Password value binding triggers updateTarget?

[jods4]

I am observing a strange beahvior that is causing me trouble.

I have a simple value binding + behavior on a password field:
<input type=password value.bind="p & validate" />.

When I type text in the input, I see updateSource is called, as expect.
When I mash keys -- or simply press two keys at the same time -- it seems updateTarget is called as well. But why? This is causing unwanted side-effects in my binding behavior.

This is easy to reproduce in IE11 with a password box: just hit two keys at the same time.
It seems it sometimes happens with a normal textbox as well, but it is a lot harder to reproduce.

[jdanyow]

does it happen in other browsers too? Whats the call stack? Can you provide a plunker?

[jods4]

It does easily happen in Edge as well. Does not seem to happen on Firefox. Untested in Chrome.

Here's a plunker: https://plnkr.co/9cF0XO8FjPK6relatTf5

There's one updateTarget at startup, which is probably because I did not define x on the model, so it gets updated from undefined to "" (I guess).

But then if you hit two keys at the same time, you very easily get a updateTarget on IE11 and Edge.

As for the callstack, I would like to take the opportunity to say that the TaskQueue makes it often quite hard to track causality. All the bindings are updated in this micro-queue... I wonder if there is a way to opt-in long stack traces like built-in async callbacks do in the Edge debugger callstack.

[jdanyow]

I see it happening in edge too. not in chrome, firefox or IE11.

https://plnkr.co/edit/ahRwzTeRjFeCYKUs0U1c?p=preview

[jods4]

@jdanyow Good that you can reproduce it!
I experienced it first on IE11, not sure why you don't.

Any idea where it comes from?

[jdanyow]

No, not yet. I also looked at adding a "long stack traces" plugin for the TaskQueue to help with the debugging thing you mentioned.

[huw489]

Did you ever find a fix for this? I'm experiencing the same problem; I've narrowed it down to Binding.prototype.call in Aurelia-binding.js.
what I'm seeing is that in Chrome when pressing two keys at once in a textarea input that when getting the value from the textarea it only includes the first key typed, but in IE11 it shows both keys - hence it executes the code if (newValue !== oldValue).
I hope that makes sense.
I came across this from users telling me the cursor always jumped to the end of their text if they moved back into the text and typed quickly.

[EisenbergEffect]

@jdanyow Putting this back on your radar.

[Alexander-Taran]

Stale since 2016

[EisenbergEffect]

@jdanyow Can you comment?

[Alexander-Taran]

or @jods4 ? (-:

[jods4]

@Alexander-Taran
Unfortunately the plunker is gone but I think the description of the problem is clear?