Documentation updated

This commit is contained in:
Alexander S. Aganichev 2001-11-22 12:58:35 +00:00
parent f7a39d9cdf
commit d19caefb11
2 changed files with 382 additions and 279 deletions

View File

@ -12,6 +12,8 @@ ______________________________________________________________________
Notes for GoldED+ 1.1.5, /snapshot/
______________________________________________________________________
! REPLYLINK defaults changed to DIRECT.
! AREAAUTOID defaults changed to LONG.
! GoldED+ will now show hidden lines and RFC header in the message

View File

@ -1,4 +1,4 @@
<!doctype tei.2 public '-7//TEI//DTD TEI Tools 0.1//EN'>
<!doctype tei.2 public '-//TEI//DTD TEI Tools 0.1//EN'>
<!-- $Id$ -->
<TEI.2 lang='en'>
<!-- --------------------------------------------------------------------- -->
@ -202,12 +202,11 @@
<head>
The Message Database Formats
</head>
<div2>
<head>
Opus, FTS-0001
</head>
<p>
<list type=gloss>
<label>
<name/Opus/, <name/FTS-0001/
</label>
<item>
These are two variants of the same type of msgbase. It works by using
one physical file per message (<code/1.msg/, <code/2.msg/ etc.),
collecting them in a directory for each area. Depending on the
@ -215,7 +214,9 @@
way to store messages. With a clustersize of about 512 bytes, the
waste may be acceptable, but the access speed can be dramatically
slow if there are many <code/*.msg/ files, due to the file system.
Cache adjustments can improve it, but there are limits.<bl>
Cache adjustments can improve it, but there are limits.
</p>
<p>
In echomail areas, this format has a special quirk: The first message
(<code/1.msg/) is normally used to store the so-called
<q/highwatermark/. The highwatermark tells the echomail processor
@ -226,7 +227,9 @@
carefully, if at all! The highwatermark can also be <q/heated/ -
which means that it is set to the last msg in the area. This
prevents the echomail processor from finding newly entered
unscanned msgs. Use with care.<bl>
unscanned msgs. Use with care.
</p>
<p>
The variants: The <name/Opus/ format originated in the <name/Opus/
<abbr/BBS/ system. It put some <name/Fido/ undocumented (?) fields to
use as date/time stamps. The <name/FTS-0001/ (defined in FTS-0001,
@ -234,173 +237,244 @@
zone/point information for the message. To the authors knowledge, the
<name/Opus/ variant is the dominant, and the <name/FTS-0001/ variant
is doomed to oblivion. If in doubt, use the <name/Opus/ format.
</item>
<label>
<name/QuickBBS/, <name/RemoteAccess/, and <name/Hudson/
</label>
<item>
</p>
</div2>
<div2>
<head>
QuickBBS, RemoteAccess, and Hudson
</head>
<p>
This msgbase format was invented by <name/Adam Hudson/, and was first
used in his <name/QuickBBS/ package. Later several other
<abbr/BBS/'es were cloned from <name/QuickBBS/ (like
<name/RemoteAccess/ and <name/SuperBBS/).<bl>
<name/RemoteAccess/ and <name/SuperBBS/).
</p>
<p>
The format limits the total size of <code/msgtxt.bbs/ to a maximum of
16MB, which translates to about 16000 msgs of <q/average/ length.
<name/GoldED+/ automatically warns you if the limit is close to being
reached, and advises you to pack the msgbase.<lb>
reached, and advises you to pack the messagebase.
</p>
<p>
The first incarnations of <name/QuickBBS/ did not support <q/sharing/
of the msgbase. This became more and more important in later years as
multitaskers and networks got cheaper. <name/RemoteAccess/ <abbr/BBS/
was the first to implement a useful method, and later a better method
was evolved (known as <name/RA/ 1.01 or <name/RA/ 1.1x), which is now
the standard for all modern software that supports msgbase sharing.
<name/GoldED+/ fully supports the new standard of course.<lb>
<name/GoldED+/ fully supports the new standard of course.
</p>
<p>
The main virtue of this format is that it is very fast to access the
msgbase.<lb>
msgbase.
</p>
<p>
The main disadvantage is that it can be very sensitive to disk
problems, and it is a common horror story that people loose their
entire msgbase because the disk developed bad clusters or some
program went berserk and messed up the msgbase files.
</item>
<label>
<name/Goldbase/
</label>
<item>
</p>
</div2>
<div2>
<head>
Goldbase
</head>
<p>
This is an enhanced version of the <name/Hudson/ format, introduced
in <name/QuickBBS/ 2.80 by the <name/QuickBBS/ group.<lb>
in <name/QuickBBS/ 2.80 by the <name/QuickBBS/ group.
