Window dragging sometimes fails

I recently updated from KLCP 16.2.0 to 16.4.0 and got the issue that dragging MPC-HC's window by clicking inside it (so not by using the titlebar) sometimes does nothing, no matter how far I drag.
It seems very inconsistent but when I find a point where it fails, trying to click-drag close to it will almost always fail too, almost as if there's some invisible rectangle that's being dragged around or something weird like that.
This happens with KLCP 16.4.6 too and for me this is very noticable since I always shuffle the window around like this, though trying to reproduce it on the spot can take quite a few tries.

Doing some research, MPC-HC was updated from 1.9.13.14 to 1.9.15.76 and looking at the changelog I found

Added a small safezone above and below the seekbar where (accidental) mouse clicks/drags are ignored.

That seems a likely candidate, here's a link to the PR.
I can't really tell if there's something obviously wrong there, but it's the only lead I got so far.

Hm... I should probably have created an issue on MPC-HC's Github instead now that I think about it.
Should I still or are the MPC-HC devs here as well?

Aight, went ahead and opened an issue on Github.

I am not able to reproduce it. The safezone is an area of about 20 pixels above the seekbar.

When the window is maximized, then you must drag a distance at least 10% of window size.

Hm... I can reproduce it dragging anywhere like right in the middle of the window and without the window being maximized.
I think I even had it happen that I couldn't click "Play" or "Navigate" once, but I can't say for sure since that happened so fast.

I recorded it in action here:


I first found a spot by trying to drag the window at random positions (which can take 50+ tries) and once found, click-dragging that spot consistently fails.
Once I drag it somewhere else, the deadzone seems to move to another random spot.
The deadzone seems to be more wide than tall, so I still think it's related to the seekbar deadzone.

What's really odd to me however is this: I can confirm that the seekbar deadzone is definitely working (even downgraded to KLCP 16.2.0 to double check), so it seems the issue is not simply that the seekbar deadzone is placed wrong, so maybe the bug is actually caused by something else, maybe somehow related to window messaging?

I also wonder why you can't reproduce it, maybe it's some platform dependent bug? As you can see I'm still on Win7 on this machine, though the WinAPI didn't change since Win2000, so I'm not sure...

It is working here on Windows 7.

I tried with both the classic theme like you and the dark theme in the player. Tried with and without a video loaded.

Very strange problem. Certainly since it doesn't happen consistently.

Here you can find development builds:
https://github.com/clsid2/mpc-hc/releases
Test 1.9.13.21 and 1.9.13.37

So you tested on Win7 too, weird.
Hm... yeah, can reproduce on 1.9.13.37 (took a while) but not on 1.9.13.21 as expected (tried like 200 times).

Can't think of much else, tried changing a few settings like DefaultToolbarSize, the video output, also tried in compact mode but alas...

Maybe I'll try figuring out how to build the project and start debugging, that's usually what I do when something annoys me enough - though looking at the Compilation.md, this seems like a lot of effort to get going.

To build the player itself you only need Visual Studio and yasm. The other stuff is for building LAV Filters. But you can skip that. There is a "lite" build target in the project to build the player with the internal codecs disabled.

Just to let you/everyone know, this issue is now fixed in the latest dev build (1.9.16.23).

Turns out the issue was caused by an 8 year old bug, always fun when that happens!