[Info-vax] RMS and SSIO (again)

[Greg Tinkler]

Have spend a bit of time researching and thinking through the SSIO / RMS issues and it occurred to me that may be the SSIO is the right direction just not far enough.

What is critical is to ensure the RMS can access all files, and until VCC/XFC RMS did its own buffering.  SSIO appears to be about access XFC bytes directly, and use XFC to coordinate the updating of the buffers and therefore the on-disk data.  Good.

Lets assume the SSIO interface will be $QIO possibly using something like IO$M_RBYTE/IO$M_WBYTE.  This would in effect be readp/writep (linux IO routines) on VMS.  So lets externalize them as SYS$READP/SYS$WRITEP.

With the move to using XFC as the main buffering engine on VMS lets also remove all the RMS internal buffering.  NB this does not remove SET RMS/... as those are still used in setting the FAB which passes that information to XFC.  For those that don't know XFC uses the smallest MBC for a file as the buffer buffer size, and coordinates access to that buffer data across the cluster.

This would greatly simplify the current RMS code, and also by using $READP/WRITEP internally there is potential to reduce the RMS code base further, while also making RMS more Linux like.

Ahh the things we think of over the time of year...

gt