Add support for 0x50 firstByte

[mgcrea]

I've got .xlsx files that are rejected by the lib Unsupported file. Any idea what I should start looking for? Tried to add:

case 0x50: return parse_xlml(d, (o.type="buffer",o));

But it breaks at:

node_modules/xlsjs/xls.js:6366
            switch(state[state.length-1][0]) {
TypeError: Cannot read property '0' of undefined
    at parse_xlml_xml (node_modules/xlsjs/xls.js:6366:32)

Great job by the way. Would love to switch all future developments to a lib with a "clean" license.

---

Found this stackoverflow answer:

If the files are in Office 2007 format (e.g. .docx), then their internal storage is either:

1) A zip file of xml docs (if it's not password protected)

2) The old style compound file format (if it IS password protected).

Therefore you could probably do something like this:

1) Check the first few bytes of the file

2) If it's a zip file (non password protected), it'll start with 0x50 0x4b 0x03 0x04.

3) If it's not a zip file, then it's probably password protected. It will start with a different binary signature (e.g. Word 2007 docs start with 0xd0 0xcf 0x11 0xe0 in this case)

Basically, if it's a new .docx or .xlsx, and it DOESN'T start with the zip signature of 0x50 0x4b 0x03 0x04, it's probably password protected.

For other versions of MS Office, it's a bit trickier...

So it looks like it's a standard zip file.

[mgcrea]

Used the wrong npm repo xlsjs. Closing while I'm trying to reproduce against correct HEAD.

[SheetJSDev]

@mgcrea If you are trying to handle both pre-2007 and 2007+ excel files in node, I'd recommend using j: https://www.npmjs.org/package/j . The read functions embed the file type check: https://github.com/SheetJS/j/blob/master/j.js#L7-L32

[mgcrea]

Thanks! Will do that.

I've deprecated my node-xlsx module by the way. Once again, awesome work. Thanks!