You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Use TinyTeX's maintained regex.json for LaTeX error detection #13612
We could use TinyTeX's patterns as our source of truth for LaTeX error detection. This would keep us synchronized with battle-tested patterns that are updated daily by the RStudio team.
We could automate periodic sync to maintain alignment without manual effort.
Test Coverage (tests/unit/latexmk/parse-error.test.ts)
100% test coverage (19/19 error examples)
Why Not Fully Automated?
TinyTeX patterns are designed for line-by-line matching in R code, while Quarto matches against the entire log as a single string. This fundamental difference requires careful pattern adaptation:
Multiline flag adjustments
Anchor modifications
Filter function updates
With ~4 pattern updates per year, manual adaptation after automated detection has proven practical.
Current Status
We're synchronized with TinyTeX (automated detection) but not fully automated (manual adaptation). This issue remains open for potential future work on complete automation, though the current semi-automated approach works well in practice.
Activity
cderv commentedon Nov 26, 2025
We've implemented a semi-automated approach to this in PR
What Was Implemented
Daily Verification Workflow (
.github/workflows/verify-tinytex-patterns.yml)regex.tar.gzfrom tinytex-releasesPattern Implementation (
src/command/render/latexmk/parse-error.ts)Documentation (
dev-docs/tinytex-pattern-maintenance.md)Test Coverage (
tests/unit/latexmk/parse-error.test.ts)Why Not Fully Automated?
TinyTeX patterns are designed for line-by-line matching in R code, while Quarto matches against the entire log as a single string. This fundamental difference requires careful pattern adaptation:
With ~4 pattern updates per year, manual adaptation after automated detection has proven practical.
Current Status
We're synchronized with TinyTeX (automated detection) but not fully automated (manual adaptation). This issue remains open for potential future work on complete automation, though the current semi-automated approach works well in practice.