JacobA

Saw that - although that is not exactly what I want. Also that method is much slower in terms of time per iteration.

I am wanting to compare a database of recorded "perfect" (digital) images, with images that will be taken of the same objects with a camera - and finding the one that matches the photographed image in the DB. Consequently, there will be noise (and maybe not even the same exact size), although the objects being imaged themselves will be the same (e.g. blue in the perfect digitally made images in the database, won't necessarily be the same blue in the photograph due to many forcings such as abient light, glare etc.). The images will be under 400 pixels in their largest dimension. After some research I realized there are lots of Autoit3 solutions for image comparison, but I am needing a suggestion of which to try first given my needs in an attempt to at least save some time. It is already a daunting project. I noticed that the two best solutions seem to be methods with Prospeed and imagemagik. Looking at the imagemagik option, I like how it has a quantification of the differences between images. I can picture coming up with an algorithm that would simply quicksearch through the "perfect" image database, acquiring levels of variance between the images in the DB and the one taken with the camera, and choosing the one that matches best. Unfortunately, I can also picture this taking AGES. I would like to try and keep the iteration time under a second if possible for each photograph. Prospeed apparently does comparisons really quickly (I noticed a quote on the forum of a 600x800 image comparison under 2 ms). And it accepts a parameter of tolerance which is something I would clearly need - however, I am unsure however of what "tolerance" actually means in this case and have been unable to find anything in searching. Does this mean tolerance from one pixel to another, or tolerance of a color-range within a specific pixel, or both. Obviously I would want the later as every color, although within a specific color range, won't be exact given one is from a digital image (scan) and another is from an external camera. I also cannot find a working download of the most recent .dll for prospeed as all of the links were switched to a different domain. Any conversation/suggestions would be greatly appreciated!

Here are some snippets of the code that are doing what I described above. Firstly, I create a GUI interface with an IE object GLobal $oIE = _IECreateEmbedded () Global $main_GUI = GUICreate("Program Name", 1024, 768+$height_changer, _ (@DesktopWidth - 1024) / 2, (@DesktopHeight - 768) / 2, _ $WS_OVERLAPPEDWINDOW + $WS_VISIBLE + $WS_CLIPSIBLINGS + $WS_CLIPCHILDREN) I build the GUI with buttons etc. One button I create as Global to change GUI modes: Global $GUI_Mode_2 = GUICtrlCreateButton("Second Mode", 690+20, 15+75, 100,20) In the While loop - I check for the case of $msg for when the user clicks the button (this is all simple GUI stuff...). It calls a function that destroys the main gui and creates a new one Case $msg = $GUI_Mode_2 MsgBox(0,"Second Mode", "Program will now go into it's second mode.") $end = 2 _second_gui() The function starts off as such: Func _second_gui() ;disable the old GUI ;GUISetState(@SW_DISABLE) ;GUISetState(@SW_HIDE) GUIDelete() Global $oIE_2 = _IECreateEmbedded() $new_GUI = GUICreate("Secondary Mode - Program Name", 1024, 768+$height_changer, _ (@DesktopWidth - 1024) / 2, (@DesktopHeight - 768) / 2, _ $WS_OVERLAPPEDWINDOW + $WS_VISIBLE + $WS_CLIPSIBLINGS + $WS_CLIPCHILDREN) I go on to embed the new IE object into the new GUI and build the new secondary GUI. When the user switches GUI's when running the program, the virus scanner triggers.

I understand Xprotec (Val it's a great script thanks) and have it running 99% the way I want it to. However, testing XProtec using 0 days as the license time, works perfectly fine. When I test for anything else however, it triggers and asks for payment, but if the user simply clicks Cancel or hits the X in the upper right, the Xprotec message box goes away and my script runs. Obviously, this doesn't happen when I tested for 0 days. Any help would be greatly appreciated. I posted on the Xprotec thread, but it seems somewhat dead, so I thought I would try here. Thanks!

