[nvxbug]

id() only quotes identifiers that aren't plain alphanumeric, with one narrow exception for the 0x hex-address prefix. Any other alphanumeric name that starts with a digit still goes out unquoted, so a function or module named like 3d, 1e5 or 123abc lands unquoted where a dot node id belongs. dot lexes a digit-leading token as a numeral, so 3d splits into 3 and d: the node declaration and the edges referencing it resolve to different phantom nodes, and the real node is dropped or mislabeled.

Widen the existing 0x guard to reject any identifier whose first character is a digit, routing these names through the quoted path. Keeping the check in id() covers every node id, edge endpoint and label from the single place they are all written. Pure integers get quoted too, which is harmless since a declaration and its references stay consistent.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.98%. Comparing base (d1ba70a) to head (ccc929f).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #116   +/-   ##
=======================================
  Coverage   84.98%   84.98%           
=======================================
  Files           1        1           
  Lines        2337     2337           
=======================================
  Hits         1986     1986           
  Misses        351      351           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nvxbug]

any update?

[jrfonseca]

Makes sense. Thanks.