Resolve issue where the program aborts when running a large dataset with rotated text boxes in DOC1 Generate

When running Generate with a specific HIP, HIM, and a large input data file, doc1gen aborts partway through the records without any error message. If the input data is cut down in size, it runs to completion. No "bad" record could be found in subsequent tests.
 
The exact size which is able to process successfully is different for different environments; however, no particular resource (memory, disk space, etc.) seems to be exhausted during processing.
 
When run with only one device active the AFP run behaved the same.
However, a run with PDF-only aborted with the following error shown:
--------------------------------------------------
GEN0107I Action Print Message 'Entering Message Area Stuffer 7PDEX15PD3567048_000'
GEN0400A Failed to rasterize area of the page. Image creation failed. ImageMagick returned: 'UnableToOpenBlob `C:\Users\CH008KU\AppData\Local\Temp\savk.vvu.doc1.tmp.doc1.tmp.doc1.tmp.doc1.tmp.doc1.tmp.doc1.t??@xj.doc1.tmp.bmp': Invalid argument'
GEN0033A Production Job failed
--------------------------------------------------

Documents are outputting text in a Rotated Text Box. Before the graphics rotation feature was added to Doc1, this used to be output as text. Now the text box is being rotated/rasterized as an image. Merged PDF combines all these image files into the PDF file at the end of the run (to conserve memory), and during the run the files are created and held in the temp directory.

The data file is for 30,000 publications of ~4 pages each, with at least 2 rotated "MRA Codes". This leads to at least 60,000 files being created in the temp directory which leads to an ImageMagick error. This causes doc1gen to fail.