Date: Tue, 15 Apr 1997 16:33:08 +0100 From: Matthias Paul Subject: Re: Usage of directory entries To: opendos-developer AT delorie DOT com Reply-to: Matthias DOT Paul AT post DOT rwth-aachen DOT de Message-id: <7BDBB14DFE@reze-1.rz.rwth-aachen.de> Organization: Rechenzentrum RWTH Aachen MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT On Tue, 15 Apr 1997, Mark Habersack wrote: > > Using some more 'intelligent' algorithms > > also decoding the context of an entry, it should be possible to > It is possible to tell the LFN from SFN entry by just looking at > the starting cluster field which in LFN is always zero. To > confirm we're dealing with LFN it is also possible to check the > attributes of an entry: 0xFF for LFN. Looking at a cluster of zero is a rather nice idea, but probably won't work any more with FAT32 partitions, where the start-cluster word is only the low-word of the actual start cluster. We'd need at least some more plausibility checks. Unfortunately, a set volume attribut is also used by DELWATCH... If we cannot develop some nice fitting decoding algorithms for the several combinations of usage, we might need to modify DELWATCH, rather than 'standard' LFNs. Changing DELWATCH might also break UNDELETE, DELPURGE, CHKDSK, and the file-io in the kernel. For future OpenDOS issues, we should at least introduce a flag/type byte in the directory entry, describing the future usage of a directory entry under OpenDOS, and the kernel and all affected utilities would need several case decisions for 'old' and 'new' variants of usage. Doing so, we'd need to 'guess' the actual usage of an entry only for old entries not using this type byte... > > decide LFN-entries from normal directory entries, while still > > maintaining compatibility with "DR DOS"-features, which use the > > same entries, but in a different context. However, for full > > LFN-support we should better slightly modify the usage of these > > structures... > I've just started to write a TSR to provide LFN support. As soon as > it has some shape, I will make it available for all to review and > improve. Hey, now that's good news!!! ;-) Are you utilizing DPMS? > > More detailed info on the usage of these structures can be found in > > NWDOSTIP.TXT (from my MPDOSTIP.ZIP) and in Ralf Brown's interrupt > > list INTER53+. > Great info! Thanx! :-) Well, INTER53 still does not cover LFN in full details, but INTER54 or INTER55 should become much better... ;-) Also, I'm just writing new LFN infos down for Ralf, you'll get a CC: if ready for release. > P.S. Could you send me MPDOSTIP.ZIP by email? I was unable to > ftpsearch it anywhere?? Opps, it is located on several FTP servers in Germany, namely the three sites directly supported by me are: ftp://ftp.rhrz.uni-bonn.de/ ftp://ftp.uni-stuttgart.de/ ftp://ftp.leo.org/ also via http://www.leo.org/ Most of the time, the last one (leo) is several weeks in back with its updates, since they have to create some extra info files for each release. I would recommend you looking at my web-page - see the URL in my signature (I thought, I had mentioned this serveral times in the past). There you can find MPDOSTIP.ZIP and several other tools, some of them especially written for DR DOS, Novell DOS, and OpenDOS. Also included are links to other OpenDOS info providers and related programs. Of course, if you cannot get the file from there (for what reasons ever), I could also send it to you by email, but does your in-box allow for mega-bytes of mails? (The current version of MPDOSTIP.ZIP as of 1997-04-14 is ca. 525KB zipped...) However, if you run into any problems when downloading my files, please email me again, and I will sure help you. Bye, Matthias ------------------------------------------------------------------ Matthias Paul ! My eMail address has changed. For some time ! Ubierstrasse 28 ! mails to former ! D-50321 BRUEHL ! will be forwarded to the new address. ! eMail: WWW : URL: http://www.rhrz.uni-bonn.de/~uzs180/mpdokeng.html ------------------------------------------------------------------