Fixes: RIPD-1601
Fix intermittent failure in server stream sub/unsub test.
Root cause is LoadManager thread *sometimes* running and causing a
fee change event which got published before our test could
unsubscribe. Fixed by explicitly stopping the LoadManager for this test.
These changes use the hash of the consensus transaction set when
characterizing the mismatch between a locally built ledger and fully
validated network ledger. This allows detection of non-determinism in
transaction process, in which consensus succeeded, but a node somehow
generated a different subsequent ledger.
When a test suite starts and ends, it informs the parent process. If the parent
has received a start message without a matching end message it reports that a
child may have crashed in that suite.
Fixes: RIPD-1521
Switch to pure doxygen HTML for developer docs. Remove docca/boostbook
system. Convert consensus document to markdown. Add existing markdown
files to doxygen input set. Fix some image paths and scale images for
use with MD links. Rename/cleanup some files for consistency.
Add pipeline logic for windows slaves. Add ninja and parallel test run
option. Add make doc target build in build-and-test.sh. Cleanup README
files. Add nounity windows build. Add link to jenkins summary table.
Add rippled_classic build (win). Improve formatting of summary table.
* Rename isArray to isArrayOrNull
* Rename isObject to isObjectOrNull
* Introduce isArray and isObject
* Change as many uses of isArrayorNull to isArray as possible
* Change as many uses of isObjectorNull to isObject as possible
* Reject null JSON arrays for subscribe and unsubscribe
* Remove all builds except cmake gcc & clang debug.
* Time some dependency and build operations, using a custom format to avoid
interfering with other timers.
* Use Travis's "trusty" infrastructure.
* Install more dependencies via apt.
* Update boost version to 1.65.1.
* Do not run unit tests under gdb - several security features prevent
it from running correctly.
* Run test job with two processes.
* Replace Ubuntu scons dependency with cmake
* Remove dependency installation support for Ubuntu 12.04 Precise Pangolin.
It is EOL'd as of 2017-04-28. http://releases.ubuntu.com/12.04/
The assert in LedgerHistory.cpp was incorrect since
getLedgerByHash() can return a null object. The assert has
been seen to fire inappropriately during online_delete.
The assert in LedgerMaster.cpp was simply misguided. If a
null hash were returned by getLedgerHashForHistory(), then we
should not forward that hash to peers. The assert only fires
in a debug build, so a release build would forward the request
for a null hash to peers.
These changes augment the Validations class with a LedgerTrie to better
track the history of support for validated ledgers. This improves the
selection of the preferred working ledger for consensus. The Validations
class now tracks both full and partial validations. Partial validations
are only used to determine the working ledger; full validations are
required for any quorum related function. Validators are also now
explicitly restricted to sending validations with increasing ledger
sequence number.
Per issue #2354, when the log level of a server was configured at
"trace", sensitive keying meterial generated by the `wallet_propose`
command could be written to the server's log file, if one was
configured.
This commit improves the log scrubbing code to account for the
sensitive information generated by a `wallet_propose`.
** Important security consideration **
We still caution everyone *against* executing this command on a
server that they do not control: a malicious server operator could
intercept the generated keypair, or operate a modified server that
returns keypairs that are not securely generated.
* This change passes detailed error messages from the JSON parser
on the server side, back to the client for inclusion into the
reply's error message.
* Errors originating from the server's inability to parse are
reclassified from rpcINTERNAL to rpcINVALID_PARAMS.