</p>
<p>
The <name/Goldbase/ format removes the 16MB size limit and allows up
to 500 message areas instead of the 200 in <name/Hudson/.
</item>
<label>
<name/Squish/
</label>
<item>
</p>
</div2>
<div2>
<head>
Squish
</head>
<p>
The <name/Squish/ format was invented by <name/Maximus/ <abbr/BBS/
author <name/Scott Dudley/ in 1991, and was first used in
<name/Maximus CBCS/ v2.00.<bl>
<name/Maximus CBCS/ v2.00.
</p>
<p>
The use of a database for each area - instead of one file per msg,
or all msgs in one big database - makes this format fast, very
safe and resistant to disk problems. Even if something messed up a
<name/Squish/ area, it can almost always be fixed and recovered,
using the <name/SQFIX/ or <name/SQREIDX/ utilities that come with the
<name/Squish/ echomail processor.<lb>
<name/Squish/ echomail processor.
</p>
<p>
A special feature of <name/Squish/ areas is that they can be
self-maintaining. You can setup a <name/Squish/ area so that it may
only contain a maximum of so-and-so many msgs, and then it will
automatically re-use the space used by old msgs when the limit is
reached, and so it will practically stop growing. It will still need
packing, but not nearly as often as a <name/Hudson/ msgbase has to.
</item>
<label>
<name/Ezycom/
</label>
<item>
<name/Ezycom/ format.
</item>
<label>
<name/JAM/
</label>
<item>
</p>
</div2>
<div2>
<head>
Ezycom
</head>
<p>
<name/Ezycom/ format is supported by <name/GoldED+/. Status is not
known.
</p>
</div2>
<div2>
<head>
JAM
</head>
<p>
<name/JAM/ format as defined in the <name/JAM(mbp)/ revision 001.
Also supports <name/CrashMail/, <name/CrashEcho/, and <name/HPT/
highwater marks.<bl>
highwater marks.
</p>
<p>
<name/GoldED+/ currently ignores revision number of header structure
and assumes that future revisions will remain backward compatible.
When creating new msgs, GoldED uses the revision 1 header
structure.<bl>
When creating new msgs, <name/GoldED+/ uses the revision 1 header
structure.
</p>
<p>
<name/GoldED+/ currently doesn't support passwords to access the
message base and/or indiviual messages. When creating a new
<name/JAM/ msgbase and/or new <name/JAM/ msgs, <name/GoldED+/ sets
the password to FFFFFFFFh. If you change an existing message which
has a password, the password is <hi/NOT/ preserved, but reset to
FFFFFFFFh.<bl>
FFFFFFFFh.
</p>
<p>
The <name/JAM/ lastread file is designed such that is has to be
searched for a userid/usercrc, because one cannot assume that the
records are in a specific order. The <name/JAM/ <abbr/API/ searches
the userid field. However <name/GoldED+/ searches for the usercrc,
not the userid. In any case, it seems that <name/RemoteAccess/ 2.x
sets both the userid and usercrc to the same value.<bl>
sets both the userid and usercrc to the same value.
</p>
<p>
<name/GoldED+/ currently doesn't support the escaping described in
the <name/JAM/ specs. The specs state that the current revision of
<name/JAM/ does not support it either, so it seems to be not great
loss.<bl>
loss.
</p>
<p>
<name/GoldED+/ does not supports the following <name/JAM/ subfields:
<q/Force pickup/, and <q>Msg may not be displayed to user</q> (always
displayed).<bl>
<name/GoldED+ optionally may use tricky method invented by
displayed).
</p>
<p>
<name/GoldED+/ optionally may use tricky method invented by
<name/Odinn/ and aprooved by the <name/JAM/ developers when deleting
messages. The configuration keyword
<ref target=JAMHARDDELETE><kw/JAMHARDDELETE/</ref> specifies which
method to use.
</item>
<label>
<name/PCBoard/
</label>
<item>
<name/PCBoard/ v15.x format.<lb>
</p>
</div2>
<div2>
<head>
PCBoard
</head>
<p>
The supported <name/PCBoard/ messagebase corresponds to the v15.x of
<name/PCBoard/ <abbr/BBS/.
</p>
<p>
<name/GoldED+/ is aware of <name/PCBoard/ v15.x extended headers in
the message text. The <gi/TO/, <gi/TO2/, <gi/FROM/, <gi/FROM2/ and
<gi/SUBJECT/ extended headers are directly supported and
<q/swallowed/ when reading a message. Other extended headers are
currently treated like normal message text and is therefore not
hidden to the reader.<lb>
Passwords are not supported.<lb>
The Private, Received, Crash and Direct attributed are supported.<lb>
hidden to the reader.
