🐛 Bug Report

Files with too many merged cells will not load - they take too long.

Lib version: 4.4.0

Steps To Reproduce

1. Create an Excel book with a worksheet.
2. In the worksheet, add a large amount of rows where one cell is merged on each row. See attached file.
   toomanymerges.xlsx
3. Try to parse the workbook using:

  const buffer = fs.readFileSync(filename);
  const workbook = new Workbook();
  console.time('Buffer');
  await workbook.xlsx.load(buffer);

4. Experience how it takes a long long time to parse.

The expected behaviour:

It should not be so slow. The attached file takes 1.5 minutes to parse in my Node environment. I have encountered worse examples from users of our product, and it runs in the browser, so those instances have just never finished.

Possible solution (optional, but very helpful):

I have noticed that the slow culprit is the _mergeCellsInternal function inside worksheet.js, specifically the check that cells aren't already merges. Which makes sense, because it is built with n squared loops. See my discussion here: #2568

// worksheet.js, ._mergeCellsInternal
    // check cells aren't already merged
    _.each(this._merges, merge => {
      if (merge.intersects(dimensions)) {
        throw new Error('Cannot merge already merged cells');
      }
    });

The above code is the culprit. When I comment out this code, parsing is very fast again. Is there a way this could be more efficient? Is it even necessary? Is it necessary on first parse of the document?