2 Apr
2012
2 Apr
'12
12:30 a.m.
The problem is that there may be work patterns where there's a significant risk of getting very long identical strings, e.g. if a file is read and cached for some time and then read again from another part of the program, or if the same file is read concurrently by different threads.
What's worrying is that it could occur suddenly in systems that has been running smoothly for a long time.