This changeset validates state at startup, switching the
assumption from state being valid to state being invalid.

Further, once state transfer completes, the replica will
now broadcast a checkpoint for the state it transferred to.

https://jira.hyperledger.org/browse/FAB-379



Change-Id: I491451f4829f14c9167ceed5d29befe3a3b08521
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

FIX FAB-158

https://jira.hyperledger.org/browse/FAB-158



Change-Id: I79aae2b4e537a20a72b2fcc1bf83a40087cfeed6
Signed-off-by: Satheesh Kathamuthu <satheesh.ceg@gmail.com>

Currently, the Node.js SDK unit tests are cloning the
master branch from Gerrit in order to include it as a
dependency inside pre-built chaincode packages. That
causes the unit tests to fail on all other branches
except the master branch. Until the fabric dependency
is entirely removed from the SDK unit tests, the test
script will instead use the local copy of the fabric
project to avoid cloning the dependency directly from
master. This avoids unit test failures due to
incompatible versions.

Change-Id: I995d16b98f1e8d337a8cd44c87133c92542437d0
Signed-off-by: Anna D Derbakova <adderbak@us.ibm.com>

Modify the transaction id for a deployment transaction
to the chaincode hash. The Txid was being incorrectly
set to a generated UUID.

Fixes FAB-288

Change-Id: I69736e05e2240b79ae3bb5bf378abf519ebbf993
Signed-off-by: Anna D Derbakova <adderbak@us.ibm.com>

Remove remaining references to "account" parameter,
which was previous removed from the ECert and the
registration request.

Change-Id: Ia4233af3f711175910911a3469b29789c5238e44
Signed-off-by: Anna D Derbakova <adderbak@us.ibm.com>

Change 899 introduced a new unit test for the node sdk
which causes the node sdk unit tests to fail.  The fix was to
change the chaincodeID from mycc3 to mycc4 for the new test.
This amends FAB-285

Fixes FAB-285

Change-Id: I8bdfec16f7e46727995cb755c9e51a5e27658de0
Signed-off-by: Gari Singh <gari.r.singh@gmail.com>

Currently only local deployment of Java chaincode is supported.

Change-Id: Id7651f6eeae232260f76f2280b935f7c9a15d169
Signed-off-by: Satheesh Kathamuthu <satheesh.ceg@gmail.com>

There have been a number of instances of BDD tests which were
starting correctly but due to high load on the Jenkins host were
unable to complete prior to the timeout expiry.

This can been seen in action during
https://jenkins.hyperledger.org/view/fabric/job/fabric-verify-behave-x86_64/19/


in the chaincode authorizable counter test. Looking at the peer logs
it can be seen that peers vp1, vp2, and vp3 started up in 9, 1 and 5
seconds respectively. vp0 however timed out and didn't start.

Increasing the timeout from 10 to 20 seconds for peers to connect to
membersrvc and from 15 to 30 seconds overall to combat this.

Change-Id: I5c45697a63b95dd31371f423d274de69f0dc5cd3
Signed-off-by: Julian Carrivck <cjulian@au1.ibm.com>

The log prefix should be "stop" in DockerVM.Stop.

Change-Id: I551b37f62238791ce9506aad2c558c7bca73c81f
Signed-off-by: jiangyaoguo <jiangyaoguo@gmail.com>

Move the BDD Docker compose files into a new sub folder so they are all
in one place and do not clog up the root of the bddtests/ folder.
While altering logic to determine the container names I also refactored
the Docker-Compose parsing code to make it more clear as to what it was
doing (parseComposeOutput in bdd_compose_util.py).

I needed to add a dummy compose-defaults.yml file so that CI would
successfully pass. When this change makes it into all branches then it
can be removed and the CI scripts updated.

I also changed the permissions on a couple of the bdd files from 755 to
644.

Signed-off-by: Julian Carrivick <cjulian@au1.ibm.com>
Change-Id: Iefd1007e79ee3827d39ed9b0bdec327f30e0c39f

when receive a rejection event with no "-listen-to-rejections",
adapter of block-listener push a nil to notfy channel. This will
cause nil pointer panic. Fix the handle logic.

Change-Id: I8391af03acf32c6d99d613db6e2e7dc13d2424cc
Signed-off-by: jiangyaoguo <jiangyaoguo@gmail.com>

Introduces the behavior required in FAB-189:
https://jira.hyperledger.org/browse/FAB-189



Change-Id: I1a0e3809c5f58b7e256e5e27e0e8722b6f674725
Signed-off-by: Kostas Christidis <kostas@christidis.io>

There is an issue with using the latest version of the sjcl package.
This fix locks the versions of both sjcl and sjcl-codec
(since they are tied based on how sjcl-codec was built) rather
than pulling the latest.

Fixes FAB-221

Change-Id: I1410dbce21642c3f42647ae26b55ea9172f6b002
Signed-off-by: Gari Singh <gari.r.singh@gmail.com>

Change-Id: I4ce07658b1ee63756fdd113c1713d9cf411180ff
Signed-off-by: Kostas Christidis <kostas@christidis.io>

I'd like to nominate Gari Singh to become a maintainer for the fabric
project. Gari has really stepped up his level of contribution and
reviews over the past few weeks. His reviews are always thoughtful
and constructive. He has also been very active on Slack, helping
newcomers with their questions about how things work, and are
put together.

Gari, for those that do not know, is CTO for the IBM blockchain
development team, so he definitely has a vested interest in
helping us create an awesome open source foundation.

Change-Id: If888565389ac440815a00ca3fdd564bc51df0f17
Signed-off-by: Christopher Ferris <chrisfer@us.ibm.com>

Print out ChaincodeName as doc said so that we can reference it
when invoking or querying deployed chaincode.

Change-Id: Id264b00d694d5d303f035acbe9e212a020a03da0
Signed-off-by: jiangyaoguo <jiangyaoguo@gmail.com>

Before receiving a hello message from fanin message channel, we can't get
peerID of fanin message channel. So RegisterChannel always gets
RegisterChannel(nil, channel). PeerID here is supposed to do something
like session management (eg. duplicated peer connection check). But at this
time we can't get peerID. The session management is done by class "Impl"
with RegisterHandler and DeregisterHandler. Here PeerID is unnecessary, and
it will always be nil. Incorrect warning of "Received duplicate connection
 from nil, switching to new connection" will be emitted everytime there is
a new connection.

Change-Id: I9f1bf6287576497acefb46a543fd4ac62d062665
Signed-off-by: jiangyaoguo <jiangyaoguo@gmail.com>

Refactor the duplicated code for calling the peer REST endpoints into
standalone functions that can be reused.

Also refactor the 'Wait for transactions to be committed' BDD steps
to utilize the newly created high level functions.

Remove the msg parameter from the JSON checking function and generate
the message at invoke time.

Reduce some timeouts from 5 or 3 mins to a much more reasonable 1
min.

I have tested all behave tests locally to verify they work in lieu of
them being run on Jenkins while they are stabilised.

Change-Id: I484c1cde677461a885863a492be786bae3be44f3
Signed-off-by: Julian Carrivick <cjulian@au1.ibm.com>