Add test case with a huge file

[papandreou]

Looks like 095c82f broke an essential feature of the maxRows PR. It is delaying the teardown of the stream to the next tick by moving it to promise.catch(...) here: 095c82f#diff-678e4bd9f1bcf0cc0c17eee7684618abR101

That means that the parser will still busyloop on the remaining rows of the current unzipped chunk, which can be quite a lot.

I've added a test with a huge .xlsx that runs for over 30 seconds on my machine, but less than a second when I revert the tidy commit.

I'm not sure exactly how to fix it without essentially reverting most of it, so for now I'm just submitting the test :)

[guyonroche]

@papandreou yes - I should have read your comments more closely ;-)
I've fixed it and as soon as Travis agrees, I'll publish

[guyonroche]

@papandreou Travis didn't agree until I upped the timeout to 20s. I'm not overly happy with that but it was only a problem in node v6 & v7 so I'll let it slide for now.
published in 1.4.5

[papandreou]

@guyonroche, thank you so much for the prompt response and the fix 🎉