A few inverted expectiations

[krlmlr]

• expect_success() -- if failure should not abort the test
• expect_less_than_or_equal() and expect_more_than_or_equal()

With tests. Modeled as "new-style" expectations, the arguments of the _or_equal variants differ.

Also adds expect_silent() to the "expectations" family.

expect_unnamed() seems unnecessary, as expect_named() can already test for absence of names.

@hadley: Do you want to export expect() at some point?

Fixes #303.

[hadley]

I think you accidentally included this line

[krlmlr]

Good catch. I was wondering if this function necessary at all, we already have expect_error(..., NA)

[hadley]

Yeah, that's a good point.

[krlmlr]

Removed expect_success().

Another option would be to support an optional argument: expect_less_than(..., allow_equal = FALSE). This would be more in spirit of the other functions: expect_named(..., NULL), expect_error(..., NA), expect_warning(..., NA), ...

[hadley]

I'd also feel better if they were named expect_lt, expect_lte etc. The current names are a bit verbose.

[krlmlr]

Added shorter versions: lt, lte, mt, mte

[hadley]

I think gt (greater than) is the more common abbreviation

[krlmlr]

Now using gt.

[hadley]

Thanks!