(//" was needed and  // " would pass through)

This will hopefully make it easier for apps to use bits of I2P as build-deps.

   - Add SIGNATURE_TYPE support to GENERATE and CREATE
   - Don't NPE checking dest+privkeys
   - Simplify HELLO checking
   - Don't require two params in HELLO message
   - Make MIN parameter optional too
   - Version checking fixes
   - Bump version to 3.1, only visible if requested
   - Cleanups, javadocs

   - Check for extra bytes in private key string
   - checkPrivateDestination() returns boolean instead of throws

   - Fix empty properties check
   - Overrides
   - Statics

  Version handling (ticket #1318)
   - MAX param now optional
   - 1-digit versions now accepted for MIN and MAX
   - Use VersionComparator for version tests
   - Don't require a minor version of 0
   - Fix empty properties check
   - Overrides
   - Prep for version 3.1
  Throw exception on null option key (ticket #1325)

     Move fields to top

     Use DataHelper.getASCII() for byte conversion

     add new SAMException constructor

API 9 in the next Android release.

Also move away from using external binaries when an ant task can do the job.

   as added by 'dream' in late 2011. Testing couldn't have happened,
   as they don't work and can't ever work as designed.
   Channels must have underlying file descriptors unless you implement
   your own Selector and that would probably require JNI.
   See http://zzz.i2p/topics/1229 for details.
   Also http://stackoverflow.com/questions/911780/how-do-i-define-my-own-selectablechannel

At this point it's been offline for a month with no sign of returning.

  - Display webapp name in html title (ticket #1311)
  - Use tracker from magnet link in display (ticket #1313)
  - Clarify auto-start (ticket #1293)
  - Include tracker in magnet link on details page (ticket #964)
  - Recognize .azw4 extension as ebook
  - Cleanups

 - Serializable
 - hashCode()
 - Make DataStructureImpl Serializable (removed from DataStructure in 2005)

- This came from the script from Tanuki but it does return useful information
  (as far as its use in the script) in my testing. uname -m is better for our
  needs. (The problem is only seen on certain CPUs when *all* available wrapper
  binaries are present and the script tries to resolve the correct binary to use..