How to detect end of copy to Clipboard

[PJThys]

Is there a safe way to detect when a copy to the clipboard is finished. Afterwards the contents of the clipboard need to be written to a file. This file is then parsed.

Snippet:

Send("^a^c")

$file = FileOpen($Target & "\" & $Test & "TMP.txt", 2)

; Check if file opened for reading OK

If $file = -1 Then

MsgBox(0, "Error", "Unable to open file.")

Exit

EndIf

FileWrite($file, ClipGet())

FileClose($file)

gave me an empty file, only 2 sleep(10000) (one after ^a^c and one before FileClose) gave me results in 4 out of 5 cases, but slows down the process a lot.

Then I tried trapping the wait cursor (cursor_id 15)

Send("^a^c")

MsgBox(48, "", "Wait a moment please", 3)

While 1

MouseMove(5, 10)

$MouseCursor = MouseGetCursor()

If $MouseCursor = "15" Then

While 1

MouseMove(5, 10)

$MouseCursor = MouseGetCursor()

If $MouseCursor = "2" Then ExitLoop 2

WEnd

EndIf

WEnd

$file = FileOpen($Target & "\" & $Test & "TMP.txt", 2)

; Check if file opened for reading OK

If $file = -1 Then

MsgBox(0, "Error", "Unable to open file.")

Exit

EndIf

FileWrite($file, ClipGet())

FileClose($file)

Performance is much better, but it seems not always to trap the wait cursor (I'm not sure that it always appears). Any idea on how to find out whether ^a^c is finished writing to the clipboard? Manualy opening a text document and pasting with ^v always works, but it's rather ugly, and poses problems because of different OS and installed languages on pc's so it's more difficult to save and close notepad (different commands/window titles)

Furthermore I'm not sure that the fileclose does not close the file before the clipboard is written completely to the file. Anyone knows if fileclose is waiting till filewrite is finished?

Sorry for the long post, but I've been working for 2 weeks on this script and this and the previous post represent the last problems. I could live with it myself, because I discover immediatly when it doesn't run like intended and can try again, but to give it to our secretaries when it is not 'foolproof' is something else

Thx in advance

Peter

[flyingboz]

long winded ramble re: long cut/paste operation detection

You didn't mention pertinent information re: what app you're copying from - or what text is being copied / parsed, or if there are a mininum number of bytes being grabbed. Any of those factors may be leveraged.

Any markup language has tags , etc. etc, etc. If you're using any COM-aware app you may be able to utilize COM functionality - IE and Office apps expose methods / properties you can use to increase the reliability

You could also "salt the shaft" , append an unusual string to the text prior to the ^a^c, and then loop your ClipGet() until the salt appears.

$salt = "unusual_append_string"
While 1
  $text = ClipGet()
  If StringInStr($text,$salt) then ExitLoop
  Sleep(500)
Wend

Another methodology might involve a third party clipboard util, which you could salt, then get the clipboard -1 entry after the last one occurs.

Just a few ideas off the cuff.

Edited October 28, 2005 by flyingboz

[PJThys]

Thanx for your reply.

The application is Questionmark Perception, a program to make electronic tests. It is a black box to me.

A test consists of between 160 and 200 questions.

Each separate question is made in a separate window where a lot of extra information has to be entered for electronic correction of the test. As author of a test we can make the test with QMP, but reporting or debugging the test is difficult. If we forget to fill in a correct answer to a question, we only can see that by trying out the test (fill in all questions) and at the end of the test we push the continue button and the test comes with the correct answers. Then we can check if every question has an answer. The size of the copy is unclear to me. If I cut and paste to a text editor the result is about 75 kb, but all pictures are of course removed. That's no problem for my purpouses. If I try to paste to Word it takes about 15 minutes and results in a word doc of 2 Mb. To parse that is very difficult. The size will be between 0.1 and 2 Mb I couldn't tell exactly. Writing a string to the test window is impossible, I could only write such a string to the end off the last question (which of course we will forget to remove before final publication of the test

I Only want to be able to produce:

1 Flat text file with questions and answers (find spelling errors)

2 Flat text file with questionnumber and answer (to give to the students later: these would have been your correct answers) and this shows me the questions where we forgot to enter an answer

3 CSV file with questions, answers and topics for processing statistics about test (20 questions about respiratory system, 30 questions about old people, etc)

I can produce all these files from the copied flat text file.

I'm rewriting my script now. I had a gui where some parameters could be entered, it used messageloop-mode. I'm rewriting it to use on-event mode, suspecting that that might be the reason why my cursor is not always trapped! and then my second solution might work always

btw QMP did a nice job in preventing students to cut and past or copy the final published test, I was not able to find a tool to copy the final test, when for example kleptomania, a sofisticated screengrabber was loaded, QMP refused to start, stating that I first had to remove kleptomania from memory!!

[Confuzzled]

Some ideas:

Does the QMP program support Windows Accessibility functionality (output for the blind and hearing impaired)? If it does, then there will be some sort of secondary output mechanism that you should be able to tap into somewhere.

Increase your memory availability for the program to function quicker by shutting down unnecessary programs running.

Between clipboard operations flush the clipboard buffer by copying a single character to the clipboard.

I hope you are not developing the ultimate exam cheat program!

[PJThys]

No, no more exams at my age (54), We are develloping exams!

As far as I know there is no accessibility functionality

[Confuzzled]

No, no more exams at my age (54), We are developing exams!

As far as I know there is no accessibility functionality

Why not approach the developers of QMP and ask for added functionality to make your life a little less difficult? I'm sure if they value their product they will make it as easy as possible for others to use it. They may already have added tools for educators, or be happy to use you as a development tester while they develop them...

[flyingboz]

..difficult to cut/paste text

You may wish to invest in a copy of www.screenocr.com - i have used it to defeat other obstinate programs.

copy / paste buffer 75k

au3 clipget() at one time had a hard-coded limit of 64k - you may wish to check the current version docs to see what the limit is now and ensure that you aren't running into a size limit - IIRC, the size limit was due to a win95 compatibility issue. Edited October 30, 2005 by flyingboz

[PJThys]

'flyingboz' , you set me on the right track! I gave up watching cursorchange, too tricky.

Clipboard size was not the problem, throwing in 2 times a 10 seconds wait would solve the problem. But waiting 20 seconds for a situation which only seems to happen in 20% of the cases is a shame. So I worked on your SALT idea, bij testing ClipGet() for the occurrence of a word I was sure would be in the clipboard. This seems to have solved the problem.

My code is now like this: (where "comment" is a word I know exists multiple times in the copied text)

Send("^a^c") ;copy all to clipboard

While Not StringInStr(ClipGet(), "comment ") ;wait till "comment " is found in clipboard

;(assuming that all text is copied to clipboard then)

Sleep(10)

WEnd

Allthough this word occurs a lot in the clipboard, StringInStr(Clipget(),"comment") does not seem to start before the clipboard is written.

Can anyone confirm that behavior?

I've tested it 20 times now and found no more 'timing errors'

Peter

Edited October 31, 2005 by PJThys