My scanner is Kaspersky. I will see if I can reproduce with a smaller script as I have no intention of copying over several hundred lines of code. It's basically a GUI that contains an IE object embeded in it. When a user clicks a button, it switches to a different GUI using GUIDelete and creating a new one. This causes the scanner to issue the warning when run from the compiled script.

I have partially sorted my issue as I simply modified the from parameter in the SMTP script. Don't really like doing it that way, rather it auto, but it works. I do have another issue... testing on 0 days works fine. But if I set to anything above that, it triggers okay, but if the user hits cancel or the X in the upper right, it stops asking for payment and runs the script... w t f?

I did a few searches but couldn't find anything related, so forgive me if I have missed something. I am running a script that contains a GUI and on a few instances changes to a different GUI to go into different "modes". I did this simple via GUIdelete (originally opting for state changes to disable and hide) then creating a new GUI inside a new function that is called. However, when I run the program and switch modes, my virus scanner triggers and yells at me for "Hidden Data Sending". Is there something I could add/change such that this won't trigger?

I am now getting a variant of this to work with my own source. Thanks Val and Jos! I was wondering if anyone has attempted with dif. versions of windows besides XP? UPDATE: Everything seems to be working just great - but I am unable to obfuscate using JDeB's Obfuscator. I get an error regarding an array index, which is not present if I compile the same code without obfuscating. Any assistance (from anyone) would be greatly appreciated. I am talking about obfuscating my source code - XProtect is in an include. Do I need to obfuscate both before compiling ( I know odd question, but im pulling strings here)? UPDATE 2: FIXED (mostly) - can obfuscate my source with the beginning "XProtec" headers, but not the included XProtec. When I try to obfuscate both - I get index errors. im still having a minor annoyance (if that is the right word, as it's not a bug). If I run it through the google smtp server it works great. Shows the email address it comes from perfectly, cc'es another etc. However, if I run it through a Godaddy smtp such as smtpout.europe.secureserver.net the emails go through fine, but the FROM address comes in as theapplicationname@ClickTask.com instead of from the address I added, which occurs fine with the google smtp - and is what I would rather want. Any help here would be great - I know I can modify the P_Sender to a string of my choice, but I thought that would be sloppy.

Im begging this is my last hurdle to finish this thing! I would prefer to work inside the GUI, rather than having to have to open a new IE window everytime the function is called... I owe a beer to whoever figures this one out - or sends me on the right path.

Ive tried the function alone, with a pre-built .txt file in an iecreate() window and a GUI object... it only works with the former still - anyone have any thoughts? THIS IS DRIVING ME MAD! The docwrite HTML works perfectly fine in a new IE window. I have been following the change in the HTML with DebugBar. So... it should be exactly the same when I run the function in an IE object in a GUI... BUT NOOOOO it fails miserably... it's like it's changing other HTML... or the IE DOM (?) is getting tweaked? No clue.... SOS!

Ignore - I fixed it simply by using an input instead. I don't know why the integer was being passed differently with a combo.

