JavaScript split（）仅在Firefox / IE中向数组添加额外的项

Question

I'm trying to understand the behaviour of different browsers with the JavaScript split() method, matching by linefeed (\\n) regexp.

I have a textarea input form which is to take rows pasted from Excel and split them by linefeed into an array:

var rowsplit = document.getElementById("inputfield").value.split(/\n/g);

The user selects the rows in Excel by the side bar and so the first 11 columns have data and there are a whole bunch of empty columns after those to the edge of the spreadsheet that I slice() out later. So a typical input from Excel would be:

[data][data][data][data]x11[null][null][null]etc until edge of spreadsheet

The difference is what is matched by the regexp and pushed into the array.

In both Chrome 28 and Safari 5.1, when selecting three rows of information, it correctly matches the \\n as 3, outputs them to the array and goes frolicking through a daisy field without a care in the world.

In IE10 and Firefox 19 though, it matches the first three linefeeds fine and then pulls an extra one which is blank. So the \\n expression splits it into four items, crashes into a building and kills hundreds.

I have read a lot about different browsers handling the CRLF characters differently and tried splitting by \\r\\n and just \\r but it seems to introduce more trouble. I also copied the Excel rows into Notepad++ and turned on visibility of the CRLF characters to see that there only appears to be 3, so why are IE and Firefox adding another one?

I've also read a lot about the pains people have with regexp, JS, browsers, line breaks AND Excel but most of it seems to be more relevant to older browsers so I was hoping someone with knowledge on the current state of things could provide some insight.

Thanks!

Answer 1

Presumably, IE10 and Firefix 19 see the value as having an extra, trailing line-break. .split() won't ignore it just because it's at the end.

You should be able to remove it with .trim() before .split() 'ing:

var rowsplit = document.getElementById("inputfield").value.trim().split(/\n/g);

Another option is to use a zero-width look-ahead to only match \\n that have anything after them:

var rowsplit = document.getElementById("inputfield").value.split(/\n(?=\s*\S)/g);

Example of each: http://jsfiddle.net/FChES/