fix: keep Temporal workflow when a scheduled post is edited without a time change

[giladresisi]

Problem

Editing a future scheduled post — via the dashboard or the public API (POST /public/v1/posts with the existing ids) — sent type: 'schedule', which always terminated and recreated the post's Temporal workflow, even when only the content or settings changed and the publish time stayed the same. At 100k+ sleeping workflows that is needless churn (terminate + start per edit), and it diverged from the agent's updatePostTool, which already leaves the workflow alone for non-time changes.

Fix

Decide it server-side in PostsService.createPost: capture the post's current publish time before the upsert overwrites it, and only (re)start the workflow when it's a new post or an existing QUEUE post whose time actually moved. Content and settings are re-read from the DB at publish time, so an unchanged time needs no restart; a genuine time change still restarts it.

Done in the service (not the edit modal) so it's authoritative and covers every caller — dashboard and public API — and can't be fooled by client-side date state.

Testing (repro + validate, both surfaces)

Counted Temporal runs for the post before/after an identical time-unchanged edit:

Surface	Without fix (main)	With fix
Dashboard edit (settings only, same time)	new workflow created	no new workflow
POST /public/v1/posts (same ids, same time)	1 → 2 runs	2 → 2 runs

A real date change still recreates the workflow (TERMINATE_EXISTING, one new run) in both cases. Decision logic also spot-checked against real stored dates (unchanged → skip, moved → restart).

🤖 Generated with Claude Code

[postiz-contribution]

Contribution-checker quality warning
Heuristic score: 0/100 (low). This is a non-blocking warning surfaced by the project's quality settings.

Heuristics that flagged:

• Excessive inline code references: 7 inline refs (>3)
• PR doesn't use the repo's PR template: 5 required checkbox items missing (max 1, match ≥80%)
• Body adds too many extra headers beyond the template: 3 extra headers (>1)
• AI watermark phrase: Matched: "Generated with Claude Code"
• Commit message too long: Longest: 1056 chars
• Excessive added comments: 7 added comment lines (ratio 0.44)

If this is a genuine contribution, please add detail to your PR description and tighten the diff scope before reviewers look at it.

[postiz-agent]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (0)
✅	Open Source Security	0	0	0	0	0 issues
✅	Licenses	0	0	0	0	0 issues
✅	Code Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

[sentry]

Bug: On non-UTC servers, creating a scheduledDate with dayjs().format() and parsing it with dayjs.utc() causes an incorrect time comparison for posts with type: 'now'.
_{Severity: MEDIUM}

Suggested Fix

Ensure timezone consistency. When generating the scheduledDate for type === 'now', use dayjs.utc() to create a UTC-based time string. Replace dayjs().format('YYYY-MM-DDTHH:mm:00') with dayjs.utc().format('YYYY-MM-DDTHH:mm:00') to align the generated time with the UTC-based comparison logic used later.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: libraries/nestjs-libraries/src/database/prisma/posts/posts.service.ts#L903

Potential issue: A timezone mismatch occurs when editing a post with `body.type ===
'now'` on a server not running in the UTC timezone. The code generates a `scheduledDate`
using `dayjs().format()`, which creates a string representing the server's local time
but lacks a timezone offset. This string is later parsed by `dayjs.utc()`, which
incorrectly assumes the time is in UTC. This discrepancy causes the time comparison
against the existing `publishDate` to fail, leading to an unnecessary workflow restart
even when the scheduled time has not been modified.

_{Did we get this right? 👍 / 👎 to inform future reviews.}

[giladresisi]

Fixed with the new commit

[postiz-contribution]

Contribution-checker quality warning
Heuristic score: 0/100 (low). This is a non-blocking warning surfaced by the project's quality settings.

Heuristics that flagged:

• Excessive inline code references: 7 inline refs (>3)
• PR doesn't use the repo's PR template: 5 required checkbox items missing (max 1, match ≥80%)
• Body adds too many extra headers beyond the template: 3 extra headers (>1)
• AI watermark phrase: Matched: "Generated with Claude Code"
• Commit message too long: Longest: 1056 chars
• Excessive added comments: 7 added comment lines (ratio 0.50)

If this is a genuine contribution, please add detail to your PR description and tighten the diff scope before reviewers look at it.

[giladresisi]

Fixed in a followup commit

[giladresisi]

Partially right — normal group edits update rows in place via upsert (ids preserved), so no new records are created and the existing workflow stays keyed to the same root; and when records genuinely are recreated, existingRoot is null so the skip can't fire. The real gap was an edit that promotes a child row to root (thread head removed/reordered) — fixed in 043e2da by requiring existingRoot to have no parentPostId.