How do I share a genhtml coverage report?

Publish the entire output directory. genhtml writes `index.html` plus one HTML page per source file and a shared CSS — the drill-down is the report, so a single file is useless on its own. Drag the folder into Comma and the links keep working.

Which files does lcov actually produce?

`lcov --capture` produces a `.info` tracefile, which is data, not a report. `genhtml coverage.info --output-directory out` turns it into browsable HTML: `out/index.html`, a directory page per source folder, a page per file, and `gcov.css`.

Does this work for JavaScript coverage too?

Yes. `nyc`/Istanbul's `lcov` reporter writes `coverage/lcov-report/` in the same shape — an `index.html` with per-file pages — and publishes the same way.

How do I publish it from CI?

Generate the HTML in the job and PATCH the same report id every run. The URL stays stable, each run appends a revision, and reviewers can diff two revisions to see whether a merge moved coverage.

Can someone comment on a specific uncovered line?

Yes. Threads anchor to the rendered content, so 'this branch is unreachable, delete it' sits on that line in the source view instead of in a pull-request comment nobody can map back.

Share an lcov coverage report

lcov gives you a tracefile; genhtml turns it into the thing people actually read:

lcov --capture --directory . --output-file coverage.info
lcov --remove coverage.info '/usr/*' --output-file coverage.info
genhtml coverage.info --output-directory coverage-html

What you get is coverage-html/index.html, a page per directory, a page per source file with the hit counts in the gutter, and gcov.css. The value is the drill-down — which is precisely what doesn't survive being sent as an attachment.

Publish the directory

Drag coverage-html/ (or a zip of it) into the app:

index.html becomes the report body.
Every per-file page and the stylesheet upload alongside, and the relative links between them are rewritten to the uploaded copies — so clicking through from the summary to src/parser.c still works.
Nothing needs a local web server, which is the other reason these reports usually stay on one machine.

Same shape for JavaScript: nyc report --reporter=lcov writes coverage/lcov-report/, and it publishes identically. For the Python coverage html case, see sharing an htmlcov folder.

From CI

genhtml coverage.info --output-directory coverage-html

curl -fsS -X PATCH "https://commareports.com/api/v1/reports/$REPORT_ID" \
  -H "Authorization: Bearer $COMMA_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d "$(jq -n --rawfile html coverage-html/index.html \
        --arg title "Coverage — $GIT_COMMIT" '{title: $title, html: $html}')"

One report id per branch keeps one URL and appends a revision per run. The interesting question about coverage is never the number, it's the direction — and a revision history answers that without a dashboard product. See publishing from CI.

What review adds

Anchored threads on an uncovered branch, in the source view — commenting on HTML.
Revisions, so "coverage dropped" is a diff rather than an assertion.
Access per report — private, team, domain-gated or named reviewers. See the sharing model.

Limits

Entry HTML: 5 MB. Assets: 25 MB per file, 250 MB and 500 files total. A large C++ tree can exceed 500 pages — genhtml on a subdirectory, or the --prefix and --exclude flags, keeps the published report to the part under review.
60 requests/minute per token.

Try it

Comma is free — unlimited reports, unlimited commenters, unlimited revision history.

Publish a coverage report →