If one creates a wall of dispensers, the first time around the whole wall will fire, but the consecutive times certain portions won't fire.

Here's what I did to test and openly observe the bug:

I set up a wall of dispensers like so, all full of tnt. The repeater covered by the bucket is at the full setting.

When first triggered with a button, the whole wall of dispensers activated and fired tnt.

However, the second time, large portions of two rows did not fire.

Those same areas failed to fire when I reactivated it many times.

I found that placing replacing a "stagnant" dispenser would allow that replaced dispenser and the stagnant dispensers that were to the left and right of it (presumably all that was touching it) to fire normally for one shot before going back into the previous pattern. In other words, by doing this, a total of three dispensers that were stagnant fired once, but only once more, before becoming stagnant again.

Note that the fourth dispenser from the right on the third highest row was the one replaced, and thus the one dispenser to the left and one to the right of it also fired normally as I described.

When I placed a block on a stagnant dispenser (I placed it on the same one that I earlier replaced) and removed it, the dispenser that I placed the block on would fire normally (Did not affect any surrounding dispensers) once, and then yet again returned to being stagnant.

The dispenser that was four from the right and third highest (same spot as the replaced one) had the block placed on it, and thus only that one of the stagnant dispensers (not the ones around it) fired, but only once..

These two phenomena lead me to believe that this is a problem with block updates.

Also, the portion of the third highest row that was stagnant started to fire normally for a while, except for one dispenser which was originally working fine and not stagnant. I now notice that some of the dispensers that were active are now stagnant, and, as just mentioned, vice versa. This started happening when I used a lever to rapidly pulse it, but seeing how it was delayed and extended with repeaters, I don't think the lever has much to do with it. After some time, it reverted back to the original pattern.

Please note that all of the observed stagnant dispensers were on rows that depended on the dispensers directly below them, which were directly powered by repeaters, to be powered.

Also note that I believe that the cardinal direction that the dispensers were facing does not matter. I have observed this phenomena, though not in a test like this, in all cardinal directions. However, I have not observed this bug yet when the dispensers were facing up and down, as I was never in a circumstance when I was using dispensers in such a fashion.

I also do not know if this happens with any other projectiles or droppers.

One last thing to note:
I did not observe this phenomena when using redstone lamps in a similar test to this; they all lit up repeatedly, as intended.