I am starting to find some weird quirks in autoit as I come close to finishing my latest project. You can see some of my other concerns in the other threads I recently posted - some are getting close to being fixed however (I hope). My latest one revolves around the following: $get_article = Fileopen("Article_1.txt.",0) If $get_article = -1 Then MsgBox(0, "Error", "Unable to open content") Exit EndIf $t_body = FileRead($get_article) FileClose($get_article) $frame_list = _IEFrameGetCollection($oIE,0) $frame_nums = @extended MsgBox(0,"frames", $frame_nums) _IEDocWriteHTML($frame_list, $t_body) sleep(1000) $colTags_T = _IETagNameGetCollection($oIE, "input") for $sub in $colTags_T MsgBox(0, "submiter", $sub.value) if $sub.value =="Submit" Then ;MsgBox(0, "HERE", "HERE") _IEAction($sub,"focus") _IEAction($sub,"click") EndIf Next The problem here is the Submit button - not it NOT finding it, but the submit actually going through. It modifies the text of a TinyMCE frame on the IE page and then attempts to submit the content. If I run this function alone, after a simple _IECreate() method to get a spanking new window etc, it submits just fine - which $oIE is above. However, if I take the same exact code, stick it inside a new function like _submit(), and reference an IE window (e.g. simply $oIE_4) within a GUI (so a GUI Object e.g. Global $GUIActiveX4 = GUICtrlCreateObj($oIE_4, 10, 40+75+260, 600-100, 300)) - and obviously change all $oIE's to $oIE_4's above - it correctly fills out the MCE, hits the submit button... and boom... the submit fails. It doesn't recognize that there is content inside the MCE. Is the embedded object being handled differently somehow than a purely created IE instance? I have no clue what is happening as the code is exactly the same - just one is targeting an object in a GUI, and the other not. Can someone shed light on what might be going on here? It would be hard to paste my entire code as it's quite a large project, but I can add snippets if someone had an inkling of what might be happening. Please help! SOS! Cheers.

I need to bump this... I was able to change the mce as mentioned above, but when I hit "submit" it doesn't recognize anything that I entered within it, and removes it all. This does not happen obviously if I manually write material in the mce and submit it that way.... I suspected this was going to be an issue because after I _IEDocwritehtml it, I cannot click inside it any more, so something is getting messed up. UPDATE: I added the html to my clipboard before sending it to the _IEDocwritehtml. The submit button doesnt work if I click the Submit button after using the _IEdoctwritehtml to post the html in the mce. HOWEVER, if I paste it from my clipboard instead, then hit submit, it does work... what is going on?

The gui control is a combo (chooser menu as I would rather call it). And it just has numbers from 1 to 19 as those are the indexes of the <select> I am trying to index within. It works fine if I write 6 in there... but not if I define a variable as 6, then place the variable in the method... so I am really confused.

$category_chooser = Number(GUICtrlRead($GUI_Category)) $testing = _IETagNameGetCollection($oIE_4, "select") For $u in $testing _IEAction($u, "focus") _IEFormElementOptionSelect($u, $category_chooser, 1, "byIndex") Next The above code is really basic, but it doesn't work. It works if I delete "$category_chooser" inside the ...Optionselect method and stick in an integer myself, e.g. 6. But if $category_chooser is actually 6 it simply fails stating: --> IE.au3 V2.4-0 Error from function _IEFormElementOptionselect, $_IEStatus_InvalidValue (Invalid index value, 6) What's up with that???

Ignore - Got it Thanks!

Thanks Dale - I tried this before but for some reason it didn't work, but now it (partially) is! What I have done (and it's not working 100% yet) was to first loop through the number of frames on the page to see how many there were. I found 2 (0,1). Changing index 1 does nothing, but 0 changes the text in the mceditor, yay (I first tried to use your debuggerbar to find the index of the editor, but couldn't). However, it somehow de-activated the ability to manually edit inside the editor? I mean, I can see the text in the mceditor after _IEDocWriteHTML'ing to its index, but now I can't click in there and type anymore. Did I do something wrong? I haven't clicked the submit button, but hope it will actually submit the correctly written text... For before, simply _IEDocWriteHTML to a tag collection whos tag is $otag_t2.name = "html" (this is the mce editor) doesn't work - it writes the HTML in my entire IE object in my gui, changing the entire gui... One further question - One of the components on the page is a "suggested images" area, that changes dynamically. I won't know the contents of it, but I do know the <td> tag it is located in (I used your DebugBar to find the exact location). Can I click in the middle of a <td> tag without knowing its contents and with the IE functions?

Ok fair enough... so how would I go about doing that? Would it be via the _IE*doc functions? I could really use some help on this.. I wish it was as simple as locating and setting a value to the textarea.... As an update - I re-read the source, and the MCE on the page I am trying to modify is actually listed as a tinymc.js. I can see the javascript and everything, but its huge.

