Skip to content

Conversation

npwoods
Copy link
Contributor

@npwoods npwoods commented Dec 16, 2020

This is used by BletchMAME to toggle mouse capture on and off

This is used by BletchMAME to toggle mouse capture on and off
@cuavas
Copy link
Member

cuavas commented Dec 18, 2020

That should only be set by the OSD module providing the inputs. I’ve got a better solution for this coming, hopefully in time for 0.227 but if not then for 0.228.

@cuavas cuavas closed this Dec 18, 2020
@npwoods
Copy link
Contributor Author

npwoods commented Jan 18, 2025

I'd like to reopen this PR. I tried applying this change and it does the trick. @cuavas looking at the comments from a few years ago, it sounds like a better solution is coming. Is there any word on this, or anything I can do to help?

@npwoods npwoods reopened this Jan 18, 2025
@npwoods
Copy link
Contributor Author

npwoods commented Jan 25, 2025

@cuavas any thoughts on this blast from the past? Thanks!

@rb6502
Copy link
Contributor

rb6502 commented Feb 4, 2025

@cuavas If you're too busy for the correct OSD path on this, maybe we can put this in and mark it deprecated immediately?

@cuavas
Copy link
Member

cuavas commented Feb 4, 2025

I’m trying for fuck’s sake. I completely re-wrote the handling of pointer input so we can support multi-touch input properly. I even exposed it properly for scripts. I hope you can see how important this was for stuff like synths with on-screen faders that lend themselves to a multi-touch UI. I rewrote the menu event handling to get rid of the baked in one-frame lag on input.

I’m now neck deep in DRC shit because the more I look at it, the more broken stuff I find. My reward for touching anything is having to spend months working myself into the ground trying to fix the inevitable pile of issues I find.

Fixing the interface between the UI and the OSD window objects is difficult, but I’m trying to get there. If hacks like this are added, it takes us in the opposite direction. It’s incredibly de-motivating, and it makes me feel like we’ll never actually make progress. I guarantee you things will start to depend on it, and I’ll be the bad guy for having to remove it again.

It’s like exposing show_menu – it’s only safe to call that function under very specific circumstances. For example calling it with the ImgUI debugger active or the graphics viewer visible will cause stuff to get into an inconsistent state. There’s nothing to guard against scripts doing that.

Do you want me to just give up on trying to make things better?

@npwoods
Copy link
Contributor Author

npwoods commented Feb 8, 2025

@cuavas I'm curious what can be done (whether it me @rb6502 or myself) to help you in this process. Four years (!!!) ago you said "I’ve got a better solution for this coming, hopefully in time for 0.227 but if not then for 0.228", and this didn't come to pass. This is of course understandable, as you're incredibly busy.

I don't really have a sense as to what this promised "better solution" is, but if there is anything that can be done in the meantime to get us there, or in absence of that better solution, a limited future proofed solution (perhaps enabling this change in very specific circumstances), that would be great.

Thanks!

@npwoods npwoods merged commit 0e63bce into mamedev:master Jun 17, 2025
5 checks passed
@npwoods npwoods deleted the input_class_enabled_write_accessor branch June 28, 2025 22:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants