Content matching isn't part of matchRequest anymore; so it cannot be
integration-tested with other matching functions.
matchContent it is called directly from the streamingReceiver.
matchContent has individual unit tests.

OHM-831

OHM-831 Add back content matching for request bodies

When a channel's responseBody property (which determines whether the response body is stored) is set to false or undefined, the client should still get a response. This issue has been resolved in this commit and the authentication tests are now passing. A function that is no longer used has also been removed

OHM-896

OHM-896

The order of the middlewares had to be changed as the auto retry attempt number for a rerun transaction was being set after the logic that updates the transaction in the database. This was resulting in continuous rerunning for channels that have a max number of attempts

OHM-824

The method that completes the response and updates the transaction (in the db) was not updating the error property of the transaction (if there is an error). This was resulting in failure for some integration tests.

OHM-824

The response from an openhim mediator comes in a special format, and it has to parsed so that the response body, status, etc can be extracted. This means that it has to be stored in memory. The method that collects the chunks of the mediator's response is asynchronous and it has to be waited on. This is so we can set the right status and send back the body only (without the other fields on the mediator's response) to the client

OHM-824

Calling the method was not necessay as it is called in some other places once the response has been returned

OHM-824

whenever there is an error, the method updateWithError is invoked. Calling initiateResponse is not necessary as updateWithError does the updates that initiateResponse does. Calling the two methods at once was causing a race condition. At times the error was not being updated, resulting in some tests which have assertions on the error to be inconsistent

OHM-824

The property 'responseBody' of the channel has been set to true. This is because, at the moment when the property is false or undefined, the app does not send a response to the client. This sending of the response issue (when channel's responseBody property is set to false) has been addressed on another branch. This is a temporary fix to get the tests to pass.

OHM-824

One test has an assertion that expects the error message, when the secondary route fails, to be 'ERRCONNECT' and this was failing as the app has been modified such that when ever there is an error on the primary route the socket for the secondary route request is destroyed

OHM-824

… 5.3.0-alpha.1

fix responding to the client (responseBody -false)

The app was storing request bodies even when the channel had been configured not to store the bodies

OHM-898

fix the storing of request bodies

The logic had to be modified as all transactions that failed were being queued up for auto Retry, even when the channel is set not to auto retry a failed transction. The queueing of the transactions to be retried and the auto retries should only occur when the channel's property "autoRetryEnabled" is true. This way only transactions that fail after the autoRetry is enabled are rerun

OHM-824

This was causing the tests to fail as transactions are not marked for autoRetries when this property is false

OHM-824

The response body Id was being saved on transaction, and would never show up on the console

OHM-824

The response body id was not being set in the response headers and this resulted in bodies not showing on the openhim console as the transctions' body Id would be undefined. This was happening for responses from openhim mediators, which we handle differently

OHM-824

The orchestration's response body id was not being set, resulting in the response body for the orchestrations not showing up on the console

OHM-824

The body id should be added to the response headers. Logic has been added to check whether the response headers property exists before adding the body id. One test in which the openhim mediator response did not have headers was failing as a result.

OHM-824

The response from an openhim mediator is different and has to be handled differently. The mediator response contains a response and orchestrations, which have bodies. These bodies have to be stored seperately in gridfs and the transaction updated with the body ids. The api consumed by the console expects a body id for each body (orchestration and response). Because the body ids were not being set, the orchestr<ations displayed on the console were bodiless

OHM-824

One function for storing the transaction was removed, so the tests were failing

OHM-820

Orcherstartions were not being stored in the event of a failure on the primary route

OHM-820

OHM-820

The method that stores the transaction metrics was changed, it now calculaates the responseTime using the timestamp created at the end of the response streaming and not the one created at the start of the streaming

OHM-820

When a transaction does not exist in the database any method that tries to get or update the transaction should return a callback with an error or reject with an error. If this does not happen references to the transaction will cause the app failures

OHM-820

the function setFinalStatus returns an error and a null value for the transaction when there is an error. The error message logged out had to be changed as you cant get an id of a null value. this can cause a failure in the app

OHM-820

The logic that stores the response bodies from an openhim mediator has been modified to make use of a function that streams the bodies into gridfs. This is so we dont repeat ourselves as this logic was making use of a new function that does the same as the one that is now being used

OHM-824

The method for extracting payloads has been modified to resolve early, at the start of the streaming. This method is used in the routing middleware and this is being done to reduce the length of a request

OHM-

Ohm 824 fix auto retry

Responses from openhim medaitors have to be handled diffrently. The response body contains orchestrations etc, which have to be extracted from the body. This is done for the primary route and was not being done for the secondary routes, reslting in errors

OHM-906

… OHM-820-fix-all-broken-tests

…oken-tests

A method for waiting on the gridfs streaming of the response body to finish and for extracting the body from the stream have been added. These methods are used in the tests

OHM-820

The tests were failing because the request body is passed in a stream. This has been fixed in this commit

