Known limitations#

• Comment detection is lexical: comment markers inside string literals are treated as comments (e.g. a # in a Python string, or a // in a JS string inside an embedded <script>, can define a phantom item). URL-shaped text (a scheme followed by ://) is exempt - markers inside a URL never open a comment - which conversely means a comment opener written directly against a scheme://-shaped token (e.g. note://… with no separating whitespace) is not recognized.
• Input paths are compared exactly and case-sensitively, after only lexical normalization (. and .. segments, separators): no case folding, no symlink resolution. On a case-insensitive file system, or through a symlink, the same file reached under two spellings (e.g. docs/spec.md and DOCS/spec.md) is read twice, and its items are reported as duplicates (see Text handling).