- Add WEIGHTING_SCHEMA_VERSION constant; bump to force a global lazy reindex
- Migrate legacy Record<id, string> storage to { weight, fingerprint, pluginVersion }
- Fingerprint per-assessment on status, graded, availability, score, due, title (sourced from the React fiber)
- Refetch weighting only when fingerprint or schema version mismatches
- Preserve previous weight as refreshing placeholder during background refetch
- Render subject average immediately from cache; run parseAssessments off the critical path
- Coalesce concurrent renderSubjectAverage calls instead of dropping them
- Dispatch betterseqta:weightingsChanged on refetch start and completion
- Show row-level refresh indicator and "Refreshing weightings" notice while refreshing
- Leave weightingOverrides untouched by all reindex paths
This reverts commit 2aecd63850.
Reverting so that I can solve the indexing issue. Only marked assessments are getting indexed, which is incorrect behaviour that slipped testing when the plugin was first made.
- Adds weight label below each assessment
- Correlates assessment name with its ID, so each dom element can be identified: {trimmed_title: assessmentID}
- Slightly changed regex to be a little more permissive
- Changed pdfjs src/type due to CORS issues on firefox
- Permitted weightings to be zero, but not less than zero
- Modified how assessment items are iterated through, as the previous approach assumed they're in the same order as they are in react
- Changed parseAssessments() to immediately dispatch parsing for all pdfs asynchronously, as doing it serially is painstakingly slow
- Discard useless decimals when displaying weight (.0)
Add average grade display and also fix the CORS violation caused by pdfjs trying to load PDFs from URLs that Firefox extensions can't access.
fixed by instead:
- Fetching the PDF as an ArrayBuffer directly from the URL
- Passing the ArrayBuffer to pdfjs using { data: arrayBuffer } instead of passing a URL
Jaxx7594
AdenMGB