mirror of
https://github.com/discourse/discourse.git
synced 2026-08-04 10:39:43 +08:00
5b6d87a4c9Previously, every SQL query finding translation candidates (admin progress report and backfill jobs) filtered on `LENGTH(posts.raw)`, which forced detoasting `raw` for every scanned row and made those queries slow, while the limit itself no longer guarded anything — long posts are already split into chunks sized to the model's `max_output_tokens`, and core's `max_post_length` bounds raw size. This change removes the setting and its query filters, deletes any stored override via a post-deploy migration, and bumps the translation progress cache versions since eligibility counts change. Note for review: on existing sites, posts previously excluded by the limit become eligible for backfill, so pending counts (and translation spend on long-form content) will increase accordingly. |
||
|---|---|---|
| .. | ||
| base_candidates.rb | ||
| base_translator.rb | ||
| category_candidates.rb | ||
| category_locale_detector.rb | ||
| category_localizer.rb | ||
| content_splitter.rb | ||
| entry_point.rb | ||
| language_detector.rb | ||
| localizable_quota.rb | ||
| post_candidates.rb | ||
| post_detection_text.rb | ||
| post_locale_detector.rb | ||
| post_localizer.rb | ||
| post_raw_translator.rb | ||
| progress.rb | ||
| short_text_translator.rb | ||
| sidebar_section_candidates.rb | ||
| sidebar_section_locale_detector.rb | ||
| sidebar_section_localizer.rb | ||
| tag_candidates.rb | ||
| tag_locale_detector.rb | ||
| tag_localizer.rb | ||
| topic_candidates.rb | ||
| topic_locale_detector.rb | ||
| topic_localizer.rb | ||
| topic_title_translator.rb | ||
| verbose_logger.rb | ||