</p>
<p>
Passwords are not supported.
</p>
<p>
The <gi/Private/, <gi/Received/, <gi/Crash/, and <gi/Direct/ attributes
are supported.
</p>
<p>
<name/GoldED+/ reads the <code/pcboard.dat/ file to determine whether
to use E3h or 0Dh (<gi/CR/) as the line/paragraph termination
character when reading and writing message text.<lb>
character when reading and writing message text.
</p>
<p>
The mail waiting flags are updated when you write to people that are
named in the userbase.<lb>
named in the userbase.
</p>
<p>
When changing a message, the new edition is saved as if it were a new
message, with a new message number, and then the old edition is
deleted. This behaviour is consistent with the way <name/PCBoard/
itself works when changing a message. The old edition will still
exist with the <gi/DEL/ attribute.
</item>
<label>
<name/AdeptXBBS/
</label>
<item>
</p>
</div2>
<div2>
<head>
AdeptXBBS
</head>
<p>
The implementation is based on the documentation in version 1.05,
experimentation and questions to the authors. Thanks go to
<name/Frank Jacobberger for the initial testing and prodding of the
authors to answer <name/Odinns/ questions :-)<bl>
<name/Frank Jacobberger/ for the initial testing and prodding of the
authors to answer <name/Odinns/ questions :-)
</p>
<p>
The <name/AdeptXBBS/ format does not have a quick method of finding
deleted messages via the index file. This means that deleted messages
left in the messagebase marked with the <gi/DEL/ attribute.<bl>
left in the messagebase marked with the <gi/DEL/ attribute.
</p>
<p>
Mixing of netmail and echomail or other types of mail in the same
area is not directly supported. If an area is setup as both netmail
and echomail, <name/GoldED+/ will treat it as netmail. If an area is
neither netmail nor echomail, <name/GoldED+/ wil treat it as a local
area.<bl>
area.
</p>
<p>
<name/GoldED+/ detects <name/Usenet/ (newsgroup) and <name/Internet/
e-mail areas in the <name/AdeptXBBS/ setup and uses them as such, but
it has not yet been tested if <name/GoldED+/ and <name/AdeptXBBS/ are
compatible in the way they store and process <name/Internet/ header
information.<bl>
information.
</p>
<p>
The <name/AdeptXBBS/ personal mail feature (the index in the
<code/Personal_Mail/ directory) is supported for mails you write to
other users on the <abbr/BBS/. However, personal mail for you via
this feature or by other means is not yet supported.<bl>
this feature or by other means is not yet supported.
</p>
<p>
The replylinking method used by <name/AdeptXBBS/ (whatever the method
is??!) is not yet supported. This means that links to other messages
are missing.
</item>
<label>
<name/WildCat!/
</label>
<item>
</p>
</div2>
<div2>
<head>
WildCat!
</head>
<p>
The <name/WildCat!/ 4.x format does not have a quick method of
finding deleted messages via the index file. This means that deleted
messages left in the messagebase marked with the <gi/DEL/
attribute.<bl>
attribute.
</p>
<p>
WildCat features which are not supported yet:
<list type=simple>
<item/The userbase./
<item/The message unread chain./
<item/The message from/to title./
<item>The message from/to title.</item>
<item/The message network name./
<item/The message internal and external attach./
</list>
</p>
<p>
There is a keyword
<ref target=WILDCATUSERNO><kw/WILDCATUSERNO/</ref>, which works just
like the other <kw/>&lt;msgbase&gt;USERNO</kw> keywords. By default
@ -409,31 +483,38 @@
or you are sharing the messagebase with others, you may have to
change this user number. The userbase is currently not supported
(i.e.: you can't set the number to -1 to let <name/GoldED+/ find the
correct user and lastread automatically).<bl>
correct user and lastread automatically).
</p>
<p>
<hi/Note/: The <name/WildCat!/ support is currently not very
well-tested, so use it with caution. Limited testing shown that
<name/GoldED+/ not damages the messagebases, but you should probably
test it on less important areas and/or make backups until it is
determined that it is safe.
</item>
<label>
<name/Synchronet/
</label>
<item>
</p>
</div2>
<div2>
<head>
Synchronet
</head>
<p>
<name/Synchronet/ message base as described in <name>Synchronet
Message Base Specification</name> version 1.21. <name/GoldED+/
currently linked with <name/SMBLib/ taken from the <name/Synchronet/
<abbr/BBS/ sources slightly adopted to the <name/GoldED+/ internal
definitions.<bl>
definitions.
