Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site druca.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!mtuxo!drutx!druca!dewi
From: dewi@druca.UUCP (WilliamsD)
Newsgroups: net.micro.amiga
Subject: Finding the current directory...
Message-ID: <102@druca.UUCP>
Date: Wed, 19-Feb-86 17:31:25 EST
Article-I.D.: druca.102
Posted: Wed Feb 19 17:31:25 1986
Date-Received: Thu, 20-Feb-86 08:07:38 EST
Organization: AT&T Bell Labs, Denver, Co
Lines: 39
Keywords: AmigaDOS,current directory

I'd like to bring up an issue that was discussed some time back -- finding
the current directory under AmigaDOS. As I remember it, the consensus was
that passing a null string would work. Looking at some PD C programs that
do "ls" style things, this seems to be what everybody does.

There's only one problem, though -- this doesn't work on the RAM disk!
This is what happens:

	1> cd df0:/s
	1> cd ""
	1> cd
	df0:/s
	1> cd ram:c
	1> cd ""
	Can't find

I'm waiting for my Manx C compiler to arrive and only have the assembler
package to play with at the moment (with no ROM Kernel docs yet!) so the
scope of my experimentation is limited -- especially with the worthless
assembler package documentation (I seem to have purchased a debugger 
called romWack but it's not mentioned anywhere!).

Some of the data structures in libraries/dosextens.i seem to hold a
lock on the current directory (from memory, there's a field called
pr_Currentdir or somesuch). The only problem is getting at it -- does
anybody know if DeviceProc can return info on the *calling process* ?

How about putting a GetCurrentDir() call into the dos library in the
next release -- trying to figure it out by relying on side effects is
kind of silly.
			Dewi Williams
			..!ihnp4!druca!dewi
-- 


----------------------------------------------------------------------
	Dewi Williams.
				uucp:	   ..!ihnp4!druca!dewi
				phone:     (303) 538 4884