Open Source GUI Library for Assembler

[mark_k]

Originally Posted by AGS

Maybe this happens because I take the Mousepos rather from the Window than from the Message. Because I have Popupwindows that require me to get a mousepos relative to that windows. Other explanation: when you release the mousebutton, there are still messages to process. so it looks like a delay ...

Don't know whether you solved this problem yet, but how do you handle multiple IDCMP_MOUSEMOVE messages? Could doing something like this improve things (in C-like pseudo-code):

eventloop:
  get_message();
  if (message_type == IDCMP_MOUSEMOVE) {
    update_mouse_pos();   // Update your program's mouse X and Y variables, but don't do any other processing
    goto eventloop;
  }
  // Handle other types of message
  ...

The idea of doing that is, you effectively ignore all but the most recent IDCMP_MOUSEMOVE.

Also, are you only interested in mouse events when the pointer is over a gadget? You can use GACT_FOLLOWMOUSE instead of receiving mouse events for the entire window (WA_ReportMouse).

[AGS]

I think I need to process the mousemoves because if I just update the pos, classes will never known that user moved the mouse over a gadget. I do not use intuition or boopsi gadgtes but render and manage my own. However I am ignoring mousmoves that come with the same coordinates. Sometimes I get a lot of those, dunno why.

[AGS]

Originally Posted by daxb

Colorpicker:
Freeze/crash (no hit) after press the button preview to get the color red and then change the RGB values vial keyboard. At least it happens also in the last version.

This means the crashes I described do not happen on your system? Please try 'use'-ing the config and then click 'revert'.

[mark_k]

Okay. Suppose the user has moved the mouse over a gadget, but by the time you process the mousemove messages the pointer is not over the gadget any more. Do you want to process the messages, reading the pointer position at the time of the message (i.e. from each message), or just read the current pointer position from the Window structure? If no button was pressed, could you ignore all the "older" mousemove messages? Or do your gadgets need to know all mouse moves?

Maybe if there is some way to quickly discard unneeded mousemove messages that could reduce any lag effect. Or if in some cases you don't actually need mousemove messages (maybe they are only needed for certain gadget types???), change the window's IDCMP flags so you don't get mousemove messages.

[AGS]

I discard mousemoves that come with the same coordinates as the previous ones. On slow systems this seems to happen. I need the moves always, so that I know if the mouse has entered a gadget or left it. If the mouse is not over a gadget, how will I know if it entered a gadget without reading the mousmoves.

[mark_k]

This might not be possible depending on how versatile your custom gadget system is. But have you considered using a "dummy" Intuition gadget for each of your custom gadgets? The dummy gadget wouldn't have any imagery, but you could easily be notified when the mouse moves over it or is clicked on it.

[daxb]

Originally Posted by AGS

This means the crashes I described do not happen on your system? Please try 'use'-ing the config and then click 'revert'.

Starting GUI twice or more works. Click on revert works too.

[AGS]

Thank you, the Bug occured if some file of the skin images was missing or so. Fixed. Can you provide me an Enforcer output of the colorpicker bug?

[daxb]

No MuForce hit or something else, sorry.

[PeterK]

when I use the original icon.library and open my App on an own screen, the icon colors are wrong. When I use your lib, the icon colors are correct. Can you tell me why?

Hi AGS, you can find my answer in the IconLib thread.

[AGS]

New Version. Features live Screen-Switch (added, yet on the 'application' test page in oxlibinfo). Colorpicker bug might still be there.

http://images.quicktunnels.net/openxui_latest.zip


@mark_k

I cannot really use system gadgets because I support virtual clipping areas, so called 'windows in windows' for which intuition and boopsi gadgets don't care. They are also not compatible to windows with scrollbars if the window is not GZZ (they hang out over the borders).

[AGS]

First App using this GUI-engine released: http://eab.abime.net/showthread.php?t=72695

[JPQ]

is totally unsable using for example c language*?*looks nice and maybe even east way make modern gui.

[Samurai_Crow]

It might work in high level languages with the right wrappers. Have you ever written an fd file for the library?

[JPQ]

Originally Posted by Samurai_Crow

It might work in high level languages with the right wrappers. Have you ever written an fd file for the library?

No. btw i dont need this if i found nice way make gui with C.

[Samurai_Crow]

An FD file could be converted into a pragma file using FD2PRAGMA (found on Aminet and elsewhere) for use with any C compiler and other converters like FD2MODULE (which comes with AmigaE) for use with the AmigaE compiler. All the information that's needed to be used to create an FD file is already included in ox_lib.i in the comments and LVOs.

I guess my next question should be: Would you be interested in getting OX to work on C, AmigaE, and others?

[AGS]

Yet new features include:

* new working window menus
* alpha png images using pngflux.library
* on the fly calculated color gradients
* screen change and iconify
* tooltips for any GUI elements (even for menu entries)
* auto saving/loading of options in applications

Do you have an idea what little tool to code using this engine? I am looking for ideas.

[oRBIT]

This looks really promising! I have to check this out.

[AGS]

Please look into the XoXo-Thread. There you can find a working app that uses the GUI.

[wXR]

Hey AGS, this looks great. What open source license is this released under?