Distribution of residuals envelops normal distribution when plotting the distribution of the obtained residuals (from a regression) in stata to see if they fit a normal distribution, I get following result:

This seemed very odd to me at first but when I investigated it further, I found following reasons for this behavior:
1) The dependent variable can only take on the values 1, 2, 3, 4, 5, 6 and 7
2) all fitted valaues lie between 3.6 and 4.2
So naturally, some values of residuals can't ever exist, namely all values with .5-.7. My question is, what kind of implications does this have when assessing normality? Other than the up and downs of the distribution, it does seem to follow a normal distribution. But can I assume that residuals are normally distributed based on this?
 A: 
All normally distributed variables take values on the range (−∞,∞) and
  are continuous (though there's more to them than that of course). If
  your distribution is bounded in the values it can take, it is plainly
  not normal. If the values taken by your response are small positive
  integers, they are plainly not normal. If the original response
  variable is discrete, then the conditional distributions are all
  discrete; it is the conditional distributions that are assumed normal
  if you make a normality assumption in regression. 
Since it's clear before we even get data that it cannot possibly be
  actually normal, we must ask ourselves what is the purpose of the
  assessment? You presumably want to answer a different question than
  "is this variable normal" -- the answer to that question would almost
  never be yes. A more useful question might be "is the situation I am
  in one where the non-normality I have is not going to affect my
  tests/confidence intervals etc more than I am prepared to accept?" ...
  and a hypothesis test won't address that. (This is addressed many
  times on site.)

-Glen_b