I cannot modify an MCE, that is what I meant. I tried on other sites with similar MCE editors. Although I can return the form value and name correctly in a console write, I cannot change the inner text.

Whoever helps me find the solution to this dilemma is a legend.

im loosing it over here... this is the worlds most frustrating thing, and it should be so simple given the _IE functions and how easy it is to grab inputs and other forms...

I cannot... is there a a precedent for using autoit3 to modify mce?

As an addition - here is more of the HTML that contains the textarea I am trying to get to: <td class="formInputLabel" id="editorLabelCell"> Article Body &nbsp;&nbsp; <a class="smoothbox formWhatsThisLink" title="How do I use the Online Editor?" href="#TB_inline?height=450&width=480&inlineId=editorHelpDiv">Online Editor Help</a> </td> </tr> <tr> <td> <table cellpadding="0" cellspacing="0" border="0"> <tr> <td id="editorInputCell" style="height:100%;"> <textarea name="html" id="html" class="mceEditor"></textarea> </td> <td style="height:100%;"> <div class="suggestImagesButtonDiv"> <input class="smallButton" id="getImageSuggestionsButton" type="button" onclick="$('suggestPicsDiv').innerHTML=''; doPaneAction('post','submit','suggest-pics','suggestPicsDiv',{'search-str':$('title').value +' '+ $('description').value +' '+ $('tags').value +' '+ tinyMCE.activeEditor.getContent()},true);" value="Get Image Suggestions"> </div> <div id="suggestPicsDiv" class="suggestedPicsDiv"></div> </td> </tr> <tr><td></td></tr> </table> </td> And again, I have tried the normal routes because the name is quite obviously "html". I have tried many things - I will include my messy bit of code below - the things I have tried are actually commented, and I don't know where to go.... $colTags = _IETagNameGetCollection($oIE_4, "textarea") For $oTag_T2 In $colTags MsgBox(0, "textarea", $oTag_T2.name) if $oTag_T2.name == "description" Then ;CLipPut($art_title_complete) _IEFormElementSetValue($oTag_T2, $introduction_inserter_1) ;_IEAction($oTag,"paste") sleep(1000) ;_ClipBoard_Empty() EndIf if $oTag_T2.name == "html" Then ;$html_input = _IEFormGetObjByName($oTag_T2, "textarea") ;$this = _IEFormElementGetObjByName($html_input, "html") ;$test = _IEPropertyGet($oTag_T2, "innerhtml") ;_IEPropertySet($oTag_T2, "innerhtml", "this is it!") ;ConsoleWrite($test) ;ConsoleWRite("FOUND IT!!!!!") ;ClipPut("checking to see if this works") ;_IEFormElementSetValue($oTag_T2, "cocker_pooper") ;_IEAction($oTag_T2, "focus") _IEAction($oTag_T2, "click") sleep(250) send("hi") ;Send("this is going to rock") sleep(1000) ;_ClipBoard_Empty() EndIf Next I have also tried the basic "focus" route and that also fails.....

The actual object I am attempting to edit is a textarea but of type HTML (mceEditor). Consequently, it throws the INPUT TYPE=FILE issue with my attempted focus work-arounds. Using the basic focus work-around itself, also doesn't work (as mentioned above). I do not think it is an iframe, although that was my first guess also. You can see the specific form I am looking at, at the URL posted. Any help would be great... getting really frustrated here.

Hello! Unfortunately I forgot my old username as I have not logged in in ages... so had to create another one. Apologies. Anyway, I am having significant issues adding text to a weird IE Form object. It is a text area, but its of type: <textarea name="html" id="html" class="mceEditor"></textarea> so, when I do the normal tactic of retrieving the form or tag collection, I can pinpoint the name = "html", but I cannot "focus" or propertyset this form element. I even attempted to do the _IEAction focus and click options, but since this is an mceEditor object, I cannot get it to add text to the text area. A simple "click" and Send() also doesn't work... so any help would be MADLY appreciated.