@ngraham is there an admin that can review this or should I abandon and make a MR on gitlab?

https://invent.kde.org/plasma/plasma-pa/-/merge_requests/2

In D29827#673551, @ngraham wrote:

The name needs to be your full name ("Firstname Lastname"). This was a pre-existing requirement, it's just now enforced before patch submission, rather than afterwards.

In D29827#673541, @bcooksley wrote:

You are being rejected because those commits were made using a name that appears to just be a first name - you will need to amend your user.name in your Git configuration and amend the commits accordingly

In D29827#673539, @ngraham wrote:

You need to push to your fork of the plasma-pa repo, not the main location. git push fork should do it if you've set up your fork with the name fork as described in https://community.kde.org/Infrastructure/GitLab#Add_the_fork_to_your_source_checkout

@ngraham I'm following instructions, but push is rejected:

In D29827#673534, @ngraham wrote:

What email? Huh?

@ngraham I'm sending the email, but the system reject the request.

In D29663#673173, @marten wrote:

@kurmikon: do you have commit access?

Diff updated to a better version. So, I'm explaining what happens here.

In D29365#671289, @ngraham wrote:

Typically would would bind both to the same backend value and everything would just magically work. However the code in this KCM is terrible, and is quite imperative, without using bindings properly. So you'll have to manually set both anywhere the slider's value is currently set, and read the value of both anywhere the slider's value is read.

In D29365#671270, @ngraham wrote:

Actually sorry, I won't get to that refactor soon so I don't want to block your work.

Is this reviewable in its current state?

Corrected indentation (I was seeing it good, but atom fills 4 spaces by default while original code used tabs).

I will check it out when have time.

In D29431#669402, @ngraham wrote:

Got confirmation from @davidedmundson that this is okay. Can you provide an email address so I can land your patch with correct git authorship information?

@ngraham No reply after one week. I think this can be merged.

In D29365#666640, @ngraham wrote:

Indeed, the code here is a catastrophe. I'm actually in the middle of a rewrite to make everything more declarative which should make UI changes much easier. If you'd like to wait a few days until that's done, it might make your patch a lot simpler.

So I worked on this.

In D29365#664091, @ngraham wrote:

I like @ndavis's idea of using a slider and a spinbox. The slider can be used for coarse adjustment, while the spinbox can be used for fine-tuning and have a very fine step size. I think that would work.

[-1, 1] range of values?

Removed stepsize property, default value 1.

In D29365#663827, @ndavis wrote:

I had a look at the documentation for libinput pointer acceleration (https://wayland.freedesktop.org/libinput/doc/latest/pointer-acceleration.html). For anyone following this conversation:

• 0 is the default
• when greater than 0, acceleration starts sooner and pointer movement increases by larger factors
• when less than 0, acceleration starts later and pointer movement increases by smaller factors until pointer movement does not increase with mouse movement speed. The point where that happens is somewhere in the range of greater than -0.75 and less than -0.5.

In D29365#663671, @ndavis wrote:

In D29365#663395, @ngraham wrote:

Of course the problem with a spinbox is that by displaying numbers, it demands some kind of legend or label. What does a value of 100 mean? Does a value of 0 make any sense? etc. The value displayed to the user must make some sense or explained with context. In this case, perhaps a percentage would work, with the default value being "100%", and the max and min values displaying 50% and 200%, or something, or whatever the scale actually is.

Didn't we solve this already with Slider + Spinbox in the Display Configuration KCM?

In D29431#663581, @ngraham wrote:

Maybe we went too far in the direction of efficiency at the expense of familiarity, because the 24px step size suggests inaccurately that intermediate values can't be entered?

In D29365#663597, @ngraham wrote:

Hmm, 0% as a lower value would imply no movement speed, which is obviously not possible here.

Another solution. Spinbox 0% to 100%. -1 is 0% and 1 is 100%. Libinput default 0 will set 50%.

In D29431#663558, @davidedmundson wrote:

That doesn't explain why we're changing the step size.

No-one wants to click the up arrow 100 times.

No changes needed.

In D29422#663389, @ngraham wrote:

How does subpixel hinting work without anti-aliasing?

In D29365#663395, @ngraham wrote:

Of course the problem with a spinbox is that by displaying numbers, it demands some kind of legend or label. What does a value of 100 mean? Does a value of 0 make any sense? etc. The value displayed to the user must make some sense or explained with context.

Pulseaudio is an abstraction layer over alsa.

In D29312#662773, @mart wrote:

i think i would prefer a somewhat inaccurate message, rather than an obscuretecnicism which is technically accurate (

The spinbox would be a better solution.

@nicolasfella if an application is recording directly from an input device (a simple microphone), your hint would be the solution: we should check and exclude sink source monitors.

In D29312#660704, @nicolasfella wrote:

In D29312#660702, @kurmikon wrote:

In D29312#660670, @nicolasfella wrote:

but due to a lack in qt libraries

Can you elaborate on this? What is Qt lacking?

I'm not an expert, so I don't really know. Reading the bug report, there's no way to discern input devices from monitor sinks. So if you want to report applications that are using a microphone, you end up showing applications like PuseEffects that create a monitor sink. Those applications can use a microphone but in most cases don't, because PulseEffects is mostly used to apply effects to output streams (but need to record those streams effectively).

Qt is not really involved in this. What matters is what libpulse offers, and it seems to me that we can check whether a sink is a monitor (https://freedesktop.org/software/pulseaudio/doxygen/structpa__source__info.html#a5e304b796ce71c7fa54e5a88f3333630). One would need to add a method isMonitor to Sink that reads this information and then we can not show the indicator when it's a monitor

In D29312#660670, @nicolasfella wrote:

but due to a lack in qt libraries

Can you elaborate on this? What is Qt lacking?