by Jeffery Foy
It would be safe to assume that if you’re reading this article, you use or have used a QWK-compatible offline mail reader. The QWK format has emerged as the format of choice due to the relatively small size of QWK mail packets as compared to an equivalent ASCII text file.
As most users of offline mail readers know, the QWK format was designed by Mark Herring (Sparky) of Sparkware. While Mr. Herring did design the format, he only gave very sketchy details as to the specifics of the format. This is quite understandable as he is a very busy person. That is the reason why I’m writing this article.
In it’s most basic form, a QWK file is simply a compressed file. In almost all cases, the QWK file has been compressed with PKZIP from PKWARE. With most mail doors, you can usually choose your favorite archiver so your QWK file may not be in PKZIP format.
Within the compressed QWK file are quite a number of other component files. We’ll start with the one called CONTROL.DAT since it is the easiest to describe. It is an ASCII text file so if you have one handy, you can follow along.
Generic BBS ; Line # 1
Seattle, WA ; Line # 2
206-555-1212 ; Line # 3
Joe Sysop, Sysop ; Line # 4
00000,GENBBS ; Line # 5
01-01-1991,00:00:00 ; Line # 6
MARY USER ; Line # 7
MENU ; Line # 8
0 ; Line # 9
0 ; Line #10
254 ; Line #11
0 ; Line #12
Main Conf ; Line #13
... ; Line # x
254 ; Line # x
Last Conf ; Line # x
HELLO
NEWS
GOODBYE
Lines 12 and 13 are repeated for as many conferences as listed in line 11.
Anything you see after the last conference name can be ignored as that information isn’t usually provided by mail doors. One exception to this is the Markmail door.
Now we’ll talk about the message file itself. If you haven’t guessed by now, it is the MESSAGES.DAT file. This is, quite obviously, the largest file in the .QWK packet.
MESSAGES.DAT is organized very specifically into 128-byte records. The first record is the Sparkware copyright notice. The rest of the record after the copyright notice is filled with blanks (spaces). To maintain compatibility with Sparky’s Qmail Door, all mail doors reproduce the copyright notice exactly.
Following the first record begins the “meat” of the message file. Each message included in the file consists of a header followed directly by the message text itself. First we will describe the header:
Header Field
Position Length Description
-------- ------ ----------------------------------------
1 1 Message status byte
' ' = public message which hasn't been
read
'-' = public and already read
'*' = private message
'~' = comment to sysop which hasn't
been read by the sysop
'`' = comment to sysop which HAS been
read by the sysop
'%' = password protected message that
hasn't been read (protected by
sender of message)
'^' = password protected message that
HAS been read (protected by
sender of message)
'!' = password protected message that
hasn't been read (protected by
group password)
'#' = password protected message that
HAS been read (protected by
group password)
'$' = password protected message that
is addressed to ALL (protected
by group password)
2 7 Message number coded in ASCII
9 8 Date coded in ASCII (MM-DD-YY)
17 5 Time coded in ASCII (HH:MM) 24 hour
format
22 25 Uppercase name of person message is TO
47 25 Uppercase name of person message is FROM
72 25 Subject of message
97 12 Message password. Usually not anything
but spaces (to denote no password)
109 8 Message # this message refers to (coded
in ASCII)
117 6 Number of 128-byte chunks in the actual
message (includes header and is coded in
ASCII)
123 1 Determines whethere a message is live
(active) or killed. 90% of the time you
won't see a killed message in a packet.
'a' = Message is active/alive (0xE1)
'b' = Message is killed/dead (0xE2)
124 1 Least significant byte of conference
number.
125 1 Most significant byte of conference
number. NOTE: This isn't in the original
.QWK format but has become the standard
due to conference numbers greater than
255. In the original format, this byte
was space-filled.
126 3 Filler bytes for future expansion.
Space-filled and usually ignored.
Following the header record comes the message text itself. The message text is simply the body of the message. To save space, the return / linefeed combination is translated to the pi character ‘c’ (0xE3). Note that the last line of the message is padded with spaces to fill out the 128-byte record.
