According to dinnerbone (https://twitter.com/Dinnerbone/status/313237462757543937), this is not (as you have specifically described it here) a bug as it is a consequence of intended behaviour, although I do admit that it does seem counter-intuitive. Please allow me to explain:
It's first worth noting that this applies to all pistons, not just sticky pistons. This is also not a duplicate of MC-108 in the precise sense, but rather an inverse of it, for the leftmost pistons are actually showing the intended states of the pistons in this set up - placing/updating adjacent blocks around the piston changes nothing. However, if you place a piston in any orientation on the ground and then immediately place a redstone block two blocks above it, then the pistons require an update of the adjacent blocks to extend. For all practical purposes however, I feel it is understandable label it as a duplicate.
The reason as to why (according to dinnerbone) this specific issue is not a bug as you have described it is as follows: it has always been the case that a piston would accept a signal from two blocks above it (regardless of orientation). Since the redstone block in your first two pistons is two blocks above the piston and since redstone blocks send signals downwards (placing redstone dust immediately below a redstone block will power the dust), the block is powering the piston and thus it remains extended. A rather unusual consequence of this behaviour is that if you have a piston in any other orientation (e.g. to the side) and you have a redstone block two blocks above it, then the piston should extend even though there is a block of air permanently separating the two. Personally, I feel this is a bit silly and shouldn't be considered normal behaviour, but then again, I'm not a developer.
As for the rightmost piston in the picture, I can only offer conjecture as to why it does not extend and why this too is most likely intended behaviour. When dealing with pistons in a horizontal layout - i.e. solely within 1 block of height - powering a piston with redstone dust from the block immediately in front of the pushing arm will not power the piston. The reason for this is most likely to prevent the piston from destroying it's own power source. Attempting to recreate this scenario in a vertical layout however, does cause the piston to extend (even if you avoid powering it diagonally). Whether this in itself could be considered a bug is a matter for debate. My guess is that with the addition of redstone blocks, this behaviour of not powering the piston immediately in front of the pushing arm was extended to the redstone blocks, including in the vertical orientation, thus explaining the third piston in your picture.
Hmmm. What you say does make sense, but I've just tested this with version 1.4.7 and this specific issue wasn't present in that version whereas MC-108 is. Whatever's causing this is new to 1.5 (and is also present in 1.5.1).
Drew Duffield
According to dinnerbone (https://twitter.com/Dinnerbone/status/313237462757543937), this is not (as you have specifically described it here) a bug as it is a consequence of intended behaviour, although I do admit that it does seem counter-intuitive. Please allow me to explain:
It's first worth noting that this applies to all pistons, not just sticky pistons. This is also not a duplicate of MC-108 in the precise sense, but rather an inverse of it, for the leftmost pistons are actually showing the intended states of the pistons in this set up - placing/updating adjacent blocks around the piston changes nothing. However, if you place a piston in any orientation on the ground and then immediately place a redstone block two blocks above it, then the pistons require an update of the adjacent blocks to extend. For all practical purposes however, I feel it is understandable label it as a duplicate.
The reason as to why (according to dinnerbone) this specific issue is not a bug as you have described it is as follows: it has always been the case that a piston would accept a signal from two blocks above it (regardless of orientation). Since the redstone block in your first two pistons is two blocks above the piston and since redstone blocks send signals downwards (placing redstone dust immediately below a redstone block will power the dust), the block is powering the piston and thus it remains extended. A rather unusual consequence of this behaviour is that if you have a piston in any other orientation (e.g. to the side) and you have a redstone block two blocks above it, then the piston should extend even though there is a block of air permanently separating the two. Personally, I feel this is a bit silly and shouldn't be considered normal behaviour, but then again, I'm not a developer.
As for the rightmost piston in the picture, I can only offer conjecture as to why it does not extend and why this too is most likely intended behaviour. When dealing with pistons in a horizontal layout - i.e. solely within 1 block of height - powering a piston with redstone dust from the block immediately in front of the pushing arm will not power the piston. The reason for this is most likely to prevent the piston from destroying it's own power source. Attempting to recreate this scenario in a vertical layout however, does cause the piston to extend (even if you avoid powering it diagonally). Whether this in itself could be considered a bug is a matter for debate. My guess is that with the addition of redstone blocks, this behaviour of not powering the piston immediately in front of the pushing arm was extended to the redstone blocks, including in the vertical orientation, thus explaining the third piston in your picture.
I hope this helps.