PixelCheckSum

[trids]

I'm trying to determine where in a window's client area a certain image appears. It would be in one of three places, and I'm trying to use PixelCheckSum on these three areas to see which one contains the target.

But it looks like the checksum is not always the same for a given image, based on my tests so far . I seem to get 2 distinct values for the same target based on a trial of about 20 tests: one value appears about 80% of the time, but the other appears too often to dismiss as chance.

Question-1: Can I fairly expect PixelCheckSum to return the same value for a given image no matter where it is in a client area, or does its algorithm not allow for this?

Question-2: Is there a better way than PixelCheckSum to find which one of three areas contains a target image? Or to rephrase the goal a little differently, is there a better way to find a target image within a given area?

Question-3: It is a huge pain (and may even be the source of my problem) to determine the checksum of a target image at design time. Is there any chance of having a function (not a UDF, please, a built-in one intrinsic to AU3!) that can determine the checksum from a file? This would enable me to save a section of a screen dump to file (GIF / PNG / BMP / whatever), whose PiXelCheckSum I can then determine and use in checking the 3 regions.

Larry? Anyone?

TIA

[Valik]

Can you create a simple reproduction script that shows PixelCheckSum() returning different results for the same pixels? I'm pretty sure that's not supposed to happen.

[Jon]

If it's for a game then there are all sorts of alpha blending things that can screw it up.

[trids]

I'm afraid it's not even as noble as a game :">

For purely academic purposes, I'm trying to find the "Use Evaluation Version" button on the popup that displays when you start Winzip. You can't rely on the button name or ID, cos they change from time to time, and even the button's position changes randomly.

Happily though, I have now found an easier solution by approaching from a slightly different angle. The script below seems to work all the time, so the effect of different checksums for the same image (which gave rise to my first post) must have been due to slight discrepancies in my aim when I was trying to determine the checksum manually. The second script overcomes manual shakes by ALWAYS looking in the same places for the three buttons, and reports on the PixelCheckSum for them. To be used when determining the hardcoded value for the first script.

Thanks for the kind attention, guys

;Waits for the popup to appear, then click the disabled button.
;If there is no disabled button, find the right one with PixelCheckSum

    Opt("PixelCoordMode",2)
    Opt("MouseCoordMode",2)

    $sTitle = "WinZip"
    WinWaitActive( $sTitle )

    $nCtl = 0     
    If ControlCommand( $sTitle, "", 2901, "IsEnabled", "" ) = 0 Then $nCtl = 2901
    If ControlCommand( $sTitle, "", 2902, "IsEnabled", "" ) = 0 Then $nCtl = 2902
        
    if $nCtl <> 0 then
        ;target has been identified as the disabled button
        ControlEnable( $sTitle, "", $nCtl )
        ControlClick ( $sTitle, "", $nCtl )
    else
        ;hmmm, they're all active, so we have to resort to pixelchecksums
        if PixelChecksum( 12, 255,  12 + 134, 255 + 23) = 2582112889 Then
            ;Click the Left button
            MouseClick( "left", 14, 257, 1, 0 )
        else
            ;Click the Middle one
            MouseClick( "left", 157, 257, 1, 0 )
        endif
    Endif
oÝ÷ ÙK¶¬r¸©µ·¥£huë^®h§zØ^r˦¶¬z)íê'y¦è½æ®¶­scµ&W÷'G2öâFR6V6·7VÒf÷"V6'WGFöââvå¦7Æ6væF÷p £°VÄ6V6·7VÒÄTeEô%UEDôâÒ#S##²fÇC²ÒÒW6RWfÇVFöâfW'6ö⣰VÄ6V6·7VÒÔDDÄUô%UEDôâÒS3sC2²fÇC²ÒÒ'Wæ÷p£°VÄ6V6·7VÒ$tEô%UEDôâÒcC3Cc²fÇC²ÒÒVçFW"&Vv7G&Föâ6öFP£°ÒÒУ°VÄ6V6·7VÒÄTeEô%UEDôâÒS3sC2²fÇC²ÒÒ'Wæ÷p£°VÄ6V6·7VÒÔDDÄUô%UEDôâÒ#S##²fÇC²ÒÒW6RWfÇVFöâfW'6öࣰVÄ6V6·7VÒ$tEô%UEDôâÒcC3Cc²fÇC²ÒÒVçFW"&Vv7G&Föâ6öFP  ¢÷BgV÷CµVÄ6ö÷&DÖöFRgV÷C²Ã"¢÷BgV÷C´Ö÷W6T6ö÷&DÖöFRgV÷C²Ã" £·6WBFR¶W&÷G0¢÷D¶W6WBgV÷C·gV÷C²ÂgV÷Cµô6GW&RgV÷C²¢£·7F'BFRF7Æ¢ôFV'VrgV÷C²6G2gV÷C²¶÷VâÆörvæF÷p ¢b33c¶×5FFÆRÒgV÷Cµvå¦gV÷C° £¶ær&÷Væ@¢vÆR¢6ÆVW¢vVæ@ ¤gVæ2ô6GW&R£·FRVæBöbFR&Vâç6÷rFR6ö÷&G2äBFR6V6·7VТvä7FfFRb33c¶×5FFÆR¢ôFV'VrgV÷CµVÄ6V6·7VÒÄTeEô%UEDôâÒgV÷C²fײVÄ6V6·7VÒ"Â#SRÂ"²3BÂ#SR²#2¢ôFV'VrgV÷CµVÄ6V6·7VÒÔDDÄUô%UEDôâÒgV÷C²fײVÄ6V6·7VÒSRÂ#SRÂSR²3BÂ#SR²#2¢ôFV'VrgV÷CµVÄ6V6·7VÒ$tEô%UEDôâÒgV÷C²fײVÄ6V6·7VÒ#rÂ#SRÂ#r²3BÂ#SR²#2¢ôFV'VrgV÷C²ÒÒÒgV÷C²¤VæDgVæ0

The _Debug UDF is by courtesy of Cyberslug