When using Oracle DataServer we are required to recompile every program that accesses a table to which new fields have been added. The same behavior is not observed when using MSS DataServer or native OpenEdge database, where we can continue to use pre-compiled r-codes provided the new fields have been added at the end of the table.

This behavior is pointed out in the last two paragraphs of the article https://knowledgebase.progress.com/articles/Article/000065440:

"Adding a new field in the schema holder and allow pre-compiled r-code is a separate feature and at the moment of writing this article, December 28, 2015, is currently available in MSS Dataserver from 11.4 onwards.

For Oracle Dataserver adding a new field will still give a CRC mismatch error if try to run the pre-compiled r-code; recompiling r-code is required in Oracle Dataserver. There is an estimation however that this feature will be available in OpenEdge version 12.0 onwards."

This matter has also been discussed with Tech Support through cases 00333115 and 00429317.

The request is having the ability to run pre-compiled r-code after adding new fields to existing tables.

We are an ISV with a large code base and many different customer configurations and this requirement to recompile code for Oracle is causing issues, mainly around build times, dependency management and larger deliverable to our customers.

We've been able to address many issues for Progress and MSS builds, but Oracle builds still hurt.