238264 - Mozilla doesn't use GetKeyState according to spec

Status: RESOLVED FIXED

(Core Graveyard :: GFX: Win32, defect)

Description

[[e-mail]]

The MSDN docs state that on Windows the sign bit 15 of the 16-bit short return
value from GetKeyState determines the pressed state of the key. However Mozilla
uses bit 7 for some reason.

Comment 1

[[e-mail]]

Attachment: Proposed patch

Comment 2

[[e-mail]]

http://lxr.mozilla.org/seamonkey/source/embedding/tests/mfcembed/BrowserView.cpp#458
doesn't look right either but I'll let Adam figure out what he wants there...

Comment 3

[Ere Maijala (slow)]

Comment on attachment 144506 [details] [diff] [review]
Proposed patch

I've seen a few examples where 0x80 is used (such as
http://mech.math.msu.su/~vfnik/WinApi/g/getkeystate.html). I'm not sure why,
but it's definitely wrong according to any recent documentation. Anyway, better
to do it right. r=ere

Comment 4

[rbs]

Comment on attachment 144506 [details] [diff] [review]
Proposed patch

sr=rbs

Apparently the confusion is because people think that 'short' is 8 bit, whereas
it is 16 bit instead, and they meant to test 0x8000.

I checked another MSDN document which stated clearly that:
"The GetKeyState() function returns a number indicating the current state of a
specific key at the time the DLL call is made. A negative number indicates that
the key is pressed, a positive number indicates that the key is not pressed."

Comment 5

[[e-mail]]

Comment on attachment 144506 [details] [diff] [review]
Proposed patch

Seeking approval for this low-risk Windows-specific correctness patch.

Comment 6

[[e-mail]]

Hmm... it looks like that quote comes from KB article 108357.

Comment 7

[Asa Dotzler [:asa]]

Comment on attachment 144506 [details] [diff] [review]
Proposed patch

a=asa (on behalf of drivers) for checkin to 1.7

Comment 8

[[e-mail]]

Fix checked in.

Comment 9

[:deanis74]

Cool.  Every time I looked at that code I thought that it couldn't be right.