Mmm... more testing from old macros with MMM V6.7.0 has shown problems again. Switching back to V6.6 solved it. I am guessing V6.7.0 has much less delay in execution, therefore causing the problem. It would have been nice to have a converter to convert V6.6 macros to V6.7.0 macros by adding time delay automatically, such that the V6.6 macros would slow down to work with the faster V6.7.0.
I think I am keeping both the V6.6 and V6.7.0 MMM applications on the desktop to avoid debugging the old V6.6 proven macros while running on MMM V6.7.0.
How much faster are you finding execution using 6.7 rather than 6.6? Version 6.7 (like every new version) adds to improve what I like to call the 'macro processing engine' to make macros run cleaner. I haven't experienced a great deal of execution speed change with the new version. I do recommend using the newer version over the old one. Version 6.7 internally is better than 6.6 and will make for smother running and recording.
Yes your right this is now a little different. The difference is the way 6.6 defines new %integer% variables. 6.6 is happy to pass the string word %clipboard% to %integer% and then check if the contents of the clipboard is a word or a number. However 6.7 does not want to accept a string word at all when defining a new numeric value into any %integer% variable. From memory the change was made due to some tweaking of the interval variable system with the addition of all the FOR loop variables. I will look to changing this I think. As you've pointed out I think it would be better to check first the contents of a string variable when it is passed rather then cut it all together.