There's no real significance to it, its just the way the directory detection logic works today. The double slashes on Linux get interpreted as a single one, and does not induce the behaviour in any way (same within Java's CLASSPATH parser, for which this entry is created)
Here's where the classpath entry is being built: /usr/lib/hadoop/libexec/hadoop-config.sh (you can vi this file on your installation)
Sourcing the variables that are being appended to CLASSPATH ('/*' is being appended to whose values) will reveal further that some of them come from the file /usr/lib/hadoop/libexec/hadoop-layout.sh (you can vi this as well) wherein they are set to './'.
As a result the concatenation operations on these strings produce the logic:
"/usr/lib/hadoop" (base) + "/" (separator) + "./" (specific sub-dir path) + "/*" (wildcard suffix)
The output of such a thing would thereby appear as "/usr/lib/hadoop/.//*".
Similar things apply to the non wildcard question, where we look to just add the directory itself onto the classpath.
Understanding-wise this is no different in evaluation than the reduced form of:
/etc/hadoop/conf:/usr/lib/hadoop/lib/*:/usr/lib/hadoop/*:/usr/lib/hadoop-hdfs/:/usr/lib/hadoop-hdfs/lib/*:/usr/lib/hadoop-hdfs/*:/usr/lib/hadoop-yarn/lib/*:/usr/lib/hadoop-yarn/*:/usr/lib/hadoop-mapreduce/lib/*:/usr/lib/hadoop-mapreduce/*
Hope this helps in reducing the confusion.
