Only treat an invalid QWK msg block count (<2) as an error if there is
more than one msg header in the packet (stop saving so many worthless
data/file/*.rep.*.bad files).

file lists during events.

	"read %u message scan cfg/pointers in %lu ms"
The message pointer performance problems were resolved a long time ago now
and this is just noise in the logs.

reported as "Duplicate index offset", so overlapping headers means the
index offset must not be an exact duplicate, so we don't report the same
corrupt message parameter (its offset into the header file) 2 different
ways.

is 0 (logic moved from sbbs_t::postmsg()).

as it is easy for a sysop to accidentally hit [D]elete and not realize it.

packets.

So I created a single add_msg_ids() function (mostly migrated from postmsg.cpp)
to make this incredibly easy so any/everywhere that adds a message to a message
base should be able to call it. This function also creates reply-IDs (when
relevant) and program ID (Fido-PID) header fields:
- FIDOPID (always)
- FIDOMSGID (if an echomail post)
- FIDOREPLYID (if an echomail reply-post)
- RFC822MSGID (always)
- RFC822REPLYID (if a reply)

If confirm_email_address = false in the [login] section of modopts.ini, do not
ask the user to confirm their email address before sending password.
Do not display the email address back to the user (in case they did not already
know it).
Send the user a telegram for each failed login attempt (using user alias, not
number).
Send the user a telegram when their account info (password) was requested.

If the user.dat file is already opened in deny-none mode (as is the norm),
another process (e.g. useredit.exe) cannot simultaneously open it in exclusive
mode.

the read buffer may be utilized more efficiently if you do *not* perform an
(unnecessary) seek operation just before the read. By eliminating unnecessary
seeks in smb_getmsgidx() and smb_getmsghdr(), I was able to reduce the time
required to load 9000+ index and header records over a network (Samba/CIFS)
share from 15+ seconds to less than 5.

(blank names or internal codes) from the total saved to xtrn.cnf.

read the entire index file in a single read operation

This has not actually been an observed problem, but I just realized while
working on smb optimizations that this condition would not be reported by
chksmb.

Use more verbose names for global msgs.cnf properties (don't mimic the
scfgdefs.h names).

        file-prots
        file-extrs
        file-comps
        file-viewers
        file-testers
        file-dlevents
... using cfglib.js (wow, that was easy!)

locally on your BBS without using X/Y/ZMODEM (or other) actual file transfer
protocols? Well, now you can. Install this simple script
(using 'jsexec localcopy install') and you (sysops only) will get a new
transfer protocol option for "local" file transfers, which are really just
file copies between different locations on your BBS/server's file system(s)
quickly and without any special terminal capabilities (e.g. X/Y/ZMODEM).

- loginAttemptListCount()
- loginAttemptListClear()
- loginAttempts()

Now returns 0 on list lock faliure: loginFailure() and loginBanned().

events when not in "appy-time".

So now if you need to clear the failed-login/temp-ban list, just
   'touch /sbbs/ctrl/clear'
Other changes to the semfiles:
The terminal server is now identified as "term" rather than "telnet" in the
(optional) server/protocol specific semfile naming.
Removed support for the old telnet.rec semfile (long deprecated).

email/netmail messages. Does *not* support the "From:" kludge.
I know this function needs need to be completely replaced and it's silly to
have QWK kludge parsing logic in multiple places, but ugh, I didn't want to
rewrite this just yet.

sub-menus.

internal code suffixes or grp/lib internal code prefixes.
More strcpy->SAFECOPY and sprintf -> SAFEPRINTF replacements.
More use of uifcYesNoOpts.

if any sub-boards or directories have *blank* names or internal code suffixes
(a sign of a corrupted configuration), exclude them from the saved subs/dirs
automatically. I'm not sure how this can happen (use of cnflib.js?) - but
running "scfg -f" (force save) should fix this situation. For Android8675.

DOOR.ID file created by Synchronet.

plain-text portion (only) for MIME-encoded messages (emails). If/when there
are ever QWK readers that support MIME-decoding, I'll make this behavior
configurable. :-)
When passing mail between QWKnet nodes, leave the original MIME-encoding in
tact (if there is any).

If the modify callback function returned a pointer *within* the list item's
allocated buffer, the strcpy() would copy from potentialy free'd memory as
realloc may change the location of the heap buffer when resizing.
Fixed by allocating a copy of the returned pointer before freeing the original
list item buffer and then just assign the allocated pointer (no copying
needed). This likely will result in more heap fragmentation for modified list
items are now newly-allocated buffers rather than reallocated existing buffers
but the other option would have been to allocate a temporary copy of the string
before reallocating and then copying and that would've been a lot more overhead
than with this approach.

This likely fixes any crashes seen in recent revs of v3.17c (e.g. when
imported QWK or REP packets and text/*.can and the twitlist.cfg are parsed).
The sighting on Vertrauen was only in the Windows build and appeared when
importing QWK/REP packets.

If the attached file begins with a '^', ignore/skip that character in the file
path. This should resolve the error reported by Ken  of Section One BBS:
ERROR line 850, attachment file not found: ^C:/SBBS/XTRN/BRE777/...

SBBSecho v2.x included this work-around and it was not included in the v3
re-factoring/re-write. Of course, BRE is doing something wrong and is totally
at fault here, but no one expects BRE to be fixed... ever. :-)

i've made a few adjustments to xtrn.cpp for external.bat...

- remove unneeded redirection of dosemu output.
   still uses same log file name and stops zero
   byte file dosemu.log from being created in node
   directories.
- fix faulty "keystroke" dosemu -I command. it
   didn't actually do the "\r" but sent it as
   characters to the command line.
- added code for SBBS specific lredirs to be
   done more cleanly like dosemu's default
   autoexec.bat. only deletes redir if it exists.

His editor also appears to have removed trailling white-space from this file,
which is fine.

commit.

suitable for passing to findstr_in_list().
SBBSecho peformance improvement: don't open/read the twitlist.cfg file for
*each* imported message: just read it once during initialization (using the new
findstr_list() function of course).
Reversed course on the findstr()/trashcan() matching logic: significant leading
white-space was not backwards compatible (and was the cause of recent lost
messages in DOVE-Net) - so I decided to go a different route and support
C-style character escape sequences (e.g. \r, \n, \t, \x##, etc.) in findstr
comparison strings, so the new way to represent a leading space character
in a filter file (e.g. twitlist.cfg, name.can) would be: "\ ". So to match (e.g.
filter/disallow) all strings with a leading space: "\ *". "\x20 *" would also
work (0x20 is ASCII for "space").
Now, again, leading white-space in filter files (e.g. text/*.can, twitlist.cfg)
is ignored. <sigh>

a caller-supplied function (callback) which can modify each string or leave
the string in tact by returning NULL. Shrinking strings is trivial while
expanding strings (making them bigger) is a little more involved, probably
using the cbdata to pass a temporary buffer around to store each expanded
string or returning a new globally-valid pointer.
Removed the malloc.h #include as we haven't used alloca() here for a while.

Now log a warning about unexpected messages from the server and
unexpected disconnects, even when in quiet mode.