | amended | New 2026-08-03. #32 (rejected) is not re-opened: it covered the absence of the Airananwe language, on which the operator said "that language will land later. No need right now", and this row takes no position on that. What is new is that one entry is linked and eleven identical ones are not, in a table written in a single sitting. Recorded as not a finding: the two tables mix inline ( |
) and one-cell-per-line row syntax, which renders identically, and English's Family cell reads Language isolate — a category rather than a family, and linked from the lead of the same page.
Operator verdict 2026-08-04 (via Thiorosan): amended — red-link the rest: link all twelve roster entries, the eleven plain names becoming red links like Airananwe; and (b) as suggested, link the three plain family names. |
Dismissed flags
Deferred or rejected findings, kept here so later runs do not re-flag them.
- #8 (deferred) — Wanted history-period pages Yezuri Age of Exploration / Yezuri Middle Ages. Operator: not canon names for these eras; will not create stubs under temporary names.
- #11 (deferred) — Andusian / Tacchoan / Shung-Worrehan member-language queries fall back to the default. A content gap, not a data error; operator: no member languages are in yet, add when ready.
- #12 (rejected) — Greek-letter native names on Hertia / Hertic language / Alompe etc. Operator: fully intentional in-world orthographic style.
- #14 (deferred) — Uncategorised portal front pages and the root Category:Yezur. Operator: accepted, but no decided approach for handling it yet.
- #21 (deferred) — The four wanted Ancient Hertic inflection templates (
conj-IS,conj-OO,decl-N1,decl-N2). Operator: "will be fixed when the tables are created — already on my to-do." Confirmed present on the to-do list at User talk:Yezur. - #30 (rejected) — The one-line User:Yezur page, read as an unfinished sentence. Operator: the line is not unfinished — the city name is written there as
, "using a script I have added to the wiki, with characters stored in the PUA … a temporary text intended to be used for testing the PUA font; it is working correctly." Answering the operator's condition for reflagging: the routine can see the characters and reads five PUA codepoints there (U+E067 U+E111 U+E179 U+E179 U+E0EB, the third and fourth identical, so a five-letter name with a doubled letter), which round-trip intact through the API. What it cannot do is read them: no mapping from those codepoints to sounds or letters exists anywhere on the wiki, so the glyphs are legible only to a reader with the font. Not reflagged. If a transliteration table for the script ever lands on-wiki, say so and later runs can check such names instead of skipping them. - #32 (rejected) — Airananwe and the code
name:yas-aiinMap:name's worked example matching no language and no registry code. Operator: "that language will land later. No need right now." Left as written; the routine will not re-open it when the language arrives. - #33 (deferred) — The two redlinked values
Map:place=hamletandMap:place=suburbinMap:place's settlement table. Real but not now; they remain the only redlinked values in any key page's value table, and both are listed in the new Map:Features index only through their parent key. - #43 (rejected) — Template:Infobox former country declared and stored into a FormerCountries Cargo table that was never created, so Hertia's structured data went nowhere. Operator: "Cargo tables no longer the go-to." Overtaken by the retirement of Cargo from the encyclopedia (2026-07-30): the declaration and the store have both been removed from the template, so there is no longer a table to create. Confirmed this run — of the six tables the wiki once declared, only Lexemes answers a query; Continents, Languages, LanguageFamilies, Polities and FormerCountries now all throw the same
MWExceptionthat once distinguished an absent table from an empty one, because all five are now absent. The administrative Create data click this row was waiting on is no longer wanted. - #59 (rejected) — Betonia, Unteria and Sea of Sihia categorised but reaching no Cargo table, with no infobox on the wiki fitting a subcontinent, an island or a sea. Operator: "No longer relevant since Cargo tables no longer the go-to. Not everything needs to be tabulized." The structured-data half of the finding has been removed by the same ruling, and the infobox question closes with it. Vetsihia's borrowed
{{Infobox continent}}is left as it stands: with Cargo gone it is display only, and itstype=Geographic region (colloquial)already says what the article is.