Now we’ll talk about the *.NDX files that are included in the packet. Each .NDX file is formatted into records of 5-bytes each. The bytes in each record are formatted thusly:
Start Field
Byte Length Description
---- ------ --------------------------------------------
1 4 This is a floating point number in the MSBIN
format. This number is the record number of
the message header in MESSAGES.DAT that
corresponds to this message.
5 1 This byte is the conferece number of this
message. This byte can (and should) be
ignored as it is duplicated in the message
header in MESSAGES.DAT. This is especially
important for conferences numbered higher
than 255.
Let’s stray just a moment to talk about the MSBIN floating point format. This is the format used by the older Microsoft Basic compilers and interpreters. Most compiler manufacturers have switched to the more efficient IEEE floating point format. Therefore, we must have a method of converting to and from MSBIN format. Included at the end of this article are two routines in C that accomplish this quite easily.
Ok, let’s talk about the format of the .REP (reply) packet. Like the .QWK packet it is usually compressed. Inside the compressed archive is a file whose extension is .MSG. The filename itself is the same as line #5 of CONTROL.DAT. This is the BBSID. So, for example, if the BBSID is GENERIC, the complete filename in the .REP packet would be GENERIC.MSG.
The format of the .MSG file is almost exactly the same as the MESSAGES.DAT file with three differences:
In the first record, rather than a copyright notice, the first eight bytes are the BBSID as described above. The rest of the record is filled with spaces.
In the message header, rather than the ASCII-coded message number, we have the ASCII-coded conference number
Also in the message header, the conference number field (byte offset 124 & 125) may be filled with spaces OR the conference number.
In recent months a new file, DOOR.ID, has been added to the .QWK packet. I know very little about it but will attempt to explain it as best as I can.
DOOR.ID seems to be a method for individual doors to let the mail reader know how to add and drop conferences. It is a good idea and I hope more doors and readers can be made to cooperate with it.
Usually there are only five lines in this file. Here is a sample from one of my recent .QWK packets:
DOOR = TomCat! Line #1
VERSION = 2.9 Line #2
SYSTEM = Wildcat! 2.x Line #3
CONTROLNAME = TOMCAT Line #4
CONTROLTYPE = ADD Line #5
CONTROLTYPE = DROP Line #6
Here are the routines I use to convert to and from the MSBIN format. You may use them as you see fit - they are not copyrighted by me.
union Converter
{
unsigned char uc[10];
unsigned int ui[5];
unsigned long ul[2];
float f[2];
double d[1];
}
/* MSBINToIEEE - Converts an MSBIN floating point number */
/* to IEEE floating point format */
/* */
/* Input: f - floating point number in MSBIN format */
/* Output: Same number in IEEE format */
float MSBINToIEEE(float f)
{
union Converter t;
int sign, exp; /* sign and exponent */
t.f[0] = f;
/* extract the sign & move exponent bias from 0x81 to 0x7f */
sign = t.uc[2] / 0x80;
exp = (t.uc[3] - 0x81 + 0x7f) & 0xff;
/* reassemble them in IEEE 4 byte real number format */
t.ui[1] = (t.ui[1] & 0x7f) | (exp << 7) | (sign << 15);
return t.f[0];
} /* End of MSBINToIEEE */
/* IEEEToMSBIN - Converts an IEEE floating point number */
/* to MSBIN floating point format */
/* */
/* Input: f - floating point number in IEEE format */
/* Output: Same number in MSBIN format */
float IEEEToMSBIN(float f)
{
union Converter t;
int sign, exp; /* sign and exponent */
t.f[0] = f;
/* extract sign & change exponent bias from 0x7f to 0x81 */
sign = t.uc[3] / 0x80;
exp = ((t.ui[1] >> 7) - 0x7f + 0x81) & 0xff;
/* reassemble them in MSBIN format */
t.ui[1] = (t.ui[1] & 0x7f) | (sign << 7) | (exp << 8);
return t.f[0];
} /* End of IEEEToMSBIN */
Well, that is all there is to it! I hope this article has shed some light on the so-called “mysterious” .QWK format.
Jeffery Foy, April 1991