Buffering

Decrypting a PML file takes longer than reading a plain-text version, and in some circumstances PML files may be re-read many times during a session. (A new command PML STATISTICS displays information on the numbers of times each file has been read and some extra information useful to AVEVA when testing the Published PML facilities).

In order to reduce the time taken to re-read the files, Published PML files may contain a buffering directive in the header-line (the first line in the file). If a dash and a number are included directly after the three-digit encryption algorithm ID, then the software will retain the file in memory indefinitely once it has been read that many times.

You may wish to edit heavily used files to add buffering to the header by hand, or may use the -buffer 5 option of pmlencrypt.exe to include a "buffer after five reads" tag in each file encrypted.

A value of five is a good number to start with. Many files are read precisely once during module start up - there is little benefit in buffering those files, and a value of five will avoid that, but apply to all heavily used files.

If a file you are actively developing has a header including buffering, it will not be re-read as often as you are used to. To force all buffered files to be cleared from memory if they are not in current use, you can issue the PML REHASH or PML INDEX commands, or switch modules.