Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!magnus.ircc.ohio-state.edu!csn!ccncsu!mozart!barbourj
From: barbourj@mozart.cs.colostate.edu (jim barbour)
Newsgroups: comp.os.msdos.programmer
Subject: Re: Bug in TC findfirst()/findnext()!
Keywords: tc,bug,find,findfirst,findnext
Message-ID: <13255@ccncsu.ColoState.EDU>
Date: 27 Feb 91 16:59:02 GMT
References: <1991Feb25.192823.18224@ux1.cso.uiuc.edu> <1991Feb25.152152.17353@hellgate.utah.edu>
Sender: news@ccncsu.ColoState.EDU
Organization: Colorado State University
Lines: 38

>In article <1991Feb25.192823.18224@ux1.cso.uiuc.edu> gordon@osiris.cso.uiuc.edu (John Gordon) writes:
>
>	I submit the following program:
>  ...
>
>	According to the _Turbo_C_Bible_, calls to findfirst() and findnext()
> are supposed to return filenames that match *both* the given path *and* the
> given attribute.  However, the above program, which clearly specifies the
> *directory* attribute (FA_DIREC), finds both directories *and* plain files.
> Anyone have any comments/fixes/etc?
>
> John Gordon
> Internet: gordon@osiris.cso.uiuc.edu        #include <disclaimer.h>
>           gordon@cerl.cecer.army.mil       #include <clever_saying.h>

The Turbo C Bible is wrong, i.e. its a documentation error, not a bug.

Unfortunately, my TC++ manual doesn't document this "feature" very well either.

My TC++ manual says something to the effect that findfirst uses dos function
$4E to look up files.  It also says something like "for more information about
the attr field, see your dos manual."

The New Peter Norton Programming Guide to the IBM PC and PS/2 says about dos
function $4E:

"The attribute search follows a certain logic.  If you specify any combination
of the hidden, system and directory attributes bits, the search matches all
normal files, plus any files matching the specified attributes."

My Turbo Pascal 5.5 manual says "The Attr parameter specifies the special
files to include (in addition to all normal files.)"

I find this kind of documentation shortfall on Borland's part often.  It the
one biggest complaint I have again Borland.  I've never seen the Turbo C
Bible, so I don't know if it has as many omissions.

Jim Barbour (jwbarbour@clipr.colorado.edu)