Fixed off-by-one error in mouse button bit-masking

[Linearity]

getMouseButtons detects the wrong buttons because its bit mask doesn't match SDL's convention. For example, it's impossible to detect the left mouse button being held.

getMouseState, like SDL_GetMouseState, produces a set of Boolean flags encoded as the bits of an integer. We test the nth flag by AND-ing that integer with a bit mask: the integer 1 bit-shifted by n places.

The state of the left mouse button resides in the lowest-order bit—0 places away from the lowest place. So, to test that button we have to bit-shift by 0 places.

SDL provides the macro definitions SDL_BUTTON(X) and SDL_BUTTON_LEFT to do that bit-shifting. It defines SDL_BUTTON_LEFT to be 1, not 0, and it defines SDL_BUTTON(X) to subtract 1 from its argument before using it to construct the bit mask (see lines 203-204 in SDL_mouse.h):

#define SDL_BUTTON(X)       (1 << ((X)-1))
#define SDL_BUTTON_LEFT     1

So, && SDL_BUTTON( SDL_BUTTON_LEFT ) = && (1 << (1 - 1)) = && (1 << 0), which is what we want.

However, the corresponding Haskell provisions don't match that convention. Raw.SDL_BUTTON_LEFT (and hence toNumber ButtonLeft) is defined to be 1, but getMouseButtons doesn't do the same subtraction as SDL_BUTTON(X):

getMouseButtons = liftIO $
  convert <$> Raw.getMouseState nullPtr nullPtr
  where
    convert w b = w `testBit` fromIntegral (toNumber b)

Here, (`testBit` fromIntegral (toNumber ButtonLeft)) = (`testBit` 1), but that doesn't test the left button. We want (`testBit` 0).

[chrisdone]

Pity it's hard to make a test suite for this!

[chrisdone]

Thanks!

[Linearity]

Thanks for pulling this. When will it be on Hackage? There is no workaround for detecting the left mouse button with the current version. I either have to distribute my own version or use a different, much less popular button.

[ocharles]

I'll try and cut a release soon - I'll add a reminder.

[Linearity]

What's the status of this?

[ocharles]

Just to update others, this is released as 2.4.1.0.

[chrisdone]

Cheers Ollie

…

On Thu, 5 Jul 2018 at 11:19, Oliver Charles ***@***.***> wrote: Just to update others, this is released as 2.4.1.0. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#177 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAArC5lnhyrS7Cz4Cz_t9NGtLY2TBPSfks5uDegYgaJpZM4TZTba> .