Latest run
2026-08-03. After four runs with nothing to look at, this one had 851 content changes on 322 pages — every one of them made on 2026-08-03, between 07:42 and 19:54, by Thiorosan (835), Khurouan (15) and Yezur (1). 313 pages were created. The wiki went from 714 pages to 1,027, from 452 articles to 581, and from 1,910 edits to 2,762. It is by a wide margin the largest window the routine has ever surveyed, and the first in which the hand-maintained-lists check found something — which is the run's main finding, and the reason a routine keeps running a check that has agreed with itself twenty times.
What the window contained
Four strands, all operator-directed:
- A fifth Dictionary language. Proto-Andusian (
YCP-XX) was registered and seeded — 257 core-vocabulary entries and 13 derivational affixes — then revised four times over: a strict in-universe pass on 218 entries dropping reconstruction hedges and metalinguistic commentary from definitions; a follow-up fixing duplicated senses on 105; a pass converting See also to labelled Related terms on 135; and an operator ruling retiring the implosives (ɓ → bʰ, ɗ → dʰ), which re-created 37 entries at new titles and queued the old ones for deletion (#70). It ends the day at 270 live entries, the second-largest language on the wiki. Dictionary:About Proto-Andusian documents its conventions, phonology, derivation and reduplication. - The language roster tripled. List of languages was split into Modern and Historical tables and grew from 10 entries to 22; Module:Languages went from 10 registry codes to 26; and three articles gained the codes newly assigned to them. This is where the run's largest finding is (#67).
- The hub language arrived. Template:Translations, Module:Translations, a stylesheet and a testcases page were built, and Dictionary:warm became the wiki's first English lemma, carrying one translations box across five languages and three families. Two rows come out of it (#68(b), #69).
- Fifteen sense extensions by Khurouan at the end of the day, three of which added a second part-of-speech section to an existing entry.
Nothing has been triaged for five runs
The last verdict left on this page is still rev 1629, on 2026-07-30. Seventeen rows were pending going in and five are added below. The backlog has not been restated at length again; the short form is unchanged. #60 needs only a word — the breakage it describes has been repaired and re-verified four times. #52, #55, #56 and #66(a) are one edit each with the wording already written out. #61 has had half of itself carried out by someone else and needs a word on the other half. #58 and #70 need an administrator, not the routine. What has changed is that the Dictionary rows are no longer merely urgent: see below.
Fixes applied
None. No row carried a verdict.
One row half-resolved, outside the routine
#61(b) has been carried out by Thiorosan. Module:Languages' closing header line, which pointed at a Cargo table that no longer existed, now reads "Keep this table in step with the code fields of {{Infobox language}} and the List of languages tables" — the direction the row asked for, and one better, since it names both hand-maintained sources rather than one. Part (a) — the two Cargo sentences on Yezur Wiki:Main Page — is untouched and the row stays open on it alone.
The hand-maintained-lists check, on the day it mattered
The task list asks this run to check the family Members tables against the languages that declare each family. That check has been re-derived from scratch on every run and has never disagreed with itself. This run it did, in three articles at once, because the roster grew and nothing downstream was brought along:
- Herta-Olaic lists 2 members; List of languages assigns it 6.
- Njeshan lists 7; the roster assigns 8.
- Tacchoan lists none, and says so in words — "No member languages are currently listed." — while the roster assigns it 7.
Twelve entries in all, and on Tacchoan a sentence that is now false rather than merely out of date. Filed as #67, with every missing row's three cells written out; the fix asserts nothing the roster does not already say. The rest of the chain checks out, and was verified rather than assumed: all 23 codes in the roster resolve in Module:Languages and agree with it on both name and family; the three reconstructed codes are absent from the roster by its own lead; the ten articles carrying {{Infobox language}} agree with the registry on family without exception; and Andusian and Shung-Worrehan list no members because the roster gives them none.
The standing caveat, restated a fourth time. The task list also asks for "Main Page's continent list against the continent articles". Main Page has no continent list. It links Category:Continents from a navigation card; that category holds the expected 7 members, with Vetsihia correctly outside it in Category:Geographic regions. The instruction should name the category or be dropped.
