Another alternative: use `RLE` (run-length-estimation), borrowed originally from a very simple data compressor of the same name. 

Here's the code, albeit in the `R` language.  (For non-users,  `1L` just forces integer-class, `x[-k]` means all of vector `x` *except* index 'k'  )  

 Here's the base package code for the function `rle(x)` :   
First line: generate logical vector of "is x[j] == x[j-1] ? " 
  

      y <- x[-1L] != x[-n] ;

 
`which` returns index values when argument is TRUE, and `c` concatenates vectors (is.na just catches N/A values in case the input was skeevy) 
   

     i <- c(which(y | is.na(y)), n) ; 
finally, create a structure. First element calculates run lengths by comparing sequential index values in `i` ; second element returns the value of the input vector every time that run terminates 
  

      structure(list(lengths = diff(c(0L, i)), values = x[i]), class = "rle")

 

 then add a line to find `max(lengths[values==0])`.