(v13) Starting up the Harlequin RIP

 This page applies to Harlequin v13.1r0 and later; and to Harlequin Core but not Harlequin MultiRIP.

This stage starts with a call to invoke the core-RIP. Core-RIP customers know this as the C function SwStart.

• Memory is received from a structure passed to SwStart . Memory passed to the RIP through SwStart is for the core-RIP’s exclusive use, and that use begins as more or less the first action of the RIP.
• The %os% device is mounted (created) and associated with device type 0. Device type 0 must exist among those device types passed to SwStart. This causes the first call back out from the core-RIP: to the device type 0’s initialization function, which must succeed for the RIP to continue operating.
• The %null% device is mounted and associated with device type 1. %stdin%, %stdout%, and %stderr% are associated with a file on this device.

Device type 1 is a non-relative device type, and is pre-defined by the core-RIP. (However it can be overridden in custom implementations by passing a device type 1 via SwStart. This might be desirable in order to collect messages emitted by the RIP before the pseudo-devices can be remapped to a more useful file. The %null% device would then be said to “leak”).