@tikoci/centrs
    Preparing search index...

    Function analyzeCoordinates

    • Build the coordinate analysis of original: its UTF-8 encoding, the byte-count-preserving analyzed surface, and one CharRun per code point locating it in all three spaces.

      An ALL-ASCII input takes the identity path. Nothing about the contract changes there — the three spaces coincide, which is what the byte-count-preserving rule above was chosen to guarantee — so the mapping is arithmetic and the per-character CharRun array is deferred to its first reader. The analyzer re-derived that map for every nested block body it descended into, and pathresolve's own locate discards the coordinate facts of a body anyway (they are properties of the root input): on a 60 KiB document nested 64 deep it decoded 7.9 MiB and allocated a run object per byte of it, a quarter of total analysis time (#322).

      Parameters

      • original: string

      Returns CoordinateAnalysis