OpenSRF 2.5 release notes

osrf_control --diagnostic now includes the maximum number of drones allowed per service, making it easier to determine if drone usage is hitting limits.

For example:

The calculations of whether to split a large message into multiple chunks now take into account additional overhead added by the JSON- and XML-escaping of certain characters. As a consequence, large messages are now much less likely to overflow typical ejabberd max_stanza_size values.

The premature failure of a drone process that was just starting up, as might happen if certain ejabberd limits are reached, is now less likely to cause a listener process (and thus the entire service) to abort.

OpenSRF has long inspected the envelope of incoming requests for information about the client’s locale and made this information available to business logic.  This is used, among other things, to drive transparent translation of in-database strings within Evergreen.  In addition to locale, OpenSRF will now respect client-supplied adjustment to the effective time zone in which it operates, and provide that information to the business logic layer of applications built on the framework.

As most clients that have time zones which differ from that of the server on which the OpenSRF processes run are, in fact, web browsers, it is necessary to include time zone detection directly within the browser.  This will be stored in a cookie to be sent with all subsequent HTTP requests, and used in all OpenSRF-over-HTTP calls made using the JavaScript bindings for OpenSRF, including those for WebSockets communication.

For non-browser clients, such as support scripts written in Perl or other scripting languages, the local system’s mechanisms for detecting time zone is relied upon.  For instance, Perl scripts can directly read the TZ environment variable.

Additionally, the srfsh client now reads its local time zone from the environment and passes that to the server.

Within OpenSRF services implemented in Perl, this information is now passed up to the business logic layer via the TZ environment variable, and is reverted to the server’s value at the end of each request.  This allows automatic, transparent use of the client’s time zone in almost all cases, and provides a system-normal access mechanism when direct access is required.

For OpenSRF services implemented in C, the time zone information is provided as part of the request context object that is passed to implementation functions.  In particular, this allows services that interact with a database to set the time zone in which the database interprets timestamps to that of the client.

There is a pattern in the wild of using OpenSRF’s method_lookup() facility to decide between one of several local methods when delegating to pre-existing logic. Often times, we want to simply hand control over to another method, but the output of a subrequest’s run() is an array of results. The caller has to know if, and how, to restructure the result for the client.

Instead, we can now call dispatch() instead of run() and have OpenSRF session control completely passed to the delegate code. This way, the delegate code need not know anything about its caller, and vice versa.

OpenSRF 2.5 comes with example configurations for using HAProxy or NGINX as a reverse proxy for HTTP, HTTPS, and WebSockets traffic. This can be useful for Evergreen systems that wish to use port 443 for both HTTPS and secure WebSockets traffic.

Along with the example configuration files, there is now a new configure option, --with-websockets-port, to allow installers to specify the WebSockets port that JavaScript clients should use.

Add --with-perlbase option to the configure to specify an alternative location for installing the Perl modules. This can be useful for setups that want to run the Perl modules from a shared filesystem or environments that need to run multiple versions of OpenSRF simultaneously.

Users of --with-perlbase are responsible for ensuring that PERL5LIB is set appropriately.