Encryption / Encoding Differences between IDE and OS X Standalone

[Peter W A Wood]

Many thanks to Roger Eller, Jan Schenkel, Bill Vlahos and Paul Hibbert (off list) for helping me identify and workaround the issues I encountered.

Paul pointed out three things: I didn't the use "the text of Field ..."; how to overcome not being able to copy text from a field in an OS X standalone; and the final problem that was related to script local variables.

Using "the text of Field ..." instead of "Field ..." helped to more accurately copy and paste data. (I ended up displaying the encrypted data in Hex anyway.) 

I had found that I couldn't copy from a Field in an OS X standalone. Paul provide a script with group handlers to get around the problem. I have submitted a bug report which has been accepted - http://quality.runrev.com/show_bug.cgi?id=11830

At this stage, I was still getting inconsistent results when encrypting data in the IDE or a standalone. Then Paul noticed how I was initialising the encryption parameters, namely the script local variables tPassword, tSalt, tIV:

	openStack
  		put "" into Field "Encrypted"
  		put "12345678123456781234567812345678" into tPassword
  		put "12345678123456781234567812345678" into tSalt
  		put "12345678123456781234567812345678" into tIV
	end openStack

	on encryptMSG
  		...
	end encryptMSG

	on decryptMSG
  		...
	end decryptMSG

He changed this to re-setting the three script local fields each time they were used:

	on openStack
  		put "" into Field "Encrypted"
	end openStack

	command setPassCodes
  		put "12345678123456781234567812345678" into tPassword
  		put "12345678123456781234567812345678" into tSalt
  		put "12345678123456781234567812345678" into tIV
	end setPassCodes

	on encryptMSG
  		setPassCodes
  		...
	end encryptMSG

	on decryptMSG
  		setPassCodes
  		...
	end decryptMSG

After making this change, I now get identical results when encrypting data with the stack in the IDE or as a standalone.

I have also entered a bug report for this - http://quality.runrev.com/show_bug.cgi?id=11832 - I feel there is a bug but not in the encrypt/decrypt functions, it's in the handling of script local variables.

Once again, thanks for all the help.

Kind regards
Peter
http://LiveCode1001.blogspot.com