Pushing End Rods only occasionally breaks Cactus

My testing revealed the same results in 1.12.

As of 17w47b (or possibly 17w47a), this behaviour has changed. 

Edit: In future, I will refer to this as "Behavior 2", (Behaviour 1 being the one I originally described)

Now (so far as I can tell) pushing or pulling and end rod into or out of adjacency with a cactus will always break the cactus

However you can still place end rods next to cactus without breaking the cactus, (and you can still place cactus next to end rods)

End rod orientation continues to have no effect.

Should I change the title to "Moving End Rods with Pistons Breaks Cactus, but Placing Them Does Not" or should I wait until things stabilize a bit? (Or alternatively, should it be a separate report with this marked fixed?)

This new behaviour persists in 17w48a and 17w49a

As of 17w50a this behaviour has changed again -- Edit: "Behaviour 3"

Now (so far as I can tell) pushing or pulling and end rod into or out of adjacency with a cactus will NEVER break the cactus

End rod orientation continues to have no effect.

If this behaviour remains unchanged, I would consider this bug to have been fixed. (though I would still expect end rods pointing into or away from cactus to break it)

As of 1.13, behaviour 3 is still present, but in 18w30a, behaviour 2 is back. (Rods always break cactus when pushed, but not when placed, see my comments above for more details)

Please advise whether this warrants a new report or changes to the description and/or title.

This was just marked as "Cannot Reproduce". Is this based solely on the original behaviour? Because as noted in the comments, it has changed. No one ever got back to me about whether that warranted a new ticket, but I changed the name at some point to reflect.

I literally tested the bug a couple hours ago in 19w34a, and "Behaviour 2" is still present.

(Pushing or pulling and end rod into or out of adjacency with a cactus will always break the cactus, but you can still place end rods next to cactus without breaking the cactus, (and you can still place cactus next to end rods))

I am splitting the newer (ongoing) behaviour (behaviour 2) to a new ticket ( MC-160095 ) to hopefully reduce confusion, and have reverted the name of this ticket to reflect the old behaviour listed in its description.

Yes, the devs usually only look at the description of a ticket when trying to reproduce it. That's why we usually want a new ticket created, or the description to be edited, if the behaviour changes.

I've linked this ticket to MC-160095. If there's any behaviour from this ticket that still seems like a bug, please create a new ticket.

Is there any way (or reason) to add move the versions affected by "Behaviour 2" to the new bug? (17w47b through 17w49a and 18w30a and newer)

Also, should this be marked fixed in either version 17w47b or 17w50a instead of Cannot Reproduce?

Yes, we can do both, but it is not really necessary, as we only care for the latest version and the latest snapshot. We can add archived versions to the affected or fix version list, but that's rather complicated. It's possible, but it's not really worth the hassle.