To avoid pollution of that thread (it's supposed to help people catch up, not be bug central), I'm reposting Shane's original text and VI here. Further debate should be on this thread.
OK I think i might mave stumbled across a bug. Granted I'm only using 6.1, but maybe it's still lurking.
I was trying to use the "Scan tokens from strings" function to find a %nm% placeholder for a wavelength value. The problem is that it wasn't found. Not understanding this, but thinking it maybe had something to do with the "%" sign, I changed them to *nm* and it worked.
Seems logical enough. But then I just had to go and dig deeper. By replacing ONE of my token array with a *xx* palceholder, I suddenly started getting results. Wierd. Of course, I had to dig deeper.
Here's a VI demonstrating the problem. I've also found out that the caching behaviour is somewhat suspect.
Try the VI first with default values, then with caching forced "off" ("off" SHOULD be the default). The first and second results come from the same functionality, with the same inputs, but they are different. By forcing "no caching", the results are the same. I don't understand this, but it gets better.
Then try changing element 2 or 3 (which isn't present in the string to be searched) from a "%xx%" value to a "*xx*" value (simply change the "Array element to replace" from -1 (no action) to 2 or 3). Lo and behold, the %nm% token is found, and the forced caching has no influence any more.
I may be missing something with the "%" character, but can someone explain this to me?
Shane.
Using LV 6.1 on W2k (SP4) and WXP (SP2)
___________________
Try to take over the world!
