LA>> This is the usual case of MS's "Do as I say, not as I do", since 
 LA>> Command.Com continues to use FCBs for fast directlry deletions, 
 LA>> just as I do.

 LE>> Try it under MS-DOS v6.x. Then try it in a Win 95 or 98 DOS "window"
 LE>> (or with the system booted to the MS-DOS 7 that Win 9x runs on top of).

 LA> You might want to try that yourself...my routines have been used    
 LA> under those conditions for years.

I think that Leonard is trying to make the point that using FCBs for "fast
deletions" is pointless in those environments, because it *isn't* actually
faster that doing things the more up-to-date way using a simple
findfirst/findnext loop.

There was a thread about this in this very echo a couple of years ago.  I
cannot remember who actually sat down and measured it, but I know that it was
a 4DOS user.  If one uses the DEL command built into JP Software's 4DOS one
can see what difference the use of FCBs makes, because one can select between
using CP/M style FCBs and using the MS-DOS version 2.0 method of deleting
files by using the /Q switch.  On raw MS-DOS, the /Q switch and FCBs make a
difference to the speed of 4DOS' DEL command, but in a VDM on OS/2 Warp, in a
VDM on Windows NT, or in a DOS Box in DOS-Windows 98 (or 95, or even 3.1 with
the so-called "32-bit file access" enabled) it doesn't make any difference at
all.

On modern operating systems the idea that using FCBs is "faster" is something
that DOS programmers remember being true, but that actually isn't true in
practice since most people don't use raw MS-DOS to run their DOS programs any
more.

 LA> FCBs are too usefull for this little trick to *ever* go away.

Actually, it will go away, precisely because it is no longer useful.  Indeed,
it should have done so already.  Environments where using FCBs for wildcard
file deletion *isn't* faster have been around for at least 8 years now, and
have been in widespread use for at least 5.

 ¯ JdeBP ®

--- FleetStreet 1.22 NR
114/477
143/1