OHM-820

The code is made cleaner by making use of the ternary operator

OHM-906

The two functions rerunHttpRequestSend and rerunSetHTTPRequestOptions had to be modifed as some validation checks had been removed during the refactor. some of the fields for the expected ressponse were not being set. This has been fixed in this commit

OHM-820

The status message is expected in the response by the rerun functionality

OHM-820

The rerun task functionality was modifed and this required the tests to be modified
OHM-820

The method fro extracting a payload into chunks was modified to return the file id at the stat of the streaming. This was done to reduce the duration of a request. The tests for this functions have been modifed in this commit

fix the handling of openhim mediator responses

The properties 'isTcpChannel' and 'tcpChannelHasHttpRoute' are now added to the context for tcp channel requests. These properties are used in the streaminReceiver middlware, which comes after this middleware, to determine whether the middleware is bypassed or not. The streamingReceiver is used by the http routes. This will cater for cases when we route from tcp to http. These properties will be set to true for that case and then the streaming middleware is used

OHM-907

This has been doen because with the logic that was there the collectBody property was always true even when it had to be false. The collectBody is used to detemine whether the request body should be buffered in memory before anything else happens. This is for cases when we are doing request matching on content.

OHM-907

The streaming middleware is only used for http routes. THe logic will check whether we are routing from tcp to http, in which case the streaming middleware has to be used and request body has to be pushed downstream for the http router

OHM-907

This brings back rawbody reader for parsing the request body. THe body is used in the retrieveTcpTransaaction middleware to retrieve the tranaction and the channel. The request body has to be buffered in memory first. The flow of the middlewares has also been changed. The streamingReciever middleware is only used when routing from tcp to http

OHM-907

The functionality for storing the response body was being called so many times. It is now only called when the promises that do the routing have all resolved

OHM-907

The functions for storing the response and request bodies were modified (this is no longer being done in the a seperate middleware) and have been added to the function. THe functinality was not working, it is now

OHM-907

… OHM-907-fix-tcp-route

This is so the app does not break when a transation is not found. This also updates some of the app's dependencies

OHM-907

The passing of the tests was inconsistent and this is because ther was a race condition between the deleting of the bodies stored and other async methods that were doing read, update and delete operations on the boies

OHM-907

The logic for handling cases where the transaction does not exist (in the setFinalStatus method) has been modified. The callback that is returned by this function was being called before checking whether the transaction exists. The callback was being called with a value of null for the transaction which causes errors as the callback function tries to extract an id from null

OHM-907

The functionality for recording events for secondary routes had been commented out during the routing refaCTOR. This commit brings it back and also adds catching of errors to the completeResponse method which throws an error when a transaction is not found or when theres any other error

OHM-907

Eight events are recorded as the recording of secondary routes events has been brought back in. It had been commented out during the routing refactor

OHM-907

The route name is a a required property. Mongo errors were being thrown

OHM-907

The body was not being stored in gridfs and the transaction's request was not being updated with the body id for cases when we route from tcp to http

OHM-907

Some functinality had ben duplicated. It has been been moved into its own function now

OHM-907

Ohm 907 fix tcp route

Ohm 820 fix all broken tests

OpenHIM docs have their own repository now. This content has been moved
there and updated.

OHM-929

That version of Ubuntu is no longer supported. Nor do we support Ubuntu
PPA installs

OHM-929

Remove outdated script instructions

OHM-929

This script exists on master. However getting those changes could be
difficult...

OHM-929

New content

OHM-822

OHM-822 Bump alpha version

…alpha.1

Fixed afew broken tests

With changes made in Node 12, a new TLS version is used which caused some tests to break due to incorrect cipher usage.
Added a .env.test file to load environment variables when testing. This was due to an addition NODE_OPTIONS env variable that was adding which is starting to bloat the start line

OHM-938

Removed the unsupported node 8 (carbon)

OHM-938

Fixed broken TCP-> TLS tests

…alpha.1

… 5.3.0-alpha.1

…28-fix-merge-conflicts

…im-core-js into OHM-1028-fix-merge-conflicts

To make use of the newer packages which fixes bugs and security concerns.
This also fixes a bug with the module not starting due to an error with the log collection already existing

So that MongoDB can be started properly before continuing with the tests

OHM-1028

Verify mongodb is running

OHM-1028

To cater for bug fixes and security concerns
Also fixes an issue with mongodb making use of the newer typology

OHM-1028

So that the older depricated value is replaced with the new value

OHM-1028

OHM-1028

Was used for testing the travis build

OHM-1028

OHM-1031

OHM-1031

OHM-1031

OHM-1031

This is causing something to break
Will update and investigate a bit later on

OHM-1031

So that we can start testing on the newer LTS version of Ubuntu

OHM-1028

OHM-1031

An issues has been picked up that on first run of the tests, the SSL Root Certificate Authority needs to be downloaded into the ssl-root-cas module
This takes some time and test fails due to the timeout before this is completed

OHM-1028

OHM-1031

OHM-1031

OHM-1031

OHM-1031

OHM-1031

OHM-1031

OHM-1031

OHM-1031