delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp-workers/1997/11/30/06:28:02

Date: Sun, 30 Nov 1997 13:26:34 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Randy Maas <randym AT acm DOT org>
cc: djgpp-workers AT delorie DOT com
Subject: Re: proposed fsext changes
In-Reply-To: <3.0.1.32.19971128092837.007e4c80@yacker.xiotech.com>
Message-ID: <Pine.SUN.3.91.971130132614.5422S-100000@is>
MIME-Version: 1.0

On Fri, 28 Nov 1997, Randy Maas wrote:

> At 01:24 PM 11/27/97 +0200, Eli Zaretskii wrote:
> >Shouldn't `fstat' also have a hook?
> 
> Probably.  How important is it?

I don't know how important it is, but it shouldn't be too hard to add,
and doing so will make the filesystem extensions more complete.

> -- is there any concern about how the inode number is assigned by a fsext?
> Otherwise... I'd be all too tempted to use the state pointer.  
> -- should fsext be encouraged to ignore group and user id? (open and creat
> seem to ignore this at the moment)... Or is there some some posixish
> concerns, despite dos's best attempts to defeat us?
> -- currently the device field is the drive letter.... what should be
> recommended to fsext developpers for this field?  Maybe a pointer to the
> fsext's entry point?
> -- can fstat help template these fields in (and the fsext modify them if it
> wants)?  Or is this a no-no?
> -- What are the most important fields anyway?  My guess is st_mode and
> st_mtime. 

These are all valid questions, but I don't think you need to resolve
them in order to arrange a hook for `fstat'.  Whoever *installs* the
hook, should be bothered by these questions, since the hook will have
to supply all the meaningful members of struct stat.

All you have to do is to add another member to enum of FSEXT
functions, and have `fstat' call the hook if it is installed.

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019