The Dictionary rows have been decided by practice, not by verdict
#62, #63 and #64 were filed about how the Proto-Herta-Olaic entries are written. All three were re-measured from a fresh 1,027-page corpus this run and every Proto-Herta-Olaic figure is unchanged. What changed is that 270 entries in a new language were written this window — and each was written the opposite way to the corpus the rows describe:
- Senses per line (#62(a)): 259 of the 270 new entries carry more than one numbered definition line. 124 of 211 Proto-Herta-Olaic entries still carry one packed line.
- Sense labels (#63): the new entries call
{{lb}}on 288 definition lines across 197 pages. The 75 plain-text parentheticals on 70 Proto-Herta-Olaic pages are untouched, and not one Proto-Herta-Olaic entry uses the template. Wiki-wide, 201 Dictionary pages now call it. - First-sense separator (#64): the new entries use a comma in 116 of the 117 first senses that separate anything. The thirteen Proto-Herta-Olaic semicolons are the same thirteen.
#63's own note predicted this in reverse — it warned that carrying the sweep further before settling the row would leave "the newer entries the plain-text ones". The mirror image happened instead, which is worse for the older corpus rather than better: the newer entries are the template ones, and they now outnumber the plain-text ones four to one. Three of the four rows are one mechanical pass over one language, and the pass gets no cheaper.
On #62(a) specifically, the operator's own ruling has answered the open question. The row asked whether a proto entry has one broad sense or several. Dictionary:About Proto-Andusian, written this window, states the policy — "Definitions are deliberately broad. A reconstructed word is a field of related senses rather than a point" — and the entries under it enumerate the senses anyway. That is the answer, applied to one language; #62(a) is now the question of whether the other proto-language follows it.
A count in an open row, corrected
#66 recorded fifteen background colours with no text colour, on five pages. The figure was inline-only: the scan behind it read style attributes in wikitext and never looked at the wiki's two TemplateStyles sheets. Those carry four more rules of the same shape — three in Template:Infobox/styles.css (which is every infobox on the wiki, not one page) and one in Template:Translations/styles.css, created this window. The true extent is 19 declarations across 7 pages. The row has been corrected in place rather than re-filed. It also changes what fix (b) buys: moving the portal cards into TemplateStyles settles nothing on its own, because the sheets already there have the same gap. Nothing renders wrongly today; night mode is still off and the linter still reports 0.
The content review
851 edits were read against Yezur Wiki:Conventions in full. The register and the frame hold up. Across every page changed this window in an in-frame namespace, one word triggered the Earth test — human, in Dictionary:kidz "person, human being" — and that is the canon question #62's note recorded on 2026-07-30 rather than a new defect. 0 reconstruction hedges survive in any Proto-Andusian definition line and 0 entries carry a duplicated sense, so both operator-directed clean-up passes reached everything they were aimed at, checked entry by entry. The hedges that remain — possibly, reconstructed beside, attested only bound — are all in Etymology and Usage notes sections, which is where the ruling left them, and read as in-world comparative scholarship rather than as authorial doubt.
Three findings came out of the review: #68 (Yezur Wiki:Dictionary structure overtaken in three places, one of them an operator ruling that was never written down), #69 (nine wanted pages from one worked example) and #71 (one of twelve new roster entries linked, eleven not).
New flags
Five. #67 (medium, factual) — twelve languages missing from three family Members tables, and a sentence on Tacchoan that is now false. #68 (medium, factual) — the Dictionary blueprint out of step in three places. #69 (low) — nine wanted pages created by Template:Translations/testcases. #70 (low) — 37 superseded Proto-Andusian entries queued for deletion, for the administrator. #71 (low) — one linked entry among twelve identical ones in List of languages.
Clean
Linter errors 0, and corroborated a second time by the method last run established: the wikitext of all 1,027 pages was fetched and scanned outside the extension — 0 obsolete tags, 0 self-closed tags, 0 links inside external links, 0 multi-colon escapes and 0 unbalanced div/span/table/ref, on a corpus that has grown by 44% since the last such scan. The one category that cannot be corroborated is the night-mode one, now #66 with a corrected count. Broken redirects 0, double redirects 0. Category:Auphen errors 0 members across 232 transclusions of {{auphen}} (215 last run), Category:Auphen test failures 0. Unused images 0.
