OpenEuphoria

Re: Win32lib questions

• Posted by CChris <christian.cuvier at agric?lture.gouv.f?> Sep 26, 2007
• 680 views

Andy Drummond wrote:
> 
> CChris wrote:
> > That's the version I was thinking of, it had been hinted as "Patch #67" on
> > this
> > list at repeated times.
> > 
> > There is no need nor cause for any kind of friction, only for more testing
> > on
> > various flavours of Windows and various machines. If you have the same
> > problem
> > with that version, then it originates from my mods and I have to solve that.
> > 
> > 
> > This version "#67" was supposed to be a basis for code merging and start of
> > actual testing - not exactly an official release, and the latest oe caught
> > me
> > off guard. Also, the mods are extensive and hardly commented, which may
> > explain
> > why some parts apear to be complex. I'll need to tidy that up too.
> > 
> > Oh btw, you didn't mention what your version of Windows is, or I missed it.
> > 
> > I'll cobble up a stealthless keylogger ASAP for you to include in IDE, so
> > that
> > it records all kbd activity during your session and, when you get the issue
> > and email me the logfile, I'll probably get a clue. That should be saturday
> > at the latest.
> > 
> > One scenario I can think of OTOH is as follows: if Windows traps a special
> > key
> > sequence like Ctrl-Alt-Tab, the WM_[SYS]KEYUP message for the Ctrl key may
> > not
> > be received by the application that got the key down event. This would cause
> > a jammed shift key symptom. Try figuring out if this is your problem. If
> > this
> > is an issue, it's quick to fix - ask default handler for w32HLostFocus to
> > clear
> > the internal kbd buffer. Alt should be handled correctly.
> > 
> > CChris
> 
> Your key-logger sounds an excellent idea. I'm not smart enough with Windows
> (it's not my prime target for my skills) to be able to do one myself.
> In DOS, yes, in Windows XP SP2 (just so you know!), no.
> 
> I'll pick it up hopefully Monday morning here in England and give it a go.
> If you want to email it to me I am on andy at kestreltele.com.
> 
> We'll see what we can achieve. I guess your logger would need to pick up the
> information as to what shift state IDE thinks is true with each keystroke
> as well as the keystroke itself.  Sorry, of course you know that!
> 
> Andy

The logger would simply intercept the relevant WM_ messages and write the data
they carry to a file. The enhanced GetLastMessage() also records timestamp, in
case there is an isssue with this. It's straight win32lib code - I told you it
wasn't any stealthy.
Just one thing: did you check:
1/ whether sticky keys are enabled on your system;
2/ If so, whether turning them off changes the behaviour.

On my system, they are disabled; this might explain why I wasn't able to
reproduce the bug.
The newer lib checks the keyboard events at a lower level than before, so as to
report them in the new w32HKeyboard event. As a result, it may interact with
Windows keyboard handling in ways it hadn't before.

CChris