[admin]

I was preparing for this answer one day 😀

Well, I must say this idea is great. If I know it at beginning of development, then is implemented since then. It can solve various situations what I solve now with specialized widgets. Of course this idea can’t handle all situations and after all, these specialized would be implemented too.

So, question is – is this idea still worth of implementation? Is there at least one situation what can be handled significantly better with this idea then with current implementation?

Please don’t take this question wrong, I will be very happy if as many people as possible can write down here situations what can be fixed with implementation of this idea and which can’t be handled by current offer.

Why I still didn’t make decision?

Exactly like with layers implementation, I need rewrite a lot of core code. This is not fun. With layers I was 100% sure that this brings significant improvements. It took me 8 versions until the code was prepared for layers. Now it is not so dramatic, but there are various complications. Android sends really many touch events which are evaluated against collisions. With multiple layers it is yet more work, but still not bad, because I finish detection when I found first widget. But this idea needs scan all widgets over all layers below finger. This needs implement binary trees or other help for faster detection. And of course all flags like “activate on tap only”, “deactivate on leave” etc must be functional.

[admin]

If you want, send me an email to magicbox@imejl.sk

[admin]

Hi, I found these threads on web, hope it helps

[admin]

Hi, yes, you are right, I do the same, I move inner circle almost to bottom, this way I get larger front area and smaller back area. This is good for me. In addition I found more comfortable for me to not enable diagonals for 4way key mop. This is good for doom and hexen. But for duke3d 4th mop type – native with full range. Of course physical gamepad is better for fps games, touch screen is not so “handy” for fps, but I played and finished duke3d and doom well on my phone. And now I often use one button for strafe, I found it very useful.

[admin]

I didnt try to run it in kodi, i only use it for movies, but your command doesnt look good for me

android.intent.action.intent_msg1

Is incorrect i think. Magic dosbox accepts intent with name intent_msg1 and with value game id.

But you should ask on kodi forum how can kodi construct intent and send it to application.

[admin]

Hmm maybe you can try download another installation files. And you can try both cores (svn and normal), maybe there is difference

[admin]

I don’t know…i sent you my mgc file

[admin]

look at my command and yours, you added enters between images, there should be spaces

[admin]

Unfortunately I didnt add mounting floppies to ui, you must do it in expert settings, command looks like this :

“WINST” is folder with floppy images

[autoexec]
mount c: “/storage/emulated/0/Download/WINST”
imgmount a: “/storage/emulated/0/Download/WINST/1.img” “/storage/emulated/0/Download/WINST/2.img” “/storage/emulated/0/Download/WINST/3.img” “/storage/emulated/0/Download/WINST/4.img” “/storage/emulated/0/Download/WINST/5.img” “/storage/emulated/0/Download/WINST/6.img” -t floppy

change to a: and start installation

then when you are asked during installation for another floppy, press “change cd” from general settings

[admin]

Yeah, one-two times I received report about lauchers what have problems with custom shortcuts. Reason was replace in-built launcher with nova launcher (or similar). I say truth, in the past I wanted create this type of widget like a fix for shortcuts problem, but it was so rare that I worked on something else. I can’t say now how much work it is, because I never did any desktop widget, but I can look at this.

[admin]

Maybe some versions of win3.11 have problems. I use win3.11 fr workgroups and works well. I installed it from floppy images

[admin]

Cool, will try it, I played with brother similar game about horse races 🙂

[admin]

Hmm, if I understand well, you want place image (not widget) on the screen and then assign it deadly strike widget ?

Flip will be nice feature, but it is very hard and very easy for programming. It is very easy if I use continual drawing, but very hard if I use lazy drawing.

Continual drawing means draw 30-60 fps on screen every second.
Lazy drawing means draw only if dosbox sends data to draw (or if I force drawing, when needed (opening/cosing bag)

Continual drawing is bad, because drawing 30-60 frames per seconds drains battery a lot and leads to phone heating. Increased heating means less cpu power – dosbox performance is degraded. So I avoid continual drawing as much as possible.

By lazy drawing you don’t know when next frame will arrive, so you can’t return flipped button back to normal state, because you wait for dosbox drawing event. Dosbox can send 1,2 frames per second, then 0, then 35, etc. Like I said, there are situations when I force drawing again (even dosbox sent nothing), but I choose wisely when, it can’t be often. But button flips very often.

I can add flipping to lazy drawing, but you must accept graphical glitches. I mean, I will higlight and de-higlight when dosbox sends data to draw. In praxis, button may ignore highlight or will stay higlighted to next redraw. Not very nice.

In addition, you can enable “synaptic feedback” on button if your device has vibrator

[admin]

I think this is not a problem, but has it real usecase? Can you give example when you would like to use it?

[admin]

Hah, monopoly 🙂 cool thank you:)

[Author]

Posts