The main theme of the release is defect fixing and that includes PMRs but also the work we have done to adjust, simplify and make more consistent some of the features we have introduced in 1.2.
We’ve also worked to improve the RPE performance, focusing on the core processing part. The RPE core processing is the document generation part that lies between data input and writing the output files.
But the core processing is just one of the document generation stages and so the overall document generation time depends on how fast the other two parts, input and output, take. Between the two, the input part usually has a greater weight so the slower the data retrieval is, the more it will weigh in the overall time and this will reduce the impact of the core processing improvements we have made.
As a first step in improving the input part we read ahead the pages in a paged data feed. Rather than request a page only when needed, RPE will now request the next page in a feed as soon as the previous pages is downloaded and while the RPE core processes it.
There are several other changes in the RPE code that should further boost performance including the ability to use the RPE Remote Generation Web Services in a load balanced environment.