The Question is:
Is there a practical way to derive which source code files were used to create
an image, apart from keeping better records?! ANA/IMAG doesn't seem to help
and DUMP gives some clues but not the version number or location.
An internally developed application was last compiled several years ago by a
long-gone team and it is now unclear which .C, .SC and .H files were
originally used in its creation. Some further enhancements are now required on
it but multiple versions of th
e potential source files exist and apart from building the (6x6x6x6x4x5)
versions and comparing with the current image, it is not immediately clear
where to start.
PS - A revised CM process has now been implemented!
The Answer is :
There are ways of reverse-engineering an image, but the effort will
involve tracking the source compilers, the input libraries, and the
source code -- and this effort is more work than it is probably worth.
If the image contains full debugging information, you can obtain
the compilation input files, but these files may or may not be the
actual source files. For example, Oracle Rdb pre-processes files
with embedded SQL, and only the information on the results of the
pre-processor are included in the debugger symbol tables.
For a product release, the OpenVMS Wizard recommends maintaining the
source files, the build procedures, the compilers and tools required,
the /FULL source listings from the compilers, and /FULL maps from the