MouseState as button

[RobotCaleb]

At https://github.com/wjakob/nanogui/blob/master/src/screen.cpp#L443 mMouseState is passed to mouseDragEvent(), which has a prototype of bool mouseDragEvent(const Vector2i &p, const Vector2i &rel, int button, int modifiers). Note that parameter is called button, not state.

mMmouseState isn't a button, which causes any code that tries to inspect it (say, in the case where the constraint on a drag action requiring mouse button 1 (https://github.com/wjakob/nanogui/blob/master/src/screen.cpp#L488) has been removed) to get the wrong value due to the actual button not being passed.

[wjakob]

The user might press multiple buttons, so passing the state generally seems like the right way to go. What about renaming the parameter from "button" to "buttonState" or something like that?

[RobotCaleb]

Mouse drag event is currently constrained to kick in on just one button. In that case, pushing the state at all seems a bit of a fallacy. That is, checking the state doesn't provide any information with meaning.

Unfortunately, I'm not offering a solution as I don't know the best way to handle it.

On Wed, Jun 15, 2016, 03:11 Wenzel Jakob notifications@github.com wrote:

The user might press multiple buttons, so passing the state generally seems like the right way to go. What about renaming the parameter from "button" to "buttonState" or something like that?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/wjakob/nanogui/issues/70#issuecomment-226131701, or mute the thread https://github.com/notifications/unsubscribe/AAG6-V5o3tq6nJ4zw44qFy5jP067tydpks5qL8HBgaJpZM4I12fU .

[wjakob]

That limitation (to drags with the left button) is likely to go away, see the commit in #69.

[RobotCaleb]

In that case, should button be the drag action originating button?

[wjakob]

That would be a reasonable change. Feel free to submit this as a PR.