mirror of
https://github.com/MBombeck/HealthLog.git
synced 2026-11-10 13:22:14 +00:00
Marc-André Bombeck
98fb6d81c0
While two pdf.js versions broke each other, every PDF preview rendered after a local text read failed and was recorded as raster-failed, which the thumbnail backfill treats as final. Migration 0478 deletes those failure rows, so the boot-time thumbnail backfill finds each document again and renders it once. A stored failure cannot tell the clash from a page that really renders blank; such a page fails once more and settles with one attempt. Only raster-failed is cleared: the clash surfaced as a failed rasterization, which the thumbnail path maps to that reason alone. Refs #1277
21 lines
1.1 KiB
SQL
21 lines
1.1 KiB
SQL
-- Give every PDF preview that failed to rasterize one more render.
|
|
--
|
|
-- Until this release the server could load two pdf.js versions. Once a local
|
|
-- text read had claimed the in-process pdf.js worker, every later page render
|
|
-- failed until a restart, and the thumbnail job recorded those documents as
|
|
-- `raster-failed`. The backfill treats that reason as final, so their previews
|
|
-- would never be tried again.
|
|
--
|
|
-- Only `raster-failed`: the clash surfaced as a rasterization that returned a
|
|
-- failure, and the thumbnail path maps that to this reason and to no other.
|
|
-- `error` comes from a throw, and the rasterizer never throws; `empty-render`
|
|
-- and `pixel-cap` are decided after a render that worked.
|
|
--
|
|
-- A stored failure cannot say whether the clash caused it or the page really
|
|
-- renders blank, so every such row goes. The boot-time thumbnail backfill
|
|
-- finds each document again, renders it once, and a page that is really blank
|
|
-- records `raster-failed` with one attempt and settles where it was. Running
|
|
-- this again costs one more render per such document and nothing else.
|
|
DELETE FROM "document_thumbnail_failures"
|
|
WHERE "reason" = 'raster-failed';
|