chore(deps): update dependency kounex/update-day to v0.12.0

[renovate]

This PR contains the following updates:

Package	Update	Change
kounex/update-day	minor	0.9.32 → 0.12.0

---

Warning

Some dependencies could not be looked up. Check the Dependency Dashboard for more information.

---

Release Notes

kounex/update-day (kounex/update-day)

v0.12.0

Compare Source

Full audit of every user-facing command - clarity, correctness, and message quality.

Added

• /delete now asks for confirmation (Delete/Cancel buttons) before actually deleting an Observe, instead of doing it immediately on a single command.

Changed

• /observe's scrape-interval is now optional (it already had a documented "Hourly is the default" - it just wasn't actually possible to omit it before).
• Clarified the keep-active description on /observe and /edit - it previously read "If you want to deactivate the Observe once it found a change, true by default", which was backwards on both the polarity (true keeps it active, it doesn't deactivate it) and the stated default (it's false).
• /scrape (manual trigger) now goes through the same persistence/notification pipeline as the background scheduler, instead of bypassing it entirely - it used to leave lastScrapeAtMS/amountScraped stale, skip auto-deactivation on a found change, and could cause the same change to be DM'd to the user twice (once never, once later from the scheduler catching up).
• Uncaught-error messages shown to users no longer include raw internal error names/messages/cause - just a friendly "something went wrong, this has been logged" (full detail still goes to the debug log).
• Removed two unused Discord gateway intents (GuildMessageReactions, GuildVoiceStates) left over from the bot this project was originally forked from.
• Boolean fields in embeds (Active, Keep Active) now show ✅/❌ instead of the literal text true/false.
• Deduplicated near-identical autocomplete logic across /delete, /edit, /list, /reactivate, and /scrape into one shared helper.
• Assorted message-copy fixes: missing spaces after inline code in a few notifications, "on of" → "one of", "a Observe" → "an Observe", "it's limit" → "its limit".

Fixed

• /admin list-observes's this-guild/active-only options were documented as "true by default" but actually defaulted to showing every guild's Observes, including inactive ones, when left unset - and explicitly passing this-guild: false did the opposite of what it should. An admin running this with no arguments, expecting just their own guild's active Observes as documented, would see every other guild's data on this bot too.
• /reactivate didn't reset consecutiveTimeouts, so an Observe deactivated for hitting the timeout limit would immediately re-deactivate after just one more timeout instead of getting a fresh run of the full limit.
• Settings embed's "last updated" footer showed a raw millisecond timestamp instead of a formatted date.
• buildObserveOverview (used by /admin list-observes) fetched each user's Discord profile one at a time in a loop instead of in parallel.
• Multi-Observe embed fields (/list, /admin list-observes) had no guard against Discord's 1024-character field-value limit - a guild/user with enough Observes would have hit an API error building the embed. Now truncates safely with a "…and N more" note, keeping all columns aligned to the same row count.

v0.11.2

Compare Source

Changed

• /edit no longer requires re-entering every field. Only current-name (which Observe to edit) is required - name, url, text, scrape-interval, css-selector and keep-active are all optional now, and any left empty keep their current value instead of being overwritten or reset to a default. This also fixes a latent bug where editing any field silently reset keep-active back to false if you didn't re-specify it.
• css-selector can be explicitly cleared back to whole-page search by typing none.

v0.11.1

Compare Source

Fixed

• puppeteer.launch() had no --no-sandbox flag, so Chrome crashed immediately in this container (missing sandbox privileges) - every single scrape was silently hitting this and surfacing as a generic timeout. This was breaking every Observe, on every site.
• Some sites (this container's User-Agent literally contains HeadlessChrome) block requests from an obviously-automated browser - now presented as regular desktop Chrome (same bundled build, just without the giveaway tag).
• Pages using CSS content-visibility: auto (a real, increasingly common perf optimization) render as empty text until scrolled into view - the target element/scope is now scrolled into view before reading its text, same as a real visitor would encounter it.

v0.11.0

Compare Source

Changed

• /observe and /edit no longer require a CSS selector. By default the bot now searches the whole page's visible text for the phrase you give it; css-selector is now optional and only needed to narrow the search if that phrase could also show up elsewhere on the page (nav, other products, etc). This removes the main pain point of finding a precise, unique selector for text buried in a deeply nested/generic tag.
• Renamed the current-text option to text and the underlying currentText column to watchText to match its actual meaning: the phrase you're watching for, which you'll be notified about once it's no longer found on the page.
• Removed the dom-element-property option entirely — it only made sense when pointing at one exact element, which conflicts with the new default of searching a broader area, and nothing was using it.

v0.10.0

Compare Source

Added

• Retry with backoff around page navigation during a scrape, instead of failing on the first transient network blip.
• Observes now actually get deactivated (and stay deactivated) once they hit the consecutive-timeout limit, instead of just claiming to be in the notification while continuing to time out forever.

Changed

• Bumped base image to node:22-slim (Node 18 is EOL) and updated discord.js, puppeteer, inversify, prisma, execa, typescript, and related dependencies to current versions.
• Puppeteer's Chrome download now happens once at image build time instead of on every container start.
• tini is now actually wired up as the container's PID 1 (ENTRYPOINT) for proper signal handling/zombie reaping — it was previously installed but unused.
• Removed a batch of unused dependencies and dead config (YOUTUBE_API_KEY, SPOTIFY_CLIENT_ID/SECRET, and other leftovers from the bot template this project was originally forked from).

Fixed

• Fixed a bug where completing a scrape early could wipe in-progress tracking for every later-scheduled scrape (Array.prototype.splice called without a delete count), causing duplicate/concurrent scrapes.
• Fixed unhandled promise rejections in the scrape scheduler and guildCreate handler that could crash the entire bot process (e.g. a user with DMs disabled).
• Fixed a Puppeteer browser/process leak when a scrape errored before reaching its own browser.close() call.
• Fixed a missing break in the /scrape command's result handling that caused a detected change to display as "No change".
• Corrected the guild-join welcome message, which linked to a different bot's (Muse) permissions wiki page.

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • "on saturday"
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.