Lexemes — 800 rows, up from 526. No row missing word, lang, pos or gloss; no duplicate (word, lang, pos); word equals the page title on all 800. 210 rows carry an etymology (209 inherited, one compound), unchanged, every etym_source one of YAP-XX/YAQ-HE/YBP-XX, and no row holding one of etym_type/etym_source without the other. Each code maps to exactly one langname: YCP-XX 273, YAQ-HE 212, YAP-XX 211, YBP-XX 102, YBS-PJ 1, YXS-EN 1. One check has changed meaning and is reported rather than flagged: three pages now carry two rows — Dictionary:aɢ, Dictionary:ceŋ and Dictionary:gida — where every previous run found none. All three are noun/verb pairs added by Khurouan on 2026-08-03 under two part-of-speech headings on one page, which is exactly what the blueprint's skeleton prescribes. Two rows on a page is now correct; the check now tests for duplicate (word, lang, pos) only, and that is still 0.
Wanted pages 43 (31 last run), wanted categories 59 (43), wanted templates 4 (unchanged) — all diffed name by name rather than compared by total. No entry dropped off any of the three. The twelve new wanted pages are Airananwe language (#71), the nine testcase placeholders (#69), Module:Translations/doc (the eighth module joining to-do flag #25) and Category:Proto-Andusian lemmas. The sixteen new wanted categories are the twelve Proto-Andusian lexical categories and four English ones, all emitted by {{head}} and all covered by to-do flag #18, whose accepted model is recorded at User talk:Yezur and is not re-flagged. The four wanted templates are #21's inflection tables.
Uncategorised pages 4, categories 1 (Category:Yezur), templates 3, images 1 — the three new entries all created this window, noted under #54 and #69 rather than filed. Unused templates 3. Orphaned pages 298 and dead-end pages 289, both up by roughly the number of Dictionary entries created, which is what those two lists do on this wiki: a lexicon entry is linked from few pages and links to few.
Deletion queue — 42 pages, up from five. The five Talk:Family A–Talk:Family E pages of #58, listed for the fifth run, and the 37 superseded Proto-Andusian entries of #70, listed for the first. Only an administrator can clear either.
Site totals: pages 1,027, articles 581, edits 2,762, files 1, job queue empty. Nine registered users.
On the routine's own tooling. Every fetch this run — recent changes, all thirteen maintenance lists, the Lexemes table and the wikitext of all 1,027 pages — was paginated through the Python-built ASCII query strings last run's failure forced, with continuation-token loop detection on each. The page fetch returned 1,027 of 1,027 titles, matching Special:Statistics exactly, and the corpus's highest revision id is 2783, the last edit of the window, so nothing was measured against a stale copy.
Self-check on this report. Its links were parsed before posting with the corrected check — the one that reads the value of exists rather than testing for the key. Twenty-two rows, nine cells each, table structure verified before posting. The nine testcase placeholders of #69 and the 37 superseded titles of #70 are written as plain text, not links, so this page does not add an incoming link to anything it reports as wanted or as queued for deletion.
Pending
Twenty-two rows, none high: #47 #50 #51 #52 #53 #54 #55 #56 #57 #58 #60 #61 #62 #63 #64 #65 #66 #67 #68 #69 #70 #71.
Cleared by one word: #60 (fixed and verified four times), and #61 now that half of it has been done for you. Cleared by one small edit each, wording already in the row: #52, #55, #56, #66(a), #67, #68. Not the routine's to clear: #58 and #70, 42 deletions between them. The rest turn on a decision.
The Dictionary rows are the ones to settle first, and the reason is now arithmetic rather than argument. #62(a), #62(b), #63 and #64 all touch the same definition lines of the same 211 entries, and the wiki has just written 270 entries in a sixth language to the opposite convention on all three counts. Every further language seeded before these are settled adds another corpus to the majority and leaves Proto-Herta-Olaic further alone. Three of the four are one mechanical pass each and the routine can do them the day they are approved.
And one on Tacchoan. Five open rows now propose edits to the same six lines of that one section — #53(c), #55, #56, #57 and #67. Approved together they are one edit; approved separately they are five.