7034: Fix rendering failure for consecutive LaTeX equations in Mermaid text nodes

[darshanr0107]

📑 Summary

This PR fixes an issue where consecutive LaTeX equations inside Mermaid text nodes fail to render correctly.
When multiple inline or block LaTeX expressions appear back-to-back within a node, the rendering pipeline
interprets them incorrectly, resulting in broken or missing output.
Resolves #7034

📏 Design Decisions

Describe the way your implementation works or what design decisions you made if applicable.

📋 Tasks

Make sure you

• 📖 have read the contribution guidelines
• 💻 have added necessary unit/e2e tests.
• 📓 have added documentation. Make sure MERMAID_RELEASE_VERSION is used for all new features.
• 🦋 If your PR makes a change that should be noted in one or more packages' changelogs, generate a changeset by running pnpm changeset and following the prompts. Changesets that add features should be minor and those that fix bugs should be patch. Please prefix changeset messages with feat:, fix:, or chore:.

[changeset-bot]

🦋 Changeset detected

Latest commit: 1031be9

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
mermaid	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[netlify]

✅ Deploy Preview for mermaid-js ready!

Name	Link
🔨 Latest commit	1031be9
🔍 Latest deploy log	https://app.netlify.com/projects/mermaid-js/deploys/6a1d67f8d9667c0008318175
😎 Deploy Preview	https://deploy-preview-7235--mermaid-js.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.
🤖 Make changes	Run an agent on this branch

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[pkg-pr-new]

Open in StackBlitz

@mermaid-js/examples

npm i https://pkg.pr.new/@mermaid-js/examples@7235

mermaid

npm i https://pkg.pr.new/mermaid@7235

@mermaid-js/layout-elk

npm i https://pkg.pr.new/@mermaid-js/layout-elk@7235

@mermaid-js/layout-tidy-tree

npm i https://pkg.pr.new/@mermaid-js/layout-tidy-tree@7235

@mermaid-js/mermaid-zenuml

npm i https://pkg.pr.new/@mermaid-js/mermaid-zenuml@7235

@mermaid-js/parser

npm i https://pkg.pr.new/@mermaid-js/parser@7235

@mermaid-js/tiny

npm i https://pkg.pr.new/@mermaid-js/tiny@7235

commit: 1031be9

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 3.27%. Comparing base (772b219) to head (1031be9).
⚠️ Report is 11 commits behind head on develop.

Additional details and impacted files

@@           Coverage Diff           @@
##           develop   #7235   +/-   ##
=======================================
  Coverage     3.27%   3.27%           
=======================================
  Files          599     599           
  Lines        61064   61064           
  Branches       921     921           
=======================================
  Hits          1997    1997           
  Misses       59067   59067           

Flag	Coverage Δ
unit	3.27% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
packages/mermaid/src/diagrams/common/common.ts	31.89% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[argos-ci]

The latest updates on your projects. Learn more about Argos notifications ↗︎

Build	Status	Details	Updated (UTC)
default (Inspect)	👍 Changes approved	1 added	Jun 1, 2026, 11:22 AM

[ashishjain0512]

[sisyphus-bot]

Thanks for chasing this one down — the diagnosis is spot on and the patch is exactly the kind of surgical fix we love to see. The original /\$\$(.*)\$\$/g is greedy, so on "From $$x(t)$$ to $$y(t)$$" it ran from the first $$ to the last, capturing the prose in between and handing KaTeX something it can't parse. Switching to a lazy .*? quantifier makes each $$...$$ match the smallest delimiter pair, which is what KaTeX actually expects.

What's working well

• 🎉 [praise] common.ts:290 — minimal, precise change. The lazy quantifier is the textbook fix here, and it also incidentally improves the fallback paths (common.ts:327, common.ts:356-358) so multiple equations get individually replaced with the "MathML unsupported" / "tiny" messages instead of being collapsed into one match.
• 🎉 [praise] cypress/integration/rendering/katex.spec.js:36-44 — nice that the test pairs a passing case (single equation per node, A/B) with the failing case (two equations in one node, C/D) so the visual diff makes the regression obvious if it ever comes back.

Things worth considering (non-blocking)

• 🟢 [nit] common.ts:290 — a focused unit test for hasKatex/katexRegex would be a nice complement to the visual test. Visual snapshots are great for confirming the rendering pipeline end-to-end, but a unit test like expect("$$a$$ x $$b$$".match(katexRegex)).toHaveLength(2) runs in milliseconds and pinpoints exactly which layer broke if someone touches the regex again. Totally optional — the cypress test already covers the regression — just a low-cost extra safety net.
• 💡 [suggestion] common.ts:352 — out of scope for this PR, but the sibling /<annotation.*<\/annotation>/g is still greedy. It happens to be safe today because, with your fix, each katex.renderToString call now runs on a single equation rather than a giant glued-together blob, so there's only ever one annotation block to strip. Mentioning it only so a future reader doesn't copy the pattern elsewhere; no need to touch it here.

Security

XSS sub-agent review: no issues identified. The captured group flows into katex.renderToString(..., { throwOnError: true, ... }) which rejects arbitrary HTML and emits its own structured output, and renderKatexSanitized (common.ts:369-374) still wraps everything in DOMPurify. The lazy quantifier doesn't introduce a new sink — if anything, the trust boundary is tighter now that each equation is rendered in isolation.

Heads up that since this adds a brand-new cypress test, the Argos baseline will need a maintainer to accept the first snapshot — that's normal for this kind of PR and not something for you to action.

Nice work. Approving — let's get this across the finish line.