What I expected to happen:
When I powered a repeater cycle, and replaced a redstone dust, the power states of either side would swap repeatedly.
What actually happened:
When I powered a repeater cycle, and replaced a redstone dust, the power states swapped once, but then both sides remained powered.
Steps to reproduce (shown in the attached video):
Make a repeater cycle using repeater configured to the same tick count.
Break one of the redstone dust and replace it once the repeater it powered unpowers.
Notice that the repeater clock will very quickly stop working.
Related issues
is duplicated by
Attachments
Comments

Can confirm.

comparator clocks don't work either

I can confirm that there is some major bugginess going on with redstone in general.
[media]
In case you haven't checked out the "duplicate" bug report I made, here is a world showing some new unwanted behaviour of repeaters introduced in this pre-release (Bug World.zip)

MC-151133 is an infoless duplicate
Have had issues in 1.14.1 of unlocked unpowered repeaters being stuck in their on state, ugh.
Now I don't know if this is a seperate 1.14.1 bug or the same redstone issue, but even pistons won't retract if you power a block over the piston, then remove that powered block or the redstone dust powering that block the piston stays extended until you restore the block's power and unpower it normally (forces nearby blocks to check their power state)
Supplemental - Yes this is a CRITICAL redstone problem, not just dust, but redstone torches, repeaters, comparators, pistons, pressure plates!!! etc.
I just dug a 2x1 hole 2 blocks deep in 1.14.1, I then placed in 2 sticky pistons, placed 2 heavy pressure plates next to the hole on the surface, stand on one and place a block on the sticky piston 1 block below you, it'll pop up, now either place a block on the other piston, or remove the raised block, moving from pressure plate to pressure plate will pop up the sticky piston nearest it, step off both plates and back the pistons stay retracted like they should, wtfbbq???

For an example of the "duplicate" bug that was previously mentioned, it is really easy to replicate and seems to be based on a repeater receiving an input signal of exactly the same length that it is set to... so a repeater set to 2 ticks receiving a 2 tick pulse will lock on until updated, a 3 tick receiving a 3 tick pulse... this can be seen pretty easily as shown:
[media]
Just input a signal which is short enough (for instance, destroying a redstone block immediately after placing it), and then the second repeater in a chain will get stuck...
[media][media]
[media]

[media]
the "stuck" repeaters act like NOT gates.

redstone doesn't update observers properly in 1.14.1pre2
@unknown, then please create a new ticket about that.

affect pre2