Better handling of repeated, identical expressions

[joshuaulrich]

This was reported as olafmersmann#7.

---

If the same expression is passed into microbenchmark() twice, then the resulting error is not very helpful.

library(microbenchmark)
microbenchmark(1 + 1, 1 + 1)
## Error in `levels<-`(`*tmp*`, value = if (nl == nL) as.character(labels) else paste0(labels,  : 
  factor level [2] is duplicated

(On older versions of R, you get the warning duplicated levels in factors are deprecated).

I can't think of a good reason why you would want to include the same expression twice, so this usually means a copy and paste error. Consequently, you should throw a human-readable error or warning before the expressions are evaluated, explaining the problem. Some thing like

The same expression, 1 + 1, appears more than once.

[MichaelChirico]

Just came across the same, proximally/somewhat unexpectedly:

microbenchmark(times = 10000L,
  2**20,
  2^20,
  bitwShiftL(1, 20)
)

Of course 2**20 parses as 2^20 (see substitute(2**20)), but still had a head tilt when I saw this error.