🚀 go-pugleaf

cornet@zen.et.tudelft.nl (Jan-Pieter Cornet) writes:

>(Oh, BTW, 4DOS-people, there wouldn't happen to be some undocumented
>4DOS replacement for DOS's find-first, find-next directory browsing
>calls?? That'd really help, I guess.)

findfirst() and findnext() have nothing to do with the shell.  They
are DOS interrupts (21/4E & 21/4F, respectively).  Unless you are
willing to invoke shell to search for each file, which would be
pathetically slow, there is nothing JP Software can do for you. 
You may use FCBs if you wish, they're faster... not any better though.

t.

>findfirst() and findnext() have nothing to do with the shell.  They
>are DOS interrupts (21/4E & 21/4F, respectively).  Unless you are
>willing to invoke shell to search for each file, which would be
>pathetically slow, there is nothing JP Software can do for you. 
>You may use FCBs if you wish, they're faster... not any better though.

BorlandC  has it's own, non-dos findfirst find-next.  I didn't do any
reading on the difference.

I think the original poster was looking for a OBJ or LIB file that
provides the 4dos wildcard capability [e.g. "*y*.*" is legal while
dos' has only y*.*].


	Nevo

-- 
Nevo Y.	 Hed              nyh@wpi.wpi.edu
++1-508-754-2491          CS student, Worcester Polytechnic Institute.
"I like my cigar too, but I take it out of my mouth from time to time" G.Marx

In article <b8zmck_.messina@netcom.com> messina@netcom.com (Tony Porczyk) writes:
>
>findfirst() and findnext() have nothing to do with the shell.  They
>are DOS interrupts (21/4E & 21/4F, respectively).  Unless you are
>willing to invoke shell to search for each file, which would be
>pathetically slow, there is nothing JP Software can do for you. 
>You may use FCBs if you wish, they're faster... not any better though.

Well.. Couldn't the 4DOS shell simply just replace the DOS-0x21 vector
with another one that checks for 0x4E or 0x4F, and redirect the calls
to findfirst/findnext to a 4DOS-routine, and all other calls to the
original DOS-vector?

-- 
kennu@mits.mdata.fi
<here's where the dumb quote is supposed to be>

In article <Bsq3rs.Dz0@wpi.WPI.EDU> nyh@wpi.WPI.EDU (Nevo Y Hed) writes:
>>findfirst() and findnext() have nothing to do with the shell.  They
>>are DOS interrupts (21/4E & 21/4F, respectively).  Unless you are
>>willing to invoke shell to search for each file, which would be
>>pathetically slow, there is nothing JP Software can do for you. 
>>You may use FCBs if you wish, they're faster... not any better though.
>
>BorlandC  has it's own, non-dos findfirst find-next.  I didn't do any
>reading on the difference.
>

That's not quite correct.  I was just looking at the source
code for firstfirst, findnext, _dos_findfirst, _dos_findnext.
Each one makes calls to the appropriate dos call (21/4e or
21/4f).  The only thing that they do on top of that is set
the DTA, restore the DTA, and translate the error message.
The only difference between the _dos_find's and the find's
is that they call different functions to translate from DOS's
error code to their own return values.

I don't know why they use different structure definitions
for the two types of functions.  They pass each structure
directly to DOS, then directly back to the caller.

					--Phil