</p>
<p>
Just like <name/WildCat!/ format, <name/Synchronet/ format does not
have a quick method of finding deleted messages via the index file.
This means that deleted messages left in the messagebase marked with
the <gi/DEL/ attribute.<bl>
the <gi/DEL/ attribute.
</p>
<p>
<hi/Note/: The <name/Synchronet/ support is currently not tested at
all.
</item>
</list>
</p>
</div2>
</div1>
<!-- --------------------------------------------------------------------- -->
<div1>
@ -1024,7 +1105,7 @@ RDDT path.rpt -n2:5020/604.19 -l2:5020/604 2:5020/604.19 -p</eg>
</list>
</p>
<p>
When the using <ref targe=AREAFILE><kw/AREAFILE/</ref> feature to read
When the using <ref target=AREAFILE><kw/AREAFILE/</ref> feature to read
external area configuration from other programs, the individual
<ref target=AREAFILE><kw/AREAFILE/</ref>'s may use specific environment
variables to find the files. Please read the
@ -1331,7 +1412,7 @@ ENDIF</eg>
the <term/4D/ format <ident>zone:net/node.point</ident>. Most modern
mailers and mail processors now supports the <term/4D/ format,
but if you are a point, you should always consult your boss about
which format to use.<bl>
which format to use.<lb>
The optional <ident/@domain/ part can be used to specify a
<term/fifth/ dimension to the <term/4D/ address. It is normally not
necessary to specify a domain. Domains are never shown in the header.
@ -1409,28 +1490,28 @@ ADDRESS 2:236/77@fidonet ; Node address with domain</eg>
following criterias:
<list type=simple>
<item>
it's a <kw/ROBOTNAME/
It's a <ref target=ROBOTNAME><kw/ROBOTNAME/</ref>
</item>
<item>
it's a <kw/USERNAME/
It's a <ref target=USERNAME><kw/USERNAME/</ref>
</item>
<item>
it's the <kw/WHOTO/
It's the <ref target=WHOTO><kw/WHOTO/</ref>
</item>
<item>
it's the mailinglist's sender address
It's the mailinglist's sender address
</item>
<item>
it's an UUCP
It's an UUCP
</item>
<item>
it's already an email address
It's already an email address
</item>
<item>
address/aka is unknown
Address/aka is unknown
</item>
<item>
it's an <kw/ADDRESSMACRO/
It's an <ref target=ADDRESSMACRO><kw/ADDRESSMACRO/</ref>
</item>
</list>
</item>
@ -1544,14 +1625,9 @@ ADDRESS 2:236/77@fidonet ; Node address with domain</eg>
<ref target=NAMESFILE><kw/NAMESFILE/</ref> for details. However, you
should not use the syntax with the attributes in the <code/names.fd/
file, because <name/FrontDoor/ and <name/Maximus/ do not know this
syntax.
</item>
<label>
Bugs:
</label>
<item>
When used in the <name/Internet/ area, name should <hi/not/ contain
<q/@/ as the first character in the <ident/name/ field.
syntax.<lb>
When used in the <name/Internet/ area, <q/@/ as the first character
in the <ident/name/ field may be omitted.
</item>
<label>
Processed by:
@ -1578,7 +1654,7 @@ ADDRESSMACRO dn,@INTERNET/david@csource.oz.au,2:241/999]]></eg>
</item>
</list>
</div2>
<div2>
<div2 id=ADEPTXBBSPATH>
<head>
ADEPTXBBSPATH
</head>
@ -1690,7 +1766,7 @@ ADDRESSMACRO dn,@INTERNET/david@csource.oz.au,2:241/999]]></eg>
<item>
Your network address, <name/FidoNet/-style. More than one address can
be specified if you are member of more than one network. The keywords
<kw/ADDRESS/ and <ref target=AKA><kw/AKA/</ref> can be used
<ref target=ADDRESS><kw/ADDRESS/</ref> and <kw/AKA/ can be used
interchangeably.
</item>
<label>
@ -1705,7 +1781,7 @@ ADDRESSMACRO dn,@INTERNET/david@csource.oz.au,2:241/999]]></eg>
the <term/4D/ format <ident>zone:net/node.point</ident>. Most modern
mailers and mail processors now supports the <term/4D/ format,
but if you are a point, you should always consult your boss about
which format to use.<bl>
which format to use.<lb>
The optional <ident/@domain/ part can be used to specify a
<term/fifth/ dimension to the <term/4D/ address. It is normally not
necessary to specify a domain. Domains are never shown in the header.
