I went into a rabbit hole after reading Surma's article "Is postMessage Slow?" and became curious to know the feasibility of using an ArrayBuffer with PROXX for performant postMessage() calls. My thoughts were to provide native JS property access while maintaining behavior of interface Cell to allow usage like: grid.cells[col][row].hasMine = true.
Do you think the approach below could provide benefits or is the current patchset implementation already performant enough? Instead of cluttering this thread with code, I wrote an independent POC to test the transferability of Cell states in a 40x40 grid. Start at CellGrid and Cell in branch: master...AnthumChris:arraybuffer-poc
I went into a rabbit hole after reading Surma's article "Is postMessage Slow?" and became curious to know the feasibility of using an
ArrayBufferwith PROXX for performantpostMessage()calls. My thoughts were to provide native JS property access while maintaining behavior ofinterface Cellto allow usage like:grid.cells[col][row].hasMine = true.Do you think the approach below could provide benefits or is the current patchset implementation already performant enough? Instead of cluttering this thread with code, I wrote an independent POC to test the transferability of
Cellstates in a 40x40 grid. Start atCellGridandCellin branch: master...AnthumChris:arraybuffer-poc