bridger@Rand-Unix.ARPA (09/25/84)
From: Bridger Mitchell <bridger@Rand-Unix.ARPA> Under CP/M 2.2,..., files written by random access may contain 'holes'-- logical data records that have never been written. A number of utilities don't seem to handle this situation well. PIP copies only at most some of the written records, missing those in higher directory extents. SD reports 'virtual' file size -- the number of blocks (in Kbytes) required if the file consisted entirely of written records. It's easy to get multi-megabyte sizes reported. What 'should' a well-behaved utility do -- i.e. what's the most useful filesize report, and which records should be copied? On first thought, it seems that the size should be number of *allocated* blocks (in Kbytes) and the copy should consist of exactly those blocks, except that in the final block only the records <= last record of file should be copied. Have other utilities already worked this out? --bridger
towson@Amsaa.ARPA (09/26/84)
From: David Towson (SECAD) <towson@Amsaa.ARPA> Bridger - You have raised an interesting question. I have never used random CP/M files with "holes", but I have read about them. I think I agree with your position on what should be copied. I wonder, however, how CP/M treats such a file if you ask the CCP to TYPE it. What goes to the screen - just the written sectors, written sectors with intervening garbage, or what? And if just the written sectors are displayed, in what order do you see them? Dave towson@amsaa