fix typo

[sanga]

Apparently test coverage for this is also not what it should be, otherwise this would've been caught. I can look into providing some tests for that this weekend perhaps (time allowing).

In other news I think this takes the record for my smallest bugfix ever :)

[Jahaja]

Thanks! Yeah, we'll need to improve the test coverage.

[heyman]

Yeah, I was about to create a test case that would trigger the locust_error event the other day when you did the #114 pull request, but realized that we have a big catch-all try-catch that makes it a little tricky.

While writing this comment I couldn't help myself from implementing a test case covering this anyway :).

I introduced a private variable Locust._catch_exceptions which defaults to True, but I'm also thinking that perhaps we should add a command line option for disabling the exception catching, since it might be nice to use during development of your load testing scripts.

[sanga]

"While writing this comment I couldn't help myself from implementing a test case covering this anyway :)" Excellent. Cheers

"thinking that perhaps we should add a command line option for disabling the exception catching": well those exceptions are still logged right? I've usually just fixing such problems by dumping those logged exceptions out to the terminal (of course locust spits a few of them - so catch_exceptions=False would make that nicer).

But yeah, I think some things could be done to make scenario development (and maintainence) easier. For example dropping into a debugger on exceptions would be nice (I guess the catch_exceptions thing would help with this) as it's not always immediately obvious as to the source of the exception. Though I have some hazy recollection debugging gevent to not be so trivial.