I beleave I saw it is 1,000,000 chars, but not found.
Indeed FM14 worked with at least 2^30 = 1,073,741,824 chars...
I used string with repeating 1 char, so there might be some compressing technique?
No, working memory increased about 2GB while the variable alive.
Since you haven't found anything, my guess on client depends on how much memory can be allocated.
with PSoS it's 1 million ( 1,000,000 ) for parameters and script results.
I'd test it with the data you have, triple and quadruple it and see when it craps out.
PSOS! It should be what I saw.
I change the script I tested to increase loop 1 time = twice string length, then FM crashed.
Working/Commit size are about 2GB, there is many remains on system having 8GB RAM Windows7.
FM crashed when I change the length to 2^31.
Then, I tried to get limit of commit size, increase string length from 2^29.
But in my environment when reached about 6+GB WINDOWS hangged with no error, forced to power off.
Did FMS or FMP crash?
FMP. I tested with newly created standalone file, it is not for PSOS.
And, I did it on my personal computer, with many other program running. MS SQL Server, FMS, PostgreSQL, Firefox use for listening net radio, Thunderbird. So the reason may not be only FMP, but that occurred the time.
Total commit size is limited 15GB, then I guess reaching it crashed...