@ -1730,12 +1806,6 @@ ADDRESSMACRO dn,@INTERNET/david@csource.oz.au,2:241/999]]></eg>
<ref target=ADDRESS><kw/ADDRESS/</ref>,
<ref target=NETNAME><kw/NETNAME/</ref>
</item>
<label>
Example:
</label>
<item><eg>
AKA 2:236/77.1</eg>
</item>
</list>
</div2>
<div2 id=AKAMATCH>
@ -1858,7 +1928,7 @@ AKAMATCH 23: 22:33/44</eg>
</item>
</list>
</div2>
<div2>
<div2 id=AKAMATCHING>
<head>
AKAMATCHING
</head>
@ -1917,7 +1987,7 @@ AKAMATCH 23: 22:33/44</eg>
</item>
</list>
</div2>
<div2>
<div2 id=AKAMATCHLOCAL>
<head>
AKAMATCHLOCAL
</head>
@ -1973,7 +2043,7 @@ AKAMATCH 23: 22:33/44</eg>
</item>
</list>
</div2>
<div2>
<div2 id=AKAMATCHNET>
<head>
AKAMATCHNET
</head>
@ -2057,7 +2127,7 @@ AKAMATCH 23: 22:33/44</eg>
The <ident/programname/ parameter defines name of the software which
wants to read its configuration from <name/GoldED+/ configuration
file. <name/GoldED+/ itself will ignore <kw/APP/ lines just like
<ref target=REM><kw/remark/</ref> lines.
<ref target=REM/remark/ lines.
</item>
<label>
Processed by:
@ -2069,7 +2139,7 @@ AKAMATCH 23: 22:33/44</eg>
See also:
</label>
<item>
<ref target=REM><kw/Remark/</ref> chapter
<ref target=REM/Remark/ chapter
</item>
<label>
Example:
@ -2089,7 +2159,7 @@ APP OtherProg IRQ 5</eg>
Synopsis:
</label>
<item>
<kw/AREA/ <ident>&lt;echoid&gt; &lt;<q>desc</q>&gt;
<kw/AREA/ <ident>&lt;echoid&gt; &lt;<![ CDATA ["desc"]]>&gt;
&lt;msgbase&gt;&lsqb;type&rsqb; &lt;location&gt; &lsqb;akano&rsqb;
&lsqb;attrs&rsqb;</ident>
</item>
@ -2268,21 +2338,51 @@ APP OtherProg IRQ 5</eg>
</item>
</list>
</div2>
<div2>
<head>
AREAAUTONEXT
</head>
<list type=gloss>
<label>
Synopsis:
</label>
<item>
<kw/AREAAUTONEXT/ <ident>&lt;YES/NO&gt;</ident>
</item>
<label>
Description:
</label>
<item>
This keyword controls where to put cursor in arealist on startup
and after exiting from an area you have been reading.
</item>
<label>
Parameters:
</label>
<item>
If enabled, <name/GoldED+/ will automatically jump to the first
marked area in the arealist on startup, and the next marked area
after exiting from an area you have been reading.
</item>
<label>
Default:
</label>
<item>
<ident/YES/
</item>
<label>
Processed by:
</label>
<item>
Mail reader.
</item>
</list>
</div2>
<!-- finished here -->
<div2>
<head>
AREAAUTONEXT &lt;(yes)/no&gt;
</head>
<p>
If enabled, <name>GoldED+</name> will automatically jump to the first
marked area in the arealist on startup, and the next marked area after
exiting from an area you have been reading.
</p>
</div2>
<div2>
<head>
AREACATCHUPREAD &lt;(yes)/no&gt;
@ -5968,7 +6068,7 @@ AREASEP &excl;C "Group C" C Local]]></eg>
editing a mail in the internal message editor.
</p>
</div2>
<div2>
<div2 id=JAMHARDDELETE>
<head>
JAMHARDDELETE &lt;yes/no&gt;
</head>
@ -6611,6 +6711,7 @@ AREASEP &excl;C "Group C" C Local]]></eg>
merit and should not be used. It is provided for cosmetical
purposes and backward compatibility only.
-->
<p>dummy text</p>
</div2>
<div2>
<head>
@ -7884,7 +7985,7 @@ AREASEP &excl;C "Group C" C Local]]></eg>
Since version 2.50, a completely rewritten implementation is used.
</p>
</div2>
<div2>
<div2 id=SQUISHUSERNO>
<head>
SQUISHUSERNO &lt;index&gt;
</head>
@ -8645,7 +8746,7 @@ AREASEP &excl;C "Group C" C Local]]></eg>
This keyword can be used globally and in a Random System group.
</p>
</div2>
<div2>
<div2 id=WILDCATUSERNO>
<head>
WILDCATUSERNO &lt;userno&gt;
</head>