Message-Id:	<5.0.2.1.0.20010117211305.03654860@pop5.banet.net>
X-Sender:	usbanet DOT farley3 AT pop5 DOT banet DOT net
X-Mailer:	QUALCOMM Windows Eudora Version 5.0.2
Date:	Wed, 17 Jan 2001 23:06:59 -0500
To:	Martin Str|mberg <ams AT ludd DOT luth DOT se>
From:	"Peter J. Farley III" <pjfarley AT banet DOT net>
Subject:	Re: Fw: Patch for statfs.c
Cc:	djgpp-workers AT delorie DOT com
In-Reply-To:	<200101171759.SAA15799@father.ludd.luth.se>
References:	<5 DOT 0 DOT 2 DOT 1 DOT 0 DOT 20010116203116 DOT 027afaf0 AT pop5 DOT banet DOT net>
Mime-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

At 06:59 PM 1/17/01 +0100, Martin Str|mberg wrote:
 >According to Peter J. Farley III:
 >> Martin, would you please try WinExplorer/Properties and the various 

 >> versions of df.exe and statfs.c on the Alpha Centauri game's
 >> CDROM?  Here are the results I get on my system:
 >>
 >> WinExplorer/Properties: 483,557,376 = 472,224 * 1024 = 236,112 * 
2048
 >
 >Alas WINDOZE's Properties says: 482,476,032 = 471,168 * 1024 for my
 >Alpha Centauri disc. So it seems we have different versions. The 
label
 >on mine is SMAC-E1_0Z according to WINDOZE.

Same label as mine.  What are the most recent file dates on yours?  My 
latest 5 are:

AUTORUN  EXE       606,208  01-15-99  3:53p autorun.exe
SETUP    BMP       298,678  01-15-99  3:49p setup.bmp
DATA1    CAB       164,085  01-14-99  5:10p data1.cab
LAYOUT   BIN           353  01-14-99  5:10p layout.bin
SETUP    LID            47  01-14-99  5:10p setup.lid

Then again, you may have a European version.  Mine is the USA version.

 >Let see what Partition Magic 5.0 yields.

Agreed.

 >Linux:
 >kant:~> uname -a
 >Linux kant 2.0.33 #1 Fri Aug 21 06:21:28 CEST 1998 i586 unknown
 >kant:~> df /cdrom/
 >Filesystem         1024-blocks  Used Available Capacity Mounted on
 >/dev/hdc              141304  141304        0    100%   /cdrom
 >
 >WINDOZE's Properties:
 >144,670,720 = 141,280 * 1024.
 >
 >fil316b.zip's df:
 >141,912 kiB
 >
 >df compiled with AX=7303 and AX=7302 calls:
 >141,912 kiB
 >
 >So there's no discrepancy between df versions for this CD.
 >
 >Were we made any wiser? (If we were, I must have missed it.)

Yes we are, because your WinProperties and fil316b.zip and new df.exe 
agree.  On my system, WinProperties and May 2000 df.exe do *not* agree, 
and my test compile of your new statfs.c .

My Partition Magic 5.0 CD is *much* smaller than yours, but here are my 
results for my copy:

My WinProperties says 68,878,336 = 67,264 * 1024 = 33,632 * 2048

April 1997 df says:

O:\>f:\eudora\attach\df z:
Filesystem         1024-blocks  Used Available Capacity Mounted on
PM_40                  67888   67888        0    100%   z:/

August 1997 df says:

M:\>df z:
Filesystem         1024-blocks  Used Available Capacity Mounted on
PM_40                  67264   67264        0    100%   z:/

May 2000 df says the same as April 1997 df:

O:\bin>df z:
Filesystem         1024-blocks  Used Available Capacity Mounted on
PM_40                  67888   67888        0    100%   z:/

statfs.c from 1997 says:

O:\src\libc\compat\sys\vfs>statfs97 z:
After AX1510: bsize = 2048, free = 0, blocks = 33944.
Results for `z:':

Total blocks: 33944
Available blocks: 0
Block size: 2048

This is 67,888 blocks of 1024, the same as April 1997 df.exe.

statfs.c from Martin says:

O:\src\libc\compat\sys\vfs>statfsms z:
After AX1510: bsize = 2048, free = 0, blocks = 33944.
After AX7303: bsize = 32768, free = 0, blocks = 2102.
After AX7303+DVD chk: bsize = 2048, free = 0, blocks = 33944.
Results for `z:':

Total blocks: 33944
Available blocks: 0
Block size: 2048

AX1510 says 67,888 blocks of 1024, same as April 1997 df.exe and May 
2000 df.exe.
AX7303 says 68,878,336 = 67,264 * 1024 = 33,632 * 2048, the same as 
WinProperties and August 1997 df.exe.

I think this is beginning to establish that my system is somehow 
different or buggy.  Can someone besides Martin with Win98SE try any of 
the CD's I listed earlier, or any CD for that matter, and see if 
WinProperties returns the same results as df.exe from 
5-26-2000?  Especially if you have *NOT* upgraded to MS Internet 
Explorer 5.5, because IE5.5 updates a large part of Win98.  If others 
get identical results from these two (which my system does not), I 
would say we have proved my system is the miscreant, and somehow 
defective (gee, I wonder who's to blame for that, Mr. Bill?).

Actually, I suspect WinProperties was updated (possibly by IE5.5, 
possibly in earlier updates) to use AX7303, and that is what is being 
seen here.

I think Martin's statfs.c is OK as is, based on what we have seen so 
far.  What remains to be proven is whose cluster count (re-scaled or 
otherwise) is actually correct.  I have no tools to count the total 
clusters on a CD independently of these interrupt functions, but if 
anyone does it would be truly interesting to see the 
results.  Especially since Linux and Unix report a third count, 
different from either of the others.

---------------------------------------------------------
Peter J. Farley III (pjfarley AT dorsai DOT org OR
                      pjfarley AT banet DOT net)

- Raw text -