It is unclear what changes would be needed to FWD to support an existing Progress schema which has not been converted using the expected conventions. The FWD schema conversion and runtime persistence framework are built upon certain assumptions about the structure of a converted schema, in particular with respect to the conversion of extent fields and the addition of a surrogate primary key for every table.
OpenEdge 11.7 is supposed to have support for Java UDFs but it has not been verified or tested. The presence of these prerequisites is unclear at this time. In addition, the database must provide an API to define and allow the invocation of these UDFs, and must process parameters to these UDFs in a deterministic order for FWD to operate correctly. This library (and thus a JVM) must run within the database server process, or within whichever process executes database query plans (presumably the Progress SQL Engine in this case). FWD uses a library of UDFs implemented in Java to support the use of Progress built-in functions within WHERE clauses. There are concerns regarding the stability/reliability and performance of this access mechanism. Thus, it cannot access a Progress database directly, and would have to go through the Progress SQL Engine. This would require some prototyping and testing. There are some concerns and unknowns which would have to be investigated and addressed in order for this option to be viable.
Currently, FWD does not support the Progress database as a back end and it is not clear how feasible it would be to do this.
0 kommentar(er)
