🚀 go-pugleaf

Has anyone worked on/considered the problem of making NDBM database files
portable (ie. byte-order independent)? This could be done by
	1) a filter to change byte order between two databases
	2) a library change to swap bytes on one of the machine types

Especially in an environment like Athena, where we have numerous workstations
of opposite byte order (DEC VS2/VS2000 versus IBM APC/RT) it would be
convenient to create one set of files to be accessed by both (solution 2, with
the benefit of saving disk space).

Since conversion in the library might slow one machine type down, solution 1
would still be useful, since it should be faster than actually building two
full databases --- my personal application creates a database with many large
keys (38000 keys, average length 10 chars, max length 80 chars) and takes hours
on an 11/750 to local disk (and has never run to completion over NFS...  >6
hours though.)

If you have ideas, patches, or code, respond by email (I will summarize here.)

				Mark Eichin
			<eichin@athena.mit.edu>
		SIPB Member & Project Athena ``Watchmaker''