The e-book is out

[Martijn Verburg]

For this interested in the book, it's now out in e-book format, see http://www.manning.com/evans for details.

Thanks,

Martijn and Ben

[Ioannis Mavroukakis]

Many many congratulations to the both of you. I'm putting in my order for a dead tree copy so I can force^H^H^H^H ask you to sign it next time I see you ;)

[Martijn Verburg]

Thanks Ioannis and it will be our pleasure!

In some random Java 7 news for you all. Java 7u6 is about to go gold
and has an interesting change in the String implementation to close a
hashing issue. You can look at the webrev for the fix...

---------

Following testing and feedback the alternative hashing string keys in
hash maps which was introduced in CR#7126277 is going to be disabled
by default in 7u6. Developers can still enable the alternative hashing
but by default it will be disabled. More time is required for
developers to test their applications and correct improper usages
before this feature can be enabled as the default behaviour. The
alternative hashing of String keys feature remains the default for
Java 8 and may become the default for Java 7 in a future release.

Developers are strongly encouraged test their applications by enabling
the alternative string hashing feature before it does become the
default behaviour. The alternative Sring hashing feature is enabled by
setting the system property, jdk.map.althashing.threshold to a value
smaller than the capacity of the maps to be tested. The future default
is likely to be "512" which will have the effect of enabling
alternative hashing of string keys for all maps who's capacity is
larger than 511 entries. Small maps only encounter limited impact from
collisions and the higher threshold also masks incidental dependence
upon iteration order that may be present in those maps. For the most
rigorous testing, set the jdk.map.althashing.threshold property to "1"
which will force all maps to use alternative string hashing.

The current patch for review:

http://cr.openjdk.java.net/~mduigou/7180240/0/webrev/

This change is not a back port because jdk8 uses a different
implementation and is unaffected.

---------

Cheers,
Martijn