characters as the buffer width terminated by a linefeed would cause an
extra blank line to be inserted due to incorrect end-of-line detection.

The number of lines to allocate was calculated correctly, so this resulted
in writing past the last line in the screen buffer, corrupting memory after
the allocation.  Glibc detected this condition and crashed with a double-
free or corruption message.

doesn't actually *use* the password for anything, EchoCfg uses the nodecfg_t
definition from here (which includes the 'ticpwd') and the SBBSecho AreaManager
allows remote changes of the TicPwd by linked-nodes. For PSI-Jack.

Now supports changing the message subject (with Ctrl-S).  Ctrl-S was previously for text search, which has now been changed to Ctrl-W (word/text serach).  Updated the version to 1.67.  Changing the subject requires a Synchronet development build from July 21, 2019 or newer (that build added support for RESULT.ED even for

Use GraphicsConverter to display files in 999-bullshit.xjs for now.
Stop using bullshit.ssjs.  Use settings in 999-bullshit.xjs.

editor even if the editor was configured for QuickBBS/MSGINF drop files.
Don't use the subject line from RESULT.ED when the read-only subject mode
is used.

particularly salvageable here.  I have a work in progress to replace
these, however this may not end up being necessary.  Well, "necessary".

EventEmitter a la node.js, with an inheritance helper for adding it to
your Object. Same API as in node except:
- No concept of MaxListeners
- No prependListener/prependOnceListener
(These can be added later if they're needed. Too much hassle tonight.)
Currently a legacy-style library and should maybe be updated for newer
return-style load/require type use.

Ignore the _ though.

for ports where any incoming packet calls it.

Just a tree of objects and methods so far.
Not all AGWPE commands are currently supported.

Main TODOs:

Connections and ports should have message queues.
There should be a select() thing that grabs messages and sorts them into queues.
The queues shoud likely be processed by a node.jsly set of callbacks.

was just a an outline of the Window? Apparently I set the MainForm's
DragKind = dkDock at some point, forever ago, and didn't realize it.
I thought all this time it was C++Builder's fault. :-(

than hard-code to 60 chars - better support for wide terminals.
Write the newly-defined "3rd line" (editor details) to the result.ed file
(this is used by latest sbbs for EDITOR msg header field).
Always write the result.ed file when creating a message, even if the subject
was not changed. "edit_top==5" is a strange way to determine that you're
editing a message, but okay...

Full editor name and version details (up to 127 chars), for use in the EDITOR
msg header field which is propagated over QWK networks.
From what I could find, the WWIV RESULT.ED file format only ever contained 2
lines:
<Anonymous: 0/1>
<subject>

So adding support for an optional 3rd line should be fine and backwards
compatible. Hopefully WWIV or its message editors don't define a different
"standard" 3rd line in the future. :-)

get_file_diz() was reading the contents of the DIZ/SDI file into a local
variable and not returning it or writing to the database. Poof!
Also, read the existing extended description if there is one, when updating
existing files.

The max_paragraph* settings introduced by mcmlxxix in 1.34 would break
the bot for anyone who updated without adding these keys to ircbot.ini.
(max_paragraph_length was NaN, breaking String.slice in Server_writeout
of ircbot_functions.js.)

Visual Studio 2017 - Windows XP (v141_xp) toolset

Visual Studio 2017 - Windows XP (v141_xp) toolset

Added per-user JSON files:
- Tracks which messages they've viewed
- Tracks the time when they last read a file bulletin
Added 'newOnly' key to root of bullshit.ini.  If true:
- Unread messages will be listed
- Files timestamped newer than last view will be listed
- All other files/messages will be supressed
- If nothing can be listed, Bullshit will start & exit silently
I won't even pretend to have actually tested any of this.

Do not use LOGINDISABLED with TLS connections.
Add support for CRAM-MD5 so it's RFC-3501 compliant